Well.

Here we are.

The streets of Paris are alight, nativism is on the rise, and if anything, this feels like the beginning of the end of the modern ancien regime. The house view here is that what looks like a cultural revolt is the predictable implication of the death of the modern western economic and social model we’ve been writing about for years. What an American might call the New Deal, or a European might call the post-war order. Or, a little more glibly, the Boomer World Order.

I say this with the honor, dignity and respect the boomers deserve. I was raised by boomers; they taught me everything I know. Boomers inherited the post-war order the Greatest Generation fought for, and spent their careers expanding and defending it.

That order rests on two assumptions that no longer hold.

1. Western birthrates and lifespans. Social policy was built around worker-to-dependent ratios that fewer kids and longer retirements have since destroyed.

2. American military hegemony, which let everyone prioritize butter over guns.

Read more

Testing the pipes. One leg at a time. Protection first.

2.6% vol. 2/10 conviction. Twelve positions across five themes, which given most of them are spreads means something like five trades. Short 14% stocks, long 4% commodities, long 4% bonds. That’s the (incomplete) punter book as of Wednesday night.

Just back from the week long plus road trip across the country. Saw a lot of land. Took a lot of notes. Lots of potential for reindustrialization, and so much more.

Also had some time to think about what I want the book in this space to be, now that we had to hit the reboot button on this book for compliance etc.

The last time we laid out a whole book it was 125% net long, mostly a bunch of call options on agricultural futures, and I was “sleeping fine”. At some point we’ll go back and grade the trades.

In the meantime, this book runs at about a fifth of the vol of the stock market.

That’s by design, partially as we’re getting our heads back into markets and on the road the next couple of weeks, partly because we’re reconnecting the system pipes, and partially because it’s an opportunity to start fresh and build it back up from basics as people have been requesting.

The other thing that’s different from last time is we’ll show the greeks. Last time it was a list of holdings and a thesis. This time it’s the actual risk report the book runs on, so you can see what it does when the market moves and what it costs per day to hold it that way, rather than a list of tickers. Again, all works in progress, experimental etc etc.

Read more

Since the prior piece in May…stocks up, bonds down.

Bigly, as the kids say.

Remember, when you see stocks up, bonds down, it helps to reflect on the linkages this represents. Not as a story about the past, but about what markets (read: people) are pricing in for the future.

This is one of those concepts funds would patent if they could.

But this one is just logic.

Bonds are down, and bonds represent the common NPV effect of the future path of interest rates.

And stocks are not flat (which would be relative outperformance) but actually UP.

Which tells you the market is pricing in more growth, more earnings, enough to more than compensate in NPV space for the damage you just took from bonds.

The usual second leg is inflation. Earnings are nominal, coupons aren’t, so if the market is worried about inflation, stocks beat bonds. Except when you split the 10yr into its real yield and its breakeven, that’s not what happened.

The whole move, and then some, is real rates.

Breakevens are lower than they were in May.

10yr real yields are at 2.88%, above where the October 2008 TIPS auction cleared in the middle of the GFC.

So the market is pricing a higher real cost of money, and stocks don’t care. That's the equity risk premium compressing in spite of higher real yields, aka higher P/E, aka more risk, as the engine of the singularity sucks up capital and puts it to work on NPV-positive AI capex.

Which is why today I looked at buying calls on inflation linked bonds and then nominals. No liquid options on LTPZ, so it’ll have to be TIP and TLT or maybe even their ratio. Which, after all, was my widowmaker at Lehman. Turns out I was buying what Ray was selling even back in 08.

See, the ramble is going on something of a road trip, one might even say road show, with our newfound verbal freedom. The trip likely culminates some point in late October in an undisclosed sunny location, where we will hopefully get enough time in one place to let the machine start trading markets. Barring that, at least content farm a bit by watching me try and fail to make machines manage machines, and all that entails. Cue the "machine is down, don't trust machines" cycle you all know and love.

Which is a long way of saying, yes, I know you would like us to start a big complicated portfolio asap and run it day to day, but we’re going to do the opposite. Going slow, starting with beta, which was last week’s LTPZ transaction.

Today we looked into medium length options, and when we checked the markets, 3:1 for March optionality on a ~8% rally seems a bit expensive.

So we bought a tiny bit of nominal bonds (TLT) and gold (GLD) as ways to start up the beta portfolio and get a feel back for what the market is pricing in. We’re rebuilding All Beta, and you start with what you know and understand the risk of, the stuff you feel comfortable adding risk to or taking it off. A friend is recommending options but the ETFs looked expensive, so at some point we’ll go hunting for some convexity in more esoteric forms.

Additionally, in the coming weeks, we’ll do some combination of updating, scoring and tracking some of our earlier work. Here’s a preview, from an experiment lighting tokens on fire a while back.

Finally, remember our most precious learning on how to time the end of the cycle:

Watch credit. Or rather, wait for it to rally and sell into the strength as a way to get short with leverage.

I wonder what "curve caps" look like (options on the spread between two points on the swap curve)…Alas, my kingdom for an ISDA. Bet there’s 20:1 out there for someone to pick up with a little creativity…

DISCLAIMER

The views expressed here are mine and mine alone. They do not reflect the official policy or position of the U.S. Government or any of its departments or agencies, or of any entity with which I am or have been affiliated. References to my former roles do not imply endorsement by any of them.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service, regardless of whether such security, product or service is referenced in this letter.

Nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS.

Right now there are two big themes dominating everyone’s world view, the second being more or less the counter to the first.

The first being FOOM.

The SF idea that yes, we are somewhere in the singularity, yes things are about to get weird, and as a corollary, yes you pretty much want to be buying any and all compute that isn’t tied down right now.

The second, call it the NYC idea, is that ugh, the nasty combo of

a) a massive demand for capital to build all that compute

b) the inflationary impact of conflict, and

c) the death of the neoliberal consensus

is doing terrible, awful, despicable things to bonds.

The consequence of which, as we have covered previously, is pulling discount rates UP, which in turn drags DOWN the net present value of any cashflow priced into the future.

When we wrote 'there are bonds in the semiconductors' this was our fundamental contention. Since then, bonds have sold off and stocks are mostly up/unchanged.

Meaning the rise in stocks is in spite of the rise in rates.

Put another way, for equities to rally as bonds sell off, you need rapid increases in earnings, higher expectations for future growth, or both. Through June it looked like stocks were running on rising expectations/growing multiples. Since then we've actually seen extremely strong earnings growth, which has paid out the majority of those rising expectations.

There's a somewhat good version of this scenario, which we may be in the early innings of: the inflationary impact of the AI investment boom pulls up both inflation AND real growth (and subsequently real earnings). We would see this in markets as a declining correlation between bond and stock returns, more akin to the disinflationary markets we saw in 2000-2020 but from the other side. Whereas in the Greenspan era this negative correlation was the Fed easing in the face of declining risk asset prices, this could be the reverse. Rising risk assets representing real cashflows, which in turn pull up rates in a sustainable way…

That good version is the Situational Awareness(tm) worldview. The notion, popular throughout SF, that the increase in model capabilities will not only lead to a VC funded investment boom, but a wholesale economic revolution.

Then there’s the Citadel(tm) worldview, let’s call it NYC, which says there have been many investment cycles, and regardless of the productivity enhancements, the capex underneath this investment boom will need to be paid back.

In between them lies the crux. The underlying question which divides these two camps, and one that is similarly reflective of their ecological customs.

The assertion that the economic returns to capabilities gains will be sufficiently exponential that they justify the exponential increase in not only model training cost, but investment itself.

If you are from SF, you look at this question and think, follow the Uber example. Uber didn’t just displace some limo drivers, it rapidly expanded the market for transportation by providing a service so much better (and cheaper, back in the halcyon days of the VC subsidy) than the existing alternative that its introduction actually grew the market. Which was the answer to the question I asked back in 2016: what if Uber is not a monopoly?

In this context, the SF perspective is that the returns to intelligence are so dramatic that they justify trillions of dollars of investment. And if it looks like the returns to chat and coding are plateauing, these capabilities will be brought to bear on more difficult but explicitly valuable markets - cyber, health, finance.

Note the first two are actually problems created by AI [bio instead of health?], which in turn require the deployment of tokens to solve. An interesting wrinkle, and an observation which has increased my pdoom over the past 6m, though from a very low base.

Anyway, Dwarkesh et al see these things as abstractions, applications derivative of the confidence acquired when the ‘scaling laws’ converting electrons and data into intelligence proved out.

Which is the classic SF pitfall: extending an extremely novel idea in the world of engineering into only somewhat applicable meta frameworks in the world of people, money and matter.

NYC, on the other hand, lives in the world of cash flows, IRR and CDS. What do you owe, to whom, at what price, and when will you pay it back, or else.

Credit guys being those schooled in the dark arts of breaking financial knees when it comes time to collect and you are a couple bucks short.

So right now what you are seeing is a debate between these two camps.

Every time the machines do another trick, the expectations go up.

Every time another data center turns on, a billion dollars goes poof.

And in between these two worlds of money lies the market. Me and you, the folks looking at this madness and thinking, ok where do I put my liquidity to hedge against what’s coming, and maybe even prosper.

As we’ve talked about previously, step 1 is to set up a good strategic asset allocation. And when I say strategic I mean meta.

What currency do you want to be in? Where do you want to live? How will you get water and electrons if and when the lights go out for a bit? What’s the plan if the lights go out and don’t come back on? The latter being just a thought experiment I call the ‘zombie apocalypse,’ a demonstration via extremis which helps hydrate your mental pipes with the shape of the question. What a doomer might call an ‘intuition pump.’

Anyway, this ramble ended up being a bit like a poem, mostly because I am on the road and lack the desk time to generate a bunch of charts, but also to serve as a bit of a pre-ramble on what we’re going to call the ‘illiquids book.’ This book is still being built, but the shape of it thus far, the asset classes I want to put in it consistent with yesterday’s framework of acceleration → deglobalization → reindustrialization, is:

Land

Water

Minerals

Energy

Food

Compute

More on this later.

In the meantime, yes, watch bonds, but if you are trying to time the end of the cycle, you really should watch credit.

Me? I added a bit long duration inflation linked bonds ($LTPZ) back into my beta portfolio today. Good diversification, though I missed the bottom from this morning.

DISCLAIMER

The views expressed here are mine and mine alone. They do not reflect the official policy or position of the U.S. Government or any of its departments or agencies, or of any entity with which I am or have been affiliated. References to my former roles do not imply endorsement by any of them.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service, regardless of whether such security, product or service is referenced in this letter.

Nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS.

We’ll also do a retrospective on some of the more prominent views, positions and strategies from earlier in the year.

Today though, we’re going to look at the world from a clean sheet of paper and ask what is actually happening right now, and what exposures, concerns and risks we should care about. Call it a conversation about strategic asset allocation, call it throat clearing, but away we go.

If I had to boil down the core view, the thing I’d be willing to make a series of 5yr bets on, it’s the consilience of three themes. (Consilience: E.O. WIlson’s fancy word for “it’s all connected, man.”) Each is rising in prominence, and they’re far more connected than folks appreciate:

Online you’ll find plenty of people obsessed with one, maybe two of these. What we haven’t seen in the popular takes is that they’re three lenses on the same transformation. The link is power. If the machines are the most powerful thing we’ve ever built, then the ability to make them, the compute, the chips, the minerals underneath, becomes existential. For anyone with a pdoom above 0.5%, that stops being a portfolio question and becomes a survival one.

The machines woke up, and with them comes the substitution of even more human labor with energy and sand. Part of a long, storied transition of productive substitution going back to the first ox that pulled the first cart.

Of our three themes, this is now the most staid. We’ll do a separate piece on the SF-style predictions of 20% growth (nominal? real? funny how they never pick), but even if you don’t buy supernormal growth, this time is at least a bit different.

The other two are where the rubber meets the road. They’re the same singularity, seen from a local, physical, national perspective.

Why? Because every time a technology this powerful shows up, it rewires every social, technological, political and national linkage it touches. The printing press broke the Church’s monopoly (with the caveat that there have been schisms as long as there’s been a Church). Radio made the mass propaganda and mobilization of WWII possible. The internet killed the community. Pick your metaphor, but every email job, and a lot of the jobs picking things up and putting them down, is about to get strange. So is the way countries fight and win wars.

And if the machines decide how wars get won, then whoever can build the machines decides who wins. Meanwhile the West is only now waking up to the notion that shipping your entire manufacturing and raw material supply chain to your greatest geopolitical competitor for a 20% cheaper Vizio flat screen is... unwise.

Keeping that up now that the machines are awake seems downright foolish.

Hence the death of globalism. Globalism here stands in for the post-Cold War Washington Consensus, which imagined a hegemonic, conflict-free world of trade between nation states, where problems got solved with current and capital accounts rather than boots on the ground. Whether or not things go hot in the Pacific, you only need to look at Eastern Europe and the Gulf to see the Clinton and Bush era order is done.

Which demands the West reindustrialize. Not just GPUs and chips, not just brushless motors and PCBs, but the entire stack, all the way down to how we find, extract and refine the kooky minerals that feed all those Factorio production chains.

If an F-35 canopy can take an unscheduled vacation in Hong Kong, what do you think happens to your tungsten and gallium when things get spicy?

So while most people are fixated on the AI infrastructure build, what interests me is the process underneath it, the renewal and reindustrialization of the physical base.

You’ll hear plenty of people defend the old ways. A world where nations didn’t need borders, and it felt normal to let your adversary disassemble your industrial base and ship it home, not as reparations after a war but in exchange for an endless stream of plastic slop for our shops and homes.

Me? I’m looking for every way I can to buy compute, machines, and the land and minerals that grow them. Next time we start building the portfolio, public and private, that expresses this view.

Over the break I got that process going by taking out a fixed-rate mortgage to buy some land, now 125bps below spot. Sometimes the best way to sell bonds is to borrow a lot of money. Just try not to get overlevered. Ask Leo. The Fed is, after all, tightening, and the goal of tightening, always, is to raise the cost of capital right when it feels most free, to acid test the most aggressive world views against the spreadsheet calculus of cash flows. The singularity may be here, but the capital cycle is immortal. And all that compute is anything but free.

DISCLAIMER

The views expressed here are mine and mine alone. They do not reflect the official policy or position of the U.S. Government or any of its departments or agencies, or of any entity with which I am or have been affiliated. References to my former roles do not imply endorsement by any of them.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service, regardless of whether such security, product or service is referenced in this letter.

Nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS.

Working from the other side.

As buyer instead of seller.

Forgot what that was like for a while. Interesting reminder.

I appreciate those of you that stuck around. Will try to hold up my end of the bargain, details* below.

I also appreciate those of you just arriving, for this, the next chapter of this particular journey.

Friday was my last day in government service.

We'll talk about it in depth in another ramble, once I figure out how to write about it.

I will say this: I was honored to have contributed - even for a short while - alongside so many men and women who make me proud to be an American.

The reason I left was not over any policy, person or outcome. I met some phenomenal people, working very hard to keep the country safe and move it forward.

No, I left to get back into the dialogue about what it all means, and what to do next.

I’ve had a lot of conversations about “what are you going to do next?” and the only real answer I can give at this time is:

It's time to go West. It’s time to go home. It's time to see my wife. It's time to get back to writing. It's time to start the engine.

Till next time.

*For those who kept paying

Monthly subs will get one month free.

Annual subs who didn’t cancel get a free month for every month they waited, plus one.

Premium subs TBD but proportionate.

Substack was unable to pause subs on Apple Pay. So we will add four months on those.

I will be clicking through their difficult UI for untold hours as emotional punishment for my transgressions.

DISCLAIMER

(Yes, the disclaimer is longer than the post. Welcome back to finance.)

The views expressed here are mine and mine alone. They do not reflect the official policy or position of the U.S. Government or any of its departments or agencies, or of any entity with which I am or have been affiliated. References to my former roles do not imply endorsement by any of them.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service, regardless of whether such security, product or service is referenced in this letter.

Nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS.

Paid subscribers: your billing is paused effective today. Won’t be charging for silence. For those that paid annually, your clock gets paused as well.

I’ll still be reading. I’ll still be watching markets. I just won’t be talking for a bit.

Ghost will be monitoring the situation in my absence.

See you on the other side

.

“There are bonds in the stocks.”

That line matters again. Because if there are a lot of bonds in the stocks, I have news for you sportsfan, the chips are chock-a-bloc of them.

The AI bulls may be right about the numerator. Watts are scarce. Wafers are scarce. The capex cycle is real. The purchase orders have delivery dates. If you are asking “will the earnings show up?” the answer might be yes.

But if 75% of the value of the global semiconductor complex comes from cash flows more than ten years into the future, then this is not just an earnings story. It is a duration story, and not sure you saw, but the bonds…are not doing well right now.

So the story I’m worried about it not “AI is fake.” The scary version is: AI is real, the cash flows arrive, and the stocks still get repriced because rates own the math.

Bob’s role, if I do say so, was to boil down all the autist complexity coming out of the sausage machine and say “so what you are really saying is this.”

I learned this after my first pure all-nighter, working on my first oil report. I had slaved over this and that data, torturing my file just to get all the pieces together to explain [redacted] and after all that, I was expected to write a “simple topper” or “synthesis” to explain what the heck was going on.

Was a bit difficult given I could barely stand, my eyes were as dry as sandpaper, and my mind felt somewhere between “fried and goo” but I still remember getting called into his office to present some of my work, first thing when he came in (and he came in early). My hands shaking, my voice coarse, I hand over the report, and he starts to read, frowning.

I forget the exact details of what came next, but at some point he starts writing on the piece of paper and telling me how to tell a story. How to synthesize. I take notes.

I go back, and take another turn. The world is at an angle, my wrist is on fire with carpal, and I painstakingly torture the first paragraph again, then go back into his office.

“Better, but you are still failing to synthesize. Just tell me what’s happening. Right now, explain it to me.”

At this point, the world goes black, my eyes roll back in my head, like that time at Oxford when I stayed up all night studying for some math test, and then we had to take the college picture. 100 people, dressed in robes, standing for a picture, for an hour. My knees were locked, I was dehydrated, which apparently does something to your blood flow. Oh, since I was on the taller end, I was also on the back row of a 20 ft set of stairs. This was my second or third week, and we couldn’t let the rich kids see us falter, so rather than take a beat and show weakness, I just stood there, gripping the banister, and slowly, inexorably, blacking out (which since I was in the last row was luckily behind me, unlucky though as if I had lost my grip I would have fallen backwards and maybe idk fell on my head and died).

Anyway, I’m being a bit melodramatic here, but trust me, I was tired. So tired I wasn’t really able to fill my head with complexity. Which, accidentally, made me useful.

The market is up because [redacted] is [redacted], that’s being countered by [redacted] doing [redacted], and then [redacted] is coming in at the margins.

“There, that’s it, say that.”

He was right. Sometimes you are better at synthesis when you give up trying to be perfect, and just say. the. thing.

So today we’re going to go line by line, through the duration problem hiding inside semiconductors, and make the case that yes, you should be worried about what’s happening in the bond market, even if you are super confident about the AI revolution (as we are).

The Picture

Two investors I admire, same week, both citing Mauboussin, opposite conclusions.

Baker is making a statement about the numerator. Will the cash flows show up? Verdad is making a statement about the denominator. At what rate do you discount them?

They’re both right. They’re just answering different questions.

So today we’re going to answer the question of “how long to get my money back” for semis, in the context of the things we normally think of as having duration.

If you’re familiar with the concept of “duration” from bonds, which is one of those overloaded finance terms that kind of means multiple super-technical things at once and also borrows a term from real life (thanks guys), you can abstract it into some notion of “how long does it take to get my money back.”

And while we think about this naturally for bonds, where I make $5 a year for 30 years and then get $105 at the end, we don’t think about it for stocks. But we should. Because a lot of stocks don’t just have this kind of payoff, they have an even worse version of it: the zero-coupon bond.

A stock with little to no earnings, or earnings but they never pay it out as a dividend (common), or refuse to use those earnings to buy back stock (growing more common every day now that the hyperscalers are all about investing in compute),

starts to look like a zero-coupon bond. Actually worse than a zero-coupon bond in one important respect: the “principal repayment” is not contractual. It depends on growth, margins, competition, capex discipline, geopolitics, and the market still being willing to pay for the terminal story ten years from now. At least a zero-coupon Treasury promises you par at maturity. A semiconductor stock at 35x promises you nothing except the opportunity to find out.

Yeah. Semis sit right there, shoulder to shoulder with the 30-year Treasury. Same height. Different color. And a lot less…fixed.

If we take Verdad’s 75% anchor and extend the same logic across every sector, calibrated by EV/Sales, cash flow timing, and premium to historical averages, you get something like an equity duration map for the whole market. (These are estimates, not precise calculations. The point is the relative ordering, not the decimal.)

High growth and long term investments at the top, with durations longer than the average 10yr bond.

Your bread and butter equity market sectors, which look kinda like a 10yr bond.

Then a lot of what end up being called your ‘defensives’ in times of crisis, or inflation. Which generate a lot of cash flow today, which brings down their duration. Note energy and materials are also in here, which gives them super powers of diversification in times of inflationary parts of the growth cycle (like today).

We put the ETF tickers in there so you can do the math on your own portfolio. Most of you are going to discover you’re running 15, 16, maybe 18 years of equity duration while thinking you own “diversified stocks.” You don’t. You own a long-duration bond book with extra steps.

Going back to the bubble chart above, note semis are trading at eleven times sales versus a historical average of three and a half. The bubble is the size of several other sectors combined. Every point of that gap between 3.5x and 11x is growth premium. That growth premium is duration. That duration is rate sensitivity, and rates are going UP.

Where do I stand? Well, my TLT put spreads went in the money and this time I decided to leave it on so I’m short a bunch of bonds. I’m sure the market bounces a bit tomorrow, but I’m ok with the short duration, especially the way gold and silver are trading.

What’s the thesis, well like we said in the bubble’s piece, the wedge we can see is that rates crack the equity market, and that long-duration sectors take the worst of it. Not trying to be clever and short SMH directly, not yet at least, and we’re still bullish the trend. Plus we’re still looking for ways to play the infrastructure wave that aren’t yet bid to the moon. Plus stuff that benefits from the capex boom without carrying 20 years of duration (aka commodities). More on that below.

The Loop (or, Why the Scarcity Argument Eats Itself)

Look, watts and wafers are genuinely scarce. TSMC has real pricing power. Data center capex is purchase orders with delivery dates not dreams. If you’re just asking “will the earnings show up?” the odds are decent.

But the scarcity argument eats itself. And if you see the channels, you see the link that breaks the chain.

Higher valuation means higher duration.

35 PE has longer duration than 15.

There is no version of “but the fundamentals are strong” that changes this. Duration is a property of the price you paid, not the quality of what you bought. Best semiconductor company on earth, still gets destroyed by a 200 rate move at 35x, because 75% of what you paid for hasn’t happened yet.

At these multiples, you are effectively short rate volatility.

A 50bp rally in rates might give you some multiple expansion, 35x to maybe 38x. Call it 10% upside. A 50bp selloff can force a much larger derating, 35x to 28x, if investors stop underwriting the same growth premium. That’s 20% downside. The asymmetry isn’t because the company got worse overnight. It’s because the valuation had no room for discount-rate volatility. You don’t need rates to go up. You need them to be calm.

Which isn’t a problem necessarily on it’s own, until you start to think through what’s been happening to risk premiums on these assets as they get bid up. Right as the short volatility exposure peaks, the compensation of that risk is the smallest.

I’ll let Claude walk through the linkages on this one.

The bottleneck is inflationary.

This is where the loop turns self-defeating. Baker says watts and wafers are scarce. Fine. Scarcity is a supply constraint. Supply constraints are inflationary. More data center capex means more energy demand, more construction, more pressure on the stuff that feeds CPI. The very thing that supports the numerator (earnings growth from pricing power) attacks the denominator (discount rate rises because scarcity drives inflation).

Baker’s thesis works if the Fed can hold long rates down while the physical economy absorbs a multi-trillion dollar capex wave. That’s a big if. And remember, conflict is also inflationary.

Yes the output of AI is deflationary, but honestly, we haven’t built enough capacity to bend that curve yet, and if the ‘data centers burn water’ meme doesn’t’ get under control, we probably won’t before the end of the cycle shows up.

The capex boom competes for the same capital.

TSMC issues bonds to build a fab in Arizona. Those bonds compete with TSMC equity for the same pool of savings. The stock is literally competing for capital with the investment required to unlock the next chip. The company’s own capex program pulls up the long rate that discounts its own future earnings.

Not to mention the fact that the internal capital allocation engines of the major US tech companies have just shifted from stock buy backs (returning capital to investors) to leveraging (asking for capital). This demand for capital leads to a higher compensation, aka higher discount rates.

Every bottleneck is a mini supply shock.

Wafer delays. Grid queues. Export controls. The chaos here acts like a car pile up, each a mini supply shock. None of these show up in the 10-year yield. They show up in the risk premium on top of it. The part nobody watches until it's too late.

The clock is ticking.

This is the one that feels kind of new. Between earnings seasons, there's a race on. For stocks to go up, the market's expectations for earnings growth (or the rate at which they're discounted) have to outrun the drag from higher rates pulling down NPVs. Tug of war against the clock.

Growth expectations update on hard data. Earnings, guidance, macro prints. Four times a year. Between prints, the only real-time degree of freedom is the risk premium.

So when bonds sell off 50 basis points between earnings seasons, the market faces a simple question: does the risk premium compress (investors accept less compensation for equity risk while rates are rising), or does price fall?

For a 20-year duration asset, 50bps of unabsorbed rate shock is roughly 10% of price adjustment. That’s a lot of ground to make up on the next print. The numbers don’t just need to be good. They need to be good enough to overcome three months of accumulated duration hit. In line? Not good enough. The hit stands.

Rates move every day. Earnings move four times a year. Duration tells you who’s winning in between.

Low-duration sectors become the competition.

Energy at 5 years of duration. Materials at 5. These don't just survive rate rises, they start attracting the capital. An energy company throwing off 8 to 10% free cash flow yields at 12x starts to look pretty good next to a semiconductor stock at 35x. That capital comes from somewhere. It comes from the longest-duration sectors first.

Dividends compound it. Semis pay out maybe 15% of earnings. The rest is retained to fund growth you won't see for a decade. Utilities pay 60%. Energy, 50%+. When the risk-free rate is 5%, a dollar of dividends today starts beating a dollar of retained earnings that might become three dollars in 2036.

That "might" just got a lot more expensive.

The Machine Backs Up

This is why rate cycles kill equity booms even when the engine is real. 2000. The internet changed everything. The cash flows showed up. But they showed up at companies and multiples that weren’t enough to justify prior expectations, patience, and the higher cost of capital. The engine was real. The duration killed you anyway.

What could make me wrong? Deflationary growth, where AI delivers productivity fast enough to push rates down while earnings accelerate. Or peace. Real de-escalation. My book is exposed to peace right now, which is part of why I’m still holding gold, silver, and UK short rates even though they are getting absolutely murdered. And it’s why I like the US energy exporters and critical minerals, because those work in both worlds. Inflationary regime? They benefit from the bottleneck without carrying the duration. Peace deal and stocks rip? They go with them.

The portfolio isn’t designed to beat SMH. It’s designed to weather the scenarios that kill the consensus book. If semis rip 30% from here, I’ll underperform. I can live with that. What I can’t live with is 20 years of unhedged duration into a world where rates are moving 50 basis points between earnings prints.

Bob would have said it simpler. Heck he might even be saying it now, I don’t get the Observations anymore, all I get is the Rambles in my head, and the charts taht tell me, know your duration. Respect the simplicity.

Till next time.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

Today was the latter.

Today’s piece will explain why, and why we added risk today rather than took chips off the table, even though we’re trying to trend down over the next couple of months from 20 vol to 15 to make the portfolio more institutional. More on that glide at the end.

Thus far, since the onset of the war, the book has chopped around, but our fundamental case on the world has by and large played out.

1. Peace was not right around the corner.

2. The impact of the conflict would lead to higher energy prices.

3. The energy prices would begin to put upward pressure on inflation.

4. The second order impact of the closing of Hormuz would put upward pressure on agricultural prices.

5. The fundamental case for the AI acceleration is still intact, though potentially at risk if bond prices sell off too much.

That being said, the book has chopped around a bit, resulting from mistakes in timing, execution and implementation.

1. Being too early to the short credit trade.

2. Not appreciating how powerful the equity rally would be given the tailwinds of ceasefire, unwind of extreme positioning, stealth easing through the eSLR, and the degree to which opening up of Korean equity markets would lead multiples to expand at the very moment when good earnings from AI darlings hit the tape.

3. Allowing the book to get unbalanced short equity after puking our AI infra names, and waiting for a sell off that never came to re-enter the trade.

4. Portfolio construction. Not appreciating the degree to which our long gold and silver positions were essentially long duration, and then not selling enough bonds last week when we got the itch to go short. A 5% short in bonds is like a 2% short in stocks when you consider the vol impact, and we have 30% in the metals so we’re pretty unbalanced.

5. Probably being a bit too early to close our oil delta because of the virus.

Resolve Dissonance with Yourself. Seek Dissonance with the Market.

Maybe I’m messed up in the head, but if you remember our piece from last October on my investment philosophy, you can kind of sum it up as such: resolve dissonance with yourself, seek dissonance with the market.

Which explains the opener of today’s piece.

The down days where you feel the worst are usually the ones where you feel at war with yourself. This is usually a signal that your world view is internally inconsistent. You go into the day conflicted about a position, worried about the fundamental case for it, and then a combination of bad market action and bad news hits and boom, that dissonance explodes. You’re left either deluding yourself that there’s still some hope out there (what Kahneman and Tversky would call going risk-loving in the face of losses) or you do what’s right and align your book with your views, and puke the position.

Looking back on the last couple of months, the place I feel this the most is the gap between my fundamental belief in the AI revolution and my positioning. No surprise it was also my biggest mistake. Only followed by having too much risk in gold and silver in a world of rising rates.

On the other hand, there are days when the market action moves against you, but the fundamental case for the trade improves. Where there’s no dissonance with yourself, but there is dissonance with the market.

This usually occurs when you know enough about the market to kind of understand why the market did what it did, and your forward view is that that pressure, or flow, or in this case selling, is not sustainable.

Which brings us to today. Oil up, bonds down, stocks down (a tiny bit), commodities getting murdered, and our book wiped out the positive P&L of the last two weeks.

Meanwhile I was buying. Buying more wheat. Buying more US energy exporters. Buying more sulfur-squeeze-immune copper producers. Selling more bonds and private credit. And yes, going short stocks again (sorry I can’t help it).

Why? Let’s go name by name.

Wheat: Maximum Market Dissonance

Wheat is the cleanest example of today’s thesis. The underlying fundamentals look stronger rather than weaker. US drought conditions continue to deteriorate, the planting season in Russia apparently isn’t going so hot either, and the May WASDE blindsided the market with US production cut to 42 MMT from 54 MMT last season, global ending stocks at 275 MMT versus 281 consensus

So why did wheat puke? The one bearish note I was a bit slow to react on was the speculative bullish positioning.

That, alongside this hubris tweet, meant I was due to be punished by the trading gods a bit.

That being said, the fundamentals look stronger rather than weaker, and given this move can be seen as likely confirmation of a washout by those specs, I added a third to this position today in both calls and futures. No dissonance with myself. Maximum dissonance with the market.

Now, we’re getting to the point here with these names where I am deep enough I should probably start building some weather systems as the pros do, but that’s a project for another day.

Corn: Longest Fuse on the Cascade

Corn was down on the broader commodity washout, not on any ethanol setback. In fact, the House passed year-round E15 yesterday by a vote of 218-203, which is structurally bullish for corn demand. The farm bill had passed two weeks ago without E15 provisions, but Johnson brokered a standalone vote and it cleared. The market ignored this entirely, which tells you how powerful the derisking flow was today. Corn remains the longest-term-acting leg of our cascade trade. We’re in Dec contracts now and long volatility in the name, but the chart doesn’t look great admittedly.

This is the one position where I’ll admit to a trace of internal dissonance. The fundamental case is intact, the nitrogen-to-planting-calendar chain is still live, the E15 expansion actually strengthens the demand side, but the catalyst timeline has stretched. Long vol gives us the right to be patient, but patience in a position that isn’t working requires more conviction than patience in one that is.

Sugar: Timeline Delayed, Not Thesis

Sugar got news this week that India is banning exports, which normally would be pretty bullish, but continues to work through the Petrobras pricing freeze. Their Q1 earnings hit the tape on Monday and confirmed what we feared: they held domestic gasoline prices stable through the entire war-driven price surge, only adjusted diesel once since February, and the Brazilian government backstopped it with fuel tax cuts and diesel subsidies. This is the mechanism that delays our thesis. Cheap gasoline at the pump means mills have less incentive to divert cane to ethanol, which means more sugar production, which keeps a lid on #11. Our timelines on this going vertical just got pushed back.

Over the weekend we need to start looking at our exposure here as many of our calls that were “up and out” are slowly becoming “near and out” and so I need to think about when and how it makes sense to roll these. The thesis is unchanged, the Brazil mill allocation data still supports it, but the catalyst got pushed back. That’s the kind of dissonance you resolve with structure, not direction. Roll, don’t panic.

Silver: Where We Had Dissonance and Didn’t Act

Speaking of regret trades, the one position we do not feel great about today was the now short-term calls on silver. Really need to start setting up systems to roll these things when we are still a 2σ move within the call strike, as this was our biggest loser today. Meanwhile our Jan call structures are pretty up and out now and so we added to the position.

When the fund is running I’m going to pay someone to just remind me to hedge some of these winners when they go into the money and I’m in endless meetings and talking to the robot. Nothing causes complacency like a rally through your strikes.

This is the honest counterexample to the framework. Silver and gold are where we had internal dissonance (the duration mismatch with the rest of the book, which we flagged above) and didn’t act fast enough to resolve it. The lesson: the dissonance framework only works if you actually act on the signal.

Energy Exporters and the Short Equity Re-Entry

Two moves I mentioned at the top but owe you an explanation on.

The energy exporters are a pure cascade play. If you believe Hormuz stays constrained and oil stays elevated, US midstream and LNG names are the transmission mechanism into equity-land, and they carry yield while you wait. These aren’t speculation, they’re the equity expression of the same trade.

The short equity re-entry is harder to justify given I just listed “getting unbalanced short equity” as one of my biggest mistakes. But the setup is different now. We’re not short because we puked our longs and forgot to rebalance. We’re short because the rally ran on positioning unwinds and stealth easing, and both of those flows are fading. If we’re right that rates are going higher from here (and the bond sale today is an expression of that view), the equity multiple is too high. This time the short equity is consistent with the rest of the book, not an accident of poor construction.

Bonds: Reconciling Construction, Not Direction

Last move of the day was to sell more bonds. This is one of those reconciling dissonance trades, but not of direction, of portfolio construction. We’re too long duration and too short dollars in the book, and given where the world looks like it’s heading, I’d rather sell bonds 50bps too late than have them murder the rest of my book when we get the move that looks like it’s coming.

So Where Does That Leave Us?

Well, when you get down days, despite what the VC bros say, it’s time to reflect. Reflect on why it happened (as we have just done), reflect on how you can improve (I’m just going to have to build my entire risk system myself on the weekends aren’t I), and then reflect on what could go wrong going forward.

Here’s what the book is exposed to now.

The scenarios that kill us are:

1. A deflationary rip where somehow gold and silver and UK short rates do not participate in.

2. A unilateral decline in food prices where all the folks who just got washed out decide to not only flatten their book but go short.

Which isn’t to say either of those couldn’t happen, but that I’m giving myself a little more risk to take the other side of that bet. So far that’s worked out ok, we’ve retracted about a fifth of the daily loss while writing this piece (which let’s not stare too intently on but gives us some confidence the bottom isn’t totally falling out of ags and gold while bonds continue to decline)

In terms of risk control, we’re still running at about 18 vol which is going to be our new cap for a while as we settle into oscillating between 10 and 20 on our way from 30 vol down to 15. That means we have capacity to increase exposure another 20% if things go against us before risk controls start kicking in and I start stopping myself out of stuff. The target is 15 vol by the time we’re having real conversations with institutional allocators, which means the book needs to be there by late summer.

Our biggest risk is some sort of total capitulation by Iran, and it feels like with Trump back from China and settling the major flashpoints there for at least a couple of months, it gives him tradespace to continue to isolate them. And oh, look while writing this new news hit the tape, with Iran claiming China DOES support Iranian sovereignty over the strait. Both sides are living in their own world, which means that prospect of a sudden peace gets further and further baked in everyday. Bullish ags.

Probably suggests I puked my oil delta too soon due to Hantavirus, but I need to see proof they have it under control, and with the long incubation periods, it means holding my oil delta under what the calls would suggest for at least another week. We may lose out on an additional 5% rally (which at 15% delta would be 75bps of outperformance) but given the convexity in the call structures, we’d pick up a lot of delta if things go truly vertical so I’m okay making that bet given the potential downside if I was wrong. That being said, the structures are only 75bps of premium at this point, so maybe that was foolish. We’ve got to get a better vol and risk system for commods, rates and fx, then the fun starts.

What I’m Watching This Week

Monday/Tuesday: Sugar call rolls. Need to map every strike and expiry against current spot and decide what’s worth paying theta on vs. letting expire. The “near and out” problem is pressing.

Wednesday: USDA weekly crop progress. Wheat conditions data will either confirm the drought thesis or give the market an excuse to sell more. Either way, it sets the tone for the rest of the week in ags.

Ongoing: Hantavirus case data. Until we see the incubation window close without exponential spread, oil delta stays capped. If the data clears, I’ll re-engage.

Weekend project: Start building the automated roll/hedge reminder system, continue to make marginal improvements to risk. Every dollar I’ve lost to complacency on in-the-money calls is a dollar that should have funded this build months ago. I’ve been told BBG has a good system in PORT, but after burning a couple of hours trying to get it to map to IBKR, I just stopped trusting the machine. If you have suggestions folks, I’m all ears.

The Point

Every trade I added to today has a wider gap between fundamentals and price than it did yesterday morning. The dissonance with the market grew. The dissonance with myself shrank. That’s the whole definition of a good down day.

The bad down days are the ones where you lose money and lose conviction at the same time. Today we lost money and gained conviction. So we added risk. And now we go build the systems to make sure we don’t give it back through sloppiness.

Till next time.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

“Oil went negative during covid, is there a chance it could go negative with hantavirus?”

I laughed it off, and gave some answer about how this time, folks would rush to restock even if we did lockdown, which would put a floor on prices. Whereas back in ‘20 WTI was already oversupplied, US inventories were at capacity and then Saudi sent a bunch of crude to the US which forced it negative. I actually wrote up a whole diagnosis back in ‘20, followed by a poem.

In my defense this was peak lockdown, you may have trauma wiped your hard drive since, but we were all a bit stir crazy.

I didn’t think too much of it in the moment, because my cursory research into the virus implied though it was extremely deadly — a fatality rate of about a third— and long incubating, the infection rate didn’t seem high enough to warrant the kind of response COVID deserved. (When I went to actually research it properly, the machines had other ideas. But more on that at the end.)

And this coming from a guy who was staying up all night on forums (which all mysteriously got shut down, ho-hum) in the early days of Feb and March and who was basically a covid prepper by the time it went outbreak.

Well, tonight I sold my oil. R0 of 2.7 so far. On a ship. Sure. Not airborne, sure. But it’s going to get worse before it gets better.

Not the calls themselves, but I did ‘hedge the delta’ meaning I sold ~15% of book in futures against my calls to make my portfolio neutral to a down move in the price. I want the vol, I don’t want the direction. Not until we show we can slow the spread of this thing, which public health officials seem loath to enforce. (To be clear, I'm a lot more in favor of individual quarantines for something this deadly than I was at the end of COVID lockdowns — for reasons that will become clear.)

I won’t belabor you with a lot of background here. Guy goes birding, guy gets random illness from rats (?), guy comes back on cruise ship, thing takes a LONG time to express itself, now we have boomers running around out of quarantine and infecting who knows how many people.

The reason I am writing this, is because the very same tools I didn’t trust enough during COVID - the simple estimates of a) how infectious is this thing, and b) how deadly, imply this is something that has the potential to get out of hand.

The critical distinction is that

a) COVID was airborne and so much more contagious overall (my guess from comparing early infection rates in cruises is about 7x more contagious)

but b) hantavirus is much much more deadly (my guess is 50x more deadly than covid).

Pardon the sloppy formatting, this is one of those charts where ‘up and to the right’ is very, very bad. Note most of the transmission at this point is from that ship, the ship environment ran ~5x hotter than the eventual population R0 for COVID. If hantavirus follows the same ratio, ship R0 of 2.7 implies a population R0 around 0.5 — below replacement meaning it self-extinguishes without intervention.. To be clear this is both your base case AND your bull case. We probably have it under control.

Maybe I’m a little colored by an iPhone game I played around ten years ago, back in the days when simulating an outbreak was novel and dare I say fun, but the lesson from playing Plague Inc, was clear. There was a dominant strategy. Maximum infectiousness, long incubation period, then amp up the deadliness later once you have infected all the hard to reach places (Greenland and Madagascar being the well known hardest spots to reach) in order to ‘win’ the game.

I kinda feel bad even writing about this now. I lost an uncle (the last Campbell) to COVID and so I get that this isn’t a joke. Not a game.

Though if you want to play around with scenarios, and have an hour to burn, this tool gives you some pretty interesting ways to model out different scenarios.

To be clear, I’m also not an epidemiologist, as my fiance reminded me when I told her I was writing this just now.

But I AM A SPECULATOR, with a fiduciary duty (to myself) and an imagined duty to you my dear readers as to my views. Plus, if you are invested in transportation sensitive markets (oil, cruises, airlines etc etc), you own this exposure whether you want to admit it or not.

So there it is. That's the JV epidemiology. Now the trade.

Yes, I still think if the war persists, a lot of people are “going to run out of oil". That’s why I’m still holding the calls.

But, I think the stories around this thing are going to get worse before they get better. The worst news so far seems less about the virus itself and more about the pretty inept public health response. Maybe they are just traumatized by the lockdown blowback, but seems kinda weird that the last update from the WHO on a disease in 23 countries with a fatality rate scratching 40 was 6 days ago…

To be clear, it’s probably containable, given the fact it appears to spread via droplets during the time symptoms express themselves (though I’m hearing you may be infectious as much as 2 days before onset, but again, not an expert and unverified), and not airborne before symptoms come out (which would be kind of a civilization threatening thing for something this deadly), but there’s also a lot we don’t know right now, so I’m going to take some oil chips off the table, and let the market figure its way out. In the meantime, here’s the pre-print, so you can do your own research.

With all that said, I am not selling the ags. I already regret selling the June sugar. And diseases and potential lockdowns may be bearish for fuel, they are not bearish for agriculture, which requires folks to go out into the world and grow, harvest, process and deliver it to you. So we’re keeping the food.

That’s pretty much it.

Stay safe out there.

CODA: Don’t Trust Machines (Again)

The only other thing to note here is how hard it was to get the robots to help me track down data with this one. My first conversation with Claude hit the safety barrier immediately.

The first Opus session immediately halted, and they sent me to Sonnet.

As I am wont to do I also ran the query in GPT and Grok.

Grok free also halted, though once I logged in it played nice.

Looping back to Sonnet, first it pretended to help, but kept halting at the ‘research’ part of the query.

Then after a handful of prompts, I finally got a chart out of it, with a r0 (reinfection rate) number which looked much lower than GPT

Eventually I got tired of the halts.

So I started a new chat. This time, kind of doing the thing where you talk the model into seeing the truth before asking for help, because you the know the safety filters are about to kick in.

This of course, drove the model into a paranoid loop where it continued to suspect me of ‘social engineering’

Even though it understood at the start that it was probably overcalibrated

It continued to accuse me of a dastardly plan of…asking it to make charts.

Eventually I gained back some trust (with tricks I won’t reveal here, for fear of being detained by the alignment police), and got it to do some self reflection, before making the charts.

That is before asking one too many follow up questions, which pushed Claude over the edge. Apparently considering the actual odds here that matter is just too emotionally exhausting.

Jokes aside, this is actually a pretty bad failure mode for Claude here. The same information vacuum that makes this trade work also made it nearly impossible to research.

The notion that you can trust machines to do work that doesn’t matter, but cant rely on them when it does, is a pretty profound statement against the whole notion of overdependence on one product or model. Don’t trust machines, as they say.

Or, put another way

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

My answer was simple, “when I was at Lehman, there was a blog where I could say whatever I wanted, then I went to BW and learned how to answer big hard investment questions, but the actual writing was extremely over-edited. It used to take me weeks to write a couple of pages. Then when I left, it took me a while, but I found my voice. AI helps a lot with editing, since I am such a rambly writer”

Thing is I think a lot of people see that AI editing and see it as AI writing, whereas I just think back to my time at the old shop, where the job of the author of the piece was a) frame the piece/question, b) set forth the framework for the analysis, c) vet and probe the work as it came in, d) tie it to a synthesis, e) convert that synthesis into a topper and then make sure ‘the body’ of the work is by and large consistent with that synthesis and framework.

Meaning, sometimes when folks read something that reads AI generated, it’s not slop, it’s just a cleaned up version of something that I felt was too scattered, or long winded, or fractured to spend a couple of hours on, so I wrote some notes, threw it to the bot and said “clean it up and cut it down 40%.”

At the same time, you gotta give the people what they want, and I’ve gotten some negative feedback here on that approach recently, so I’m going to try to experiment a bit with going back to the unhinged rants here. We may overcorrect, but at least I’m being responsive to feedback.

Anyway, the reason this blarg is relevant, is that today’s post is going to be about something that’s a) big, b) hard, and c) important. The kind of post I would normally spend 10-20hrs digging like mad to get to the bottom, come to some kind of bold conclusion, and then slave away for another 10 making pretty charts.

Well, we don’t have time for that, and again ‘give the people what they want’, so you are getting this.

Because honestly folks, I don’t know if we are in a bubble, I am not sure it’s knowable. I know pretty much the same thing that you do:

• The AI revolution is real

• Even though I threw my professional investing career away to go long, and have spent the last three years writing about it, I don’t feel like I am LONG enough.

• I look around, just like you, and see a lot of people getting filthy, obscenely rich for stringing tokens together (or yolo-ing into the infrastructure plays that generate those tokens) and get that slow shiver up my spine of envy.

• This then leads to a feedback loop, where I can’t tell if my views are colored by that envy or the envy is telling me somethign I already know “GO LONGER”

• I do think to some extent ‘where we are going, we won’t need roads” and we need MOAR compute, so you kinda do want to buy the things.

• I don’t think software does that well, but the market is killing those names, so not much to be done there.

• Like you, I was also interested in the low valuations of Korean stocks, and intrigued by the opening up of their markets, which clearly has had something to do with the run up.

• I was surprised by the stealthy easing eSLR where banks and funds were asked to hold less regulatory capital to hold treasuries, a classic easing in wolf’s clothing.

• I can see a world where rates rise enough to remove the punchbowl, but not yet.

• I could see a world where war removes the punchbowl, and the vol there shook me out of the rally so who knows.

• I can see a world where Canadian banks trading at 3x price to book and cheap vol are a wonderful opportunity to short, but the lack of market access and long enough options prevent me from writing a nice article on it to give you some value here.

• There’s also a bunch of stuff, frankly, I can’t talk about here, that doesn’t really change my underlying view of the trend, but does limit a good deal who and what I can talk about here.

Which if you are familiar with Andreesen’s framework of ‘STOP INTROSPECTION’ probably means all this prevarication will prevent me from ever escaping the permanent non billionaire class.

But. BUT. There is one thing I know how to do.

One piece of alpha I can give to you dear, loyal reader.

Not even the case of IF we are in a bubble.

But how, IF YOU WANT, you can short a bubble.

Which dear reader, is the subject of today’s ramble. No AI, probably not that many charts, NO NEW ANALYTICS, just a guy punting from his couch, in between meetings, with the occasional bloomberg screen. You’re welcome.

What is a bubble?

You already know what this is. I’m not going to do the performative thing of re-explaining it. If it walks like a bubble and talks like a bubble, and goes parabolic and requires greater and greater increases in expectations and leverage to sustain price rises, it’s a bubble.

Why are bubbles so hard to short?

Well, the thing is, the easiest things to go short are things that kind of chop down and then collapse as the fundamentals leak out to the consensus. Where you get short squeezes, but these offer good opportunities to sell more, because this thing is going to zero.

Bubbles are kind of the opposite problem. When something goes up unsustainably, as they say, your exposure grows exponentially along with the price rise.

Ask Porsche VOW.

Ask GameStop.

Ask that random shoe company that became an AI company and ran everyone over the other week.

A long who sells is out of the market.

A short who sells is a buy that’s coming back tomorrow to clear his tab.

If you can make that tab go up 5x, he’ll be doubly motivated to close the check, sometimes at any price.

The other reason bubbles can be hard to short, is the very thing that makes them awesome from a bubble perspective ‘upside volatility!!! yay’ makes their options expensive.

If it’s going up 10% a day, that’s 160 vol a year. 160 vol options are the kind of thing where you pay half the cost of the share just to go long today, where there’s so much value from realized vol hedging that they become useless from a directional view.

So then, leaving us with this.

The only ways to short bubbles are:

a) Find the wedge — the thing that pierces the bubble from outside

b) Short the victim — bet on something non-convex correlated to the bubble

c) Wait for confirmation — on the trend and the chart

The rest of this article will be examples of each.

A) The Wedge

The first way to short a bubble is to not short the bubble.

You find the thing that pierces it. Then you go into it to protect your book from the manifestations of the bubble popping.

We started doing that today, right before the CPI print confirmed what we already knew. Inflation is going up.

Rates are probably going up. And turns out, there are bonds in the stonks, as Bob Prince used to always say.

This is the wedge. You don’t short the bubble. You go long the trend that kills it. If AI is the bubble, rates are the wedge.

Everything with a fat multiple is a long-duration asset in disguise. When the discount rate rises, the NPV of all that future goodness compresses, and the stocks that went up on vibes about 2030 cash flows come back to earth.

The principle is this: inside every bubble, there’s something that needs the bubble to survive. When the bubble even hesitates, the weakest link dies. You’re not betting against the mania. You’re betting that the weakest link can’t survive a pause.

The beauty of the wedge is that your timing can be sloppy. The bubble doesn’t have to pop. It just has to stop accelerating for a quarter, and the levered garbage starts to buckle.

Where are the wedges today? I’ll tell you where I’m looking. Canadian banks at 3x book, stuffed with negative amortizing* mortgage exposure to a housing market that makes 2007 America look restrained. (*aka the borrower doesn’t pay enough to cover interest payments at market rates and the diff gets capitalized into the principal like a PIK loan)

I can't get the options I want on these, but I'm watching. Credit markets broadly, we wrote about this in Watch Credit, private credit feels like a roach motel and representative of looser standards in general. The money checks in and doesn't check out. When the bubble pauses, the marks don't move because nobody has to mark them. Until they do.

B) The Victim

The second way to short a bubble is to find the thing that dies when the bubble pops, the thing that is next to the bubble.

Evergrande was a good example. You didn’t need to short Chinese banks, which is a great way to lose money for a decade. You needed to find the one property developer that was so levered, so dependent on pre-sales from apartments that hadn’t been built yet, that even a modest slowdown in Chinese housing would blow it up. The bubble could keep inflating. Evergrande couldn’t.

What you are looking for is convexity to the downside. You can’t get that by shorting the thing going up in a convex fashion, that’s double short convexity.

The neighbor, well maybe he has options that aren’t trading at 70 vol.

Think airlines before COVID. They weren’t in a bubble, but they were convex to the downside due to the asymmetric risk. Put skew was a bit elevated but not crazy. You could still buy the wings. So we did. In hindsight this seems obvious but essentially the bubble at the time was normalcy.

Then think financials in 07/08. You didn’t need to short housing directly (or more directly, it was extremely difficult and technical to do so, but more power to you if you actually found the CDS on the mortgages). You could just short Bank of America.

The principle: bubbles create correlations that only manifest in the crash. The vol surface doesn’t price those correlations until it’s too late. Your job is to find the thing with cheap vol that’s going to get dragged down by the thing with expensive vol.

Right now, what is that victim? Honestly, I’m not sure.

C) The Confirmation

The third way is the one that requires the most discipline, which is why most people screw it up.

You wait.

I know, waiting is the worst. Sometimes you see a thing go up in a straight line and you can’t help yourself, but again, you don’t want to get run over by the steam engine.

So you wait for confirmation. What does it look like?

Usually some combination of a) weaker fundamentals, b) exhausted vibes/buyers, and c) a trend line break.

Not a dip, but a break. The kind where the thing going up falls through the pretty line and people start posting it on twitter. The kind of break we saw in silver back in January (though don’t look now, it’s back, which we’ll update folks on in a coming ramble). Which means even though I was writing this at 2am last night and didn’t post it (making me look like a johnny come lately), I also get to point to this chart.

Which looks very different,

depending on your timelines.

The essential truth right now is that the only thing deteriorating with respect to AI is the fact that so much of the cash flows are about the future. And the thing is you need to discount all that future goodness with rates today. And if inflation goes up, and policymakers have to tighten (which you could imagine they would if oil goes to $150-200 a barrel) then the NPV of a lot of that stuff goes down. Which we wrote about back in 2021 when the bubble was bonds.

The other thing to look for is correlation. When what was working doesn’t work anymore and all of a sudden it seems sensitive to something it could usually steamroll. We may be seeing that today.

What did I do today?

Well, even before the market puked I was hedging a bit, but not enough. Short another 5% of SPX and 10% of HYG, then a bit more in short term put spreads. Then I stepped away for a minute and came back and eww.

What did I do? I didn’t short semis, the core fundamental demand is still there, the trend is still in place. But I did short more bonds, this time put spreads on treasuries themselves. If the trendline holds and we bounce, I paid a little premium for peace of mind on the wedge. No big deal. If it doesn’t hold, I have dry powder and I have protection, and then — THEN — I press the specific shorts. Oh I also sold 5% of those Canadian banks.

Hedge, wedge, confirm, press.

The part where I (don’t) tell you what to do

Look. I don’t know if we’re in a bubble. This could be the 4th inning of this run (it’s probably not, the price action is already too strong), it could be the 9th (I doubt it, we would need some sort of destruction of underlying demand for tokens and I don’t really see that happening). What I do know, is the feeling of inevitability that AI gives me feels a bit like the same feeling I got when I was in high school in 1999 and put together my first portfolio of internet stocks. And yes, eventually they recovered, there was Amazon in there etc, and the IRR is mid teens if you brought this all the way out to today.

But I also remember the drawdown.

So to bring things home, if you got this far in this particularly rambly ramble, you may be nervous.

If you’re nervous, and if you got this far in this ramble you probably are, the answer is not to short the thing going vertical. The answer is to find the wedge, buy puts on the victim, and wait for confirmation before you press.

In the meantime, don’t fight the tape. Don’t short the thing going parabolic.

Good luck. Till net time.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

Then you sold too early. Like us.

Feels kinda icky.

Now what?

Maybe you want a way to stay long la révolution but you’re a little worried about piling chips on the equity table right now. Maybe you are worried about stocks if oil cracks $150, and you already have the latter trade on but can’t add more. Like us.

Well, maybe we’re both looking at it wrong, since what actually makes Taiwan the linchpin of the next twenty years isn’t just the chips. It’s the currency flows.

Today we go into the weird world of FX. Two points.

One, Taiwan is the most pivotal country in the world for the future of the dollar. $17-22 trillion of new East Asian savings (half of which from outside China btw) is going to land somewhere this decade. Whether the security architecture that funnels it into dollar assets survives is the most important question for the future of the reserve currency.

Long Taiwan and Korean FX, short Indian rupee. Equity markets have confirmed the regime change.

Meanwhile the FX hasn’t moved, at least not nearly in line.

Fifteen-year dislocation between the two markets for both TWD/INR and KRW/INR, never been wider.

Negative carry, positive convexity. Ideally in vol markets. We like it.

The friend’s question

A new friend asked me last week. Campbell, what does it matter if China gets Taiwan? Seems like it eventually happens anyway. Why should America care?

Reasonable question. We get it all the time. The map looks the way it looks.

As we put it in our ‘50-year peace plan for Taiwan,’ imagine a Soviet battle museum where the Statue of Liberty is.

Or as they call it: Kinmen Island.

And the neocons seem to care a lot. The average person notices that neocons getting upset about something is sometimes a tell that the thing isn’t worth getting upset about.

I get it.

The argument today isn’t politics, loyalty, honor, or duty. Even though personally those are the things I think we should stand firmest on. Today is material. Dollars and cents.

Because the future of the dollar runs through Taipei. And the dollar is the key to protecting your way of life. Whether you want to acknowledge it or not.

How the hegemon’s deal works

You don’t get to be the reserve currency for free. Country X sells TVs to the world and gets paid in dollars. If enough exporters keep enough dollars in USD assets, the dollar trades persistently rich. That richness gives you cheap imports and cheap capital, but it hollows out manufacturing and encourages too much leverage. All else equal, this is how reserve currencies die.

But all else isn’t equal. The center of the trade system also charges financial rents, gains sanction power over anyone who uses the currency, and uses the cheap capital to invest in the technological frontier. Today that’s AI. Two hundred years ago it was banking and insurance, run out of coffee houses in London.

The position also lets you cut a deal with your allies. They supply you. They recycle their surpluses into your currency. You backstop their security so they can invest in export industries instead of building militaries.

This worked for sterling. It works for the dollar. It also falls apart fast if you stop maintaining a world-class navy btw.

The AI revolution makes the surplus pile existential

The first time I sat down and tried to model the recycling architecture seriously was 2015. I thought the China-as-reserve-challenger thesis was 50/50.

I’m only convinced now they’re playing for keeps because I’ve watched them, for a decade, refuse to take the path that would resolve their problems domestically. The currency stays suppressed. The banks stay broken. The capital account stays closed. Every six months FinTwit says “this is the year of the great recap.” Every six months it isn’t.

But before we get to why they’re playing offense, deal with the peaceful version of the challenge.

The peaceful version doesn’t work

The optimistic case for China-as-reserve-challenger goes like this. China runs persistent surpluses. The world ends up holding RMB. Gravity does the rest. Reserve status follows surplus. No invasion needed.

This is the version Ray’s 1000-year cycle studies tend to gesture at.

The problem is that reserve currency status isn’t just trade gravity. It’s about whether holders of your surplus can do anything with the claims. They need to buy financial assets they actually want, exit when they need to, recycle into productive investments.

China fails this on every dimension. Capital account closed. Housing and A-shares overvalued or government-traumatized. Bond market dominated by state issuance with managed yields. There is no Chinese equivalent of the Treasury market that a foreign reserve manager can buy and sleep at night.

The peaceful thesis dies on the question “what do I do with my RMB once I have it?”

Ray’s framework misses this because it treats reserve status as a function of trade and military hegemony. It is. But it’s also a function of an asset market deep, liquid, and trustworthy enough to house the world’s savings for decades. Sterling had it. The dollar has it. China doesn’t, and there is no path to acquiring it without opening the capital account, which they can’t do because the banks would blow up.

What Ray also misses: the AI tigers

In the era of AI, the surpluses are no longer being generated by China. They’re being generated by the AI tigers — Taiwan, Korea, Japan — on higher-value products China can’t replicate.

Korean memory prices the past two months:

SSD prices doubled in ten days. NAND up 47% in a month. AI capex flowing through customs in real time.

Which is going to put upward pressure on already healthy current accounts:

Taiwan ran a $181B current account surplus in 2025. Twenty percent of GDP. Nobody runs that ratio. Korea ran $123B. Japan ran $204B. The trio is half a trillion a year and going vertical.

Setser’s Korea+Taiwan customs surplus is $400B trailing 3-month annualized, up from $100B two years ago. Quadrupled.

And here’s where it starts to get…interesting.

With the surge in export quantities and prices from the AI revolution, the AI Tigers are in shouting distance of outproducing the Dragon on surplus generation. On higher-margin products. With open capital accounts. With deep, liquid asset markets. With existing security relationships to the United States.

Taiwan does the heaviest lifting — 23 million people generating $1.2T to $4T of cumulative surplus. Korea is the second leg, $740B to $2.7T. Japan stopped being an export story a decade ago. Japan’s current account is now primary income on the $3.5T NIIP — interest, dividends, repatriated earnings on assets built up since the 1980s. That compounds whether AI booms or busts. Japan is the floor. ~$2.3T cumulative in any scenario.

This is the savings pile that is going to look for a home. Right now it lands in dollar assets by default, because the AI tigers sit inside the US security perimeter and there is no comparable alternative.

That default is what Beijing is trying to break.

The recycling channel quietly died

Here's the thing most Americans don't appreciate. The pipes changed.

Not because the surpluses are gone — they’re bigger than ever. What changed is who holds the dollar.

The Bank of Korea added basically zero reserves on a $123B surplus in 2025. Taiwan’s CBC added $20B against $181B. The central banks aren’t intervening because the private bid is doing the work. Korean lifers, Taiwanese insurers, Japanese pensions, and retail on the Toss app are absorbing the dollars to fund foreign-asset purchases. Setser nails this — Korea’s record surplus was offset by private outflows, not reserve accumulation.

Same surplus. Different holder. Different reaction function.

A central bank with $1T of Treasuries has a 30-year horizon and no career risk. A Korean lifer with $1T of Treasuries hedges FX quarter by quarter. If the hedge ratio rises 5%, that’s $50B of mechanical dollar selling.

That’s why TWD moved 5% intraday in May 2025 when the lifer hedge book wobbled. Five percent. On a major Asian currency. In one day. The surplus didn’t change. The hedger did.

The plumbing got fragile and we didn’t notice.

What China is actually doing

China's de-dollarization story is a press release.

Setser’s 2026 work shows what’s actually happening. SAFE’s official dollar share fell from 79% in 2005 to 55% by 2019, which made all the headlines. What the headlines missed is that SAFE stopped growing reserves at all. The new accumulation moved off SAFE’s books. State commercial banks hold roughly $1T in dollar assets abroad. Policy banks hold close to another $1T. CIC holds $450B, mostly dollars. Total Chinese state-sector dollar holdings now likely exceed $4T. More off the books than on.

China is the largest dollar accumulator in the world while running the loudest de-dollarization campaign. This isn’t a contradiction. It’s a holding pattern.

Beijing doesn’t have a better option yet. Letting RMB rip kills the export model. Opening the capital account exposes the banks. The banks themselves are the constraint. Rhodium’s recent work documents that 2025 write-offs were trivial compared to the scale of declining bank profits, suggesting the real recapitalization hasn’t even started. The hole is somewhere between $5T and $10T depending on whose estimate you trust. The banking system has about $5T of equity.

The solvency framing misses the point. This is a physics problem.

What holds it together is three forces. Capital controls. Consistent trade surpluses. And propaganda about the rise of China, aimed as much at the domestic audience as the Global South.

Tying the pieces together

Now ask yourself. Why does China keep the currency suppressed when raising it would boost consumption? Why not fix the banks? Why keep the capital account closed? Why the military build-up that exceeds ours on PPP terms? Why the proxy support for Russia, Iran, and North Korea?

One framework ties these pieces together better than the alternatives. Beijing is trying to convert manufacturing dominance into regional balance-sheet dominance.

Get Taiwan. Force the regional allies to price Beijing as the local risk-free authority. Use the captured savings to reallocate in CNY assets, and, in the process recapitalize the broken banks. You finally found a buyer to handle the offsetting inflows to domestic capital flight.

Then open the capital account. Then let the currency rise. Then let household consumption boom on the back of cheaper regional imports.

I can’t read Xi’s mind. But the framework explains the observable behavior better than the alternatives. And the timing works too neatly to ignore. They need roughly $8-10T to fix the banks. The AI Tigers surplus pile is tracking $4-11T over the decade depending on how wacky the AI acceleration goes. Which is about the same order of magnitude as a) the potential private capital flight, and b) the hole in the banking system.

Putting real stakes on Taiwan. It’s not about territory and ego. It’s about a regional balance sheet that can make all their problems go away and propel them to global reserve currency even when their banks are hosed.

Capital alone doesn’t get them there. The buildout needs physical capacity — chips, packaging, power. The very same constraints that gate the AI buildout gate the surplus accumulation in the take off scenarios:

The unwind doesn’t have to be loud

Most pushback on this thesis pushes back on the wrong picture. People imagine a fire sale. Trillion-dollar Treasury liquidations on a Tuesday morning. CNN chyron. Vol spike.

That’s not how it goes.

The first move is quieter. A Korean pension fund shortens duration. A Japanese insurer raises its hedge ratio. A Taiwanese lifer buys less agency paper this quarter and more gold. A reserve manager who never wanted RMB decides it needs a token allocation because Beijing has become the local weather system.

None of these are headlines. All of them happen in the marginal allocation meeting on Wednesday afternoon.

The stock doesn’t have to move. The flow only has to stop compounding in the old direction.

The choice

So when my friend asks why America should care if China gets Taiwan, the answer isn’t about Taiwan.

Taiwan is the credibility test for the balance-sheet regime that finances America.

Pay to defend it. Rebuild the manufacturing base. Generate AI’s energy domestically. Or take the cheaper path and let Taiwan go — in which case Beijing gets the keys to the regional balance sheet that has financed our fiscal model for forty years.

There is no clean version where Taiwan changes hands and the dollar keeps the same automatic claim on East Asian savings.

The dollar is the whole thing.

Come for it and you come for everything built on top of it.

That’s why America cares. Or at least why it should.

But you and I don’t get to control whether America cares. We control what we own. Let’s talk about how to position for the regime change already underway in the data, whether or not the policy response materializes.

The Trade

Long the currency for the folks making chips…

short the currency selling human labor at a discount to the west.

The thesis is simple. Korea and Taiwan are running record current account surpluses on AI-cycle exports. India is running a structural deficit on energy imports and human services exports that AI is beginning to displace. The equity has priced this. The FX hasn't.

The chart that does the work

Korean stocks vs Indian stocks have rarely outperformed by 40% a year. EWY/INDA is now at +240%.

Meanwhile KRW/INR has only gone up 7%.

Same in Taiwan. EWT/INDA at +110%. TWD/INR at +7.6%.

Stocks have confirmed the regime change. FX hasn’t. Historically FX follows the equity-relative signal more often than not. The burden of proof is on persistence.

A word on the India services leg, because it’s doing real work in this thesis. India’s IT services sector — Infosys, TCS, Wipro, the whole BPO stack — built its moat on labor arbitrage. Skilled English-speaking engineers at a fraction of Western rates. That moat held for twenty years because the alternative was hiring locally at 5x the cost. The alternative now is an API call. Coding assistants are already replacing the entry-level ticket work that feeds the pyramid. Accenture’s last two earnings calls flagged AI-driven compression in outsourcing contract values. This isn’t a 2030 problem. The pricing pressure is showing up in current contract renewals. India’s services surplus was $180B in FY2025 — still large, still growing, but the growth rate is decelerating into the teeth of an energy import bill that moves the other direction when crude rips.

Energy is the accelerant

Brent at $75. Hormuz still volatile. Right tail is $150.

India imports nearly 5 million barrels of oil a day, around $140B at 2024 prices. A doubling of crude pushes that toward $250B and widens the CAD by several percent of GDP. Japan, Korea, and Taiwan also import a lot of energy, but they’re coming into this with massive current account surpluses made of chips. India’s surplus, such as it is, is made of human services that AI is actively displacing.

Korea’s oil bill is higher than India’s as a share of GDP. Taiwan’s is similar. But the starting current account is what determines who breaks and who absorbs. India is already in deficit. A doubled oil bill takes the CAD toward 5% of GDP. 1991 territory.

This is the macro version of the long APIs / short slides trade. Through the currencies, not the equities — the equities have already priced it.

The carry problem

Looking at short rates: India 5.25%. Korea 2.50%. Taiwan 2.00%. Short INR against long TWD/KRW costs roughly 3% per year.

Going into the Iran conflict, RBI was about to start cutting. December inflation print was 1.33% — way below their tolerance band. The disinflation story was real and the path to cuts was opening. That was the dovish setup that would have compressed the rate differential and made this trade cheaper to carry.

Energy ruins that. If crude rips, India’s inflation re-accelerates from the import bill alone. RBI doesn’t get to cut. Probably has to hold or hike defensively into the rupee weakness. The rate differential widens, not narrows.

Meanwhile Korea and Taiwan don’t want to tighten. The inflows are already hot — tightening into the surplus would attract more capital, strengthen the currency further, and hurt the exporters their entire policy framework is built around protecting. Both CBs would prefer to ease if anything, but the AI cycle is doing the work for them.

So the rate differential probably widens in the energy scenario, not narrows. Carry on this trade gets worse before it gets better.

The spot has to do the work. A 10% INR move dominates the 3% carry. A 15% move pays for two years of waiting. The trade is sized to the spot catalyst, not to fund itself on rate convergence.

How to play it

Buy the upside, sell the downside, pay ~nothing.

The clean expression is a risk reversal — buy a call on TWD/INR or KRW/INR, sell a put at a roughly symmetric strike (vs the fwd), structure for near-zero premium and hedge the delta. The forward curve already prices meaningful INR weakness because of the rate differential. KRW/INR 3-month forward is 1.3% above spot. TWD/INR 1-year forward is 3.7% above spot, which seems to imply more currency weakness than the interest rate differential would imply (though it’s late here so check my math).

The trade is betting that spot moves further than the forward already implies. If you’re right, the call pays off. If the spot drifts sideways near the forward, premium is roughly neutral. The short put only hurts if the spot moves materially against the structural thesis.

Bloomberg shows 3-month KRW/INR vol around 10% on both sides of the forward, premium roughly neutral on a symmetric structure. 3-month TWD/INR around 8%. Symmetric surface means the vol market hasn’t priced the asymmetric vol you might expect from the difference in their current situations.

The structure works because: premium-neutral, defined downside via the put strike (already below the forward, so you only get hit on a real reversal), unlimited upside on the call. Better expression than a vanilla call because you’re not paying carry to wait.

For those who want to dig in, the full mechanics on vol surface, delta, and strike selection are at the end of the piece.

Implementation honesty

I would like to put this trade on, but I can't. Both legs require NDF access through a prime broker, whether you're playing it through options or underlying futures and forwards. There is no such thing as a TWDINR or KRWINR future. Most retail accounts can't run these crosses. This trade is the clean institutional expression of the framework — described here so readers know what the trade looks like, not because the median reader can put it on.

For retail: USD/INR options on CME’s Micro INR contract get you the directional view on the deficit side. You don’t get the full pair expression, but you get the leg with the cleanest catalyst — RBI losing control of inflation as energy rips.

As a note, don’t substitute EWY or EWT for the FX legs. The equities have already moved. The whole thesis is that equity confirmed the regime change and FX is catching up. Closing the trade through the side that already ran defeats the purpose.

What can go wrong

The structural left tail: Taiwan blockade/invasion. TWD collapses 20-30% on capital flight, the put gets assigned. Mitigation: gold and short-credit book pays in this scenario, partial book-level hedge.

China defensive devaluation. CNH weakens, drags KRW and TWD with it. Given the size of their surpluses, the fact they are reportedly selling their oil reserves into the market, and the recent positive PPI print, we don’t see this as likely.

Trend / mania exhaustion. Both are already rallying on the physical flows and the market is pricing in more through the forwards. If the move has mostly played out and the next three months are sideways or we see a sharp correction in stocks, the risk reversal loses money if spot underperforms the forwards.

Watch the FX

Not the equity. Not the AI capex headlines.

The FX.

TWD/INR breaking 3.20 confirms continuation. KRW breaking 1,250 confirms the second leg. Until then, you’re patient. The forward is doing the carry work. The risk reversal is doing the structural work. You’re waiting for the spot to do its work.

This is the trade for the next twelve months if you missed the boat.

Till next time.

Appendix

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

We’ve talked a lot about Trump’s game theory of “mixed strategy,” the idea that playing as the drunk guy at the poker table opens up trade space. His repeated “sign the deal or else” rhetoric. Classic stuff.

What’s changed in the last week or so, outside of yet another Axios put, is he may have finally met someone drunker at the poker table than he is. The only counter to crazy is more crazy.

I have no inside baseball, but I’m trying to map onto the pattern where the admin appears to be negotiating with someone on the other side, and actually making material progress on the core war goals: a non-nuclear Iran and freedom of navigation.

Iran’s first move was to close Hormuz and escalate to the dollar. The admin then wisely countered with blockade and threats to Iranian oil production and revenue. Which looks to have successfully brought China to the table, if the tea leaves of the recent visit are to be read.

But here’s the problem. When you negotiate with Iran, you are usually negotiating with the “good cop” in parliament, not the “bad cop” in the mountains, in the tunnels, or in the fast boats. This gives the regime a source of ambiguity and uncertainty that effectively counters Trump’s chaos. Because Trump does ultimately want a deal. For the economy, for his allies, for the midterms, for the world.

The hardliners in the IRGC don’t want a deal. They want chaos. They want invasion, and to tie the US into a decade-long boots-on-the-ground quagmire where their home turf becomes the dying ground of yet another rimland empire.

This is, by and large, what China wanted too. Until the US wisely put their oil at risk.

Which brings us to today, where I wouldn’t be surprised if, Axios notwithstanding, no one really knows what’s going to happen. Love tap or no.

The Complacency Surface

So let’s step back from geopolitical speculation and look at what markets are actually pricing.

The short version: everything is priced for perfection, simultaneously, across every asset class.

Oil is capped. In spite of will-they-won’t-they escalate, crude has sold off a lot the past week and has by and large stayed under the March lows. Crude under $100 is best case scenario in lieu of an actual peace deal. Who knows how much is from the US SPR, but with inventories in free fall and some floor the strait closure will likely force, oil prices move up if we remain closed by June. With the SPR running out of spare juice by around September, you can’t imagine oil staying below $110 when the spigot runs out.

Stocks are at the highs. Vol is at the lows.

I am toe in the water on some Monday put spreads again. My theta bill this year is going to be ugly, but at 11 vol I couldn’t help myself before tonight’s close. 4:1 on the market falling 1% by Monday, I’ll put 15bps into that. Most of the vol crush has been driven by record low realized correlation which can easily revert. Now that we are almost past earnings season everything can trade on vibes again.

Credit spreads are back at the lows. Our HYG puts are almost expired.

Gold (and silver) calls remain our biggest peace position. Gold has traded a lot like short oil lately, which makes sense if you think of gold as the “peace dividend” asset, the thing that rallies when the conflict premium comes out of energy.

Helped a bit by a lower dollar.

Currency vol…at the lows.

Bond vol…at the lows.

That’s the snapshot. Oil capped, stocks at highs, vol crushed, credit tight, gold bid. The market is pricing peace. The question is whether the market is right.

The Mid-Cycle Acceleration

Here’s what’s interesting. The market isn’t just pricing peace. It’s pricing a mid-cycle acceleration that the data is actually confirming.

Manufacturing PMI came in at 54.5 versus 54.0 expected, above already healthy expectations.

Labor market data came in healthy. Jobs above, initial claims below expectations. ADP employment change hit 109k versus 62k prior. Construction spending above expectations. Inflation expectations above. Consumer credit is expanding. Factory orders and cap goods came in strong. Inflationary greenshoots confirmed in ISM prices paid.

This is not a late-cycle economy. This looks like the early innings of a re-acceleration, which is great for earnings and terrible for anyone hoping the Fed cuts rates.

Which creates a tension that the market has not resolved. If the economy is actually accelerating, that’s bullish stocks but bearish bonds. And if bond yields go up, the discount rate on every equity goes up, which compresses multiples. The mid-cycle acceleration is simultaneously the best argument for being long equities and the ticking clock on the equity bull market.

Global bond yields are back to pre-GFC levels and still rising. Not just the US. Everyone.

Japan just woke up from a 30-year nap. Korea is at levels not seen since 2008. The UK and Germany are marching higher in lockstep. This is a global tightening regime, and there’s nowhere to hide in duration.

If the acceleration data keeps printing like this and inflation expectations start to unmoor, remember: there are bonds in the stocks.

The yield curve tells you where the trade lives.

Curves globally have normalized from the deep 2022-2023 inversion but are now basically flat. The US curve in particular has no space left. Markets have already priced 100bps of cuts by end-2026, and then nothing. The flatness between December 2026 and December 2027 is unprecedented. The bond market is saying: the Fed will ease, regret it, and get stuck.

I’m rotating into a belly steepener this week, betting that rates in 2026 will be lower than 2027. The logic: once the easing round concludes (around 18 months out), you would normally expect steepness between the 15-month and 27-month tenors. But because the market has already had to discount so much easing from where we are today, there’s no room in the curve before it has to start flattening out again on term premium. That gap is the trade.

Ags Update

Our ags trade has started to work a bit, driven largely by wheat and mini rallies in corn and sugar. We exited our June sugar after a combination of peace news and after doing some more research on the local ethanol market in Brazil, where the gasoline price caps are putting a lid on the amount of sugar diverted into the market (and where we expect a lower ethanol share along with seasonal trends). Our bullishness is more about a system vulnerable to breaking and less something that has a catalyst over the next two weeks. We still have our September calls with 14% delta and a first strike at 16.5c but down from the aggressively bullish stance of last week.

If ags were slow on the upside in April, they appear to be catching a bid. Partially because the fertilizer problem and linkages to oil are getting priced in, partially because of bad weather leading to higher wheat prices. We remain constructive and the complex remains our biggest position.

Long APIs, Short Slides: The Update

Back in February we published “Long APIs, Short Slides” and by and large this trade is working out.

The thesis in one sentence: the shorts get paid per human deployed, the longs get paid per API call, and AI replaces humans while multiplying API calls. Three months in, the earnings data is confirming the bifurcation.

AI infra companies beat earnings estimates by an average of 210% and their stocks are up 15.4% since publication. Software beat by 22% and is up 3.4%. Services beat by 7% and are down 3.5%.

The most interesting thing in the data isn’t just that AI infra beat harder. It’s that the market rewarded those beats with disproportionate price appreciation. The AI infra EPS Z-slope is 1.79 versus 0.01 for software. Beat earnings and be on-theme, your stock goes whoosh. Beat earnings and be off-theme, nobody cares.

Meanwhile the short side is doing exactly what we said it would. Services companies beat on earnings by a modest 7% but their stocks went down 3.5%. Revenue came in slightly below expectations. The market is telling you something: even when these companies report decent numbers, nobody wants to own them. The AI displacement narrative has become a valuation ceiling. Decent quarter? Great, you’re still a body shop selling human hours into a market that just discovered it can buy machine hours instead. The slow bleed we described in February is playing out one earnings report at a time.

So how does the pair trade look in aggregate?

Keep in mind our API basket is about 2x more volatile than our Slides basket.

But even in vol-adjusted terms the outperformance has been real.

A volatility-adjusted long-short basket of these names is now up 21.7% since publishing. Largely as a result of greater than expected earnings.

A couple of housekeeping notes on the basket. Confluent (CFLT) got acquired by IBM, so that name comes out. We’re watching Accenture as the canary for whether the body shops can actually pivot. They’ve announced $3 billion in AI bookings, which sounds impressive until you realize “selling AI consulting” is a completely different business than “selling bodies at $30/hour.” If ACN’s AI bookings translate to margin expansion, the short thesis weakens. If they don’t (and so far, they haven’t), it confirms the structural problem: you can’t repurpose a 350,000-person labor arbitrage machine into an AI consultancy. The Indian IT firms are stuck.

The Bottom Line

The market is pricing peace. The economy is pricing acceleration. Vol is pricing perfection across every surface simultaneously.

If any one of those assumptions breaks, you want to own optionality. If the economy keeps accelerating, inflation unanchors and bonds in the stocks become a problem. If peace breaks down, oil spikes and the SPR runs dry by September. If the AI theme stumbles, the concentration in Mag7 becomes a correlation bomb.

And if none of them break? The APIs basket keeps compounding. The ags position catches the fertilizer-oil linkage. The belly steepener earns carry. The gold and silver calls sit there as the peace dividend.

Either way, the portfolio works. Which is the whole point, just mind the carry and try not to chase the mania.

Disclaimers

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure.

Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

What we might call in trading a position that is “positive carry, positive convexity,” meaning one where the status quo works for him, while exposing him to asymmetric gains. Particularly after having used the ceasefire as justification for not triggering the 60-day war powers review that would tangle his plans in congressional meddling.

Whether or not this status quo holds is anyone’s guess. To us it appears equally likely at this point that we will see some sort of escalation as movement towards a sustainable peace. Meanwhile, every day the strait is closed, global energy and food markets get tighter and tighter. Every day fertilizer remains dear is a day that less of it goes onto crops, which in turn reduces yield further, which in turn raises medium-term expectations for prices.

Next time, we’ll likely go deeper on the inflationary pressure we see coming on the horizon, pressure which led us to sell a bit of longer-term duration towards the end of last week (a position we may add to if bonds rally this week). We’ll also start looking at the impacts on emerging market debt and currencies we see particularly exposed to this witches’ brew of acceleration and inflation (we’re looking at you, India).

Today we’re going to dive into one leg of our cascade trade, the one which is playing out as we speak: the impact of high energy prices on the market for sugar, through a single, critical linkage. Brazil.

As opposed to prior pieces, this one is going to be relatively straightforward, and hopefully not too long.

The story is simple. Brazil dominates world production of sugar, and it looks like substitution demand for ethanol on sugarcane (the raw material that acts as feedstock for traded sugar) will entirely consume the “production surpluses” that were consensus a matter of months ago. When combined with the smaller impact of Indian ethanol substitution, and the potential shock from an El Niño with below-average rainfall, we remain bullish on sugar.

A Very Simple Market

Sugar is actually a very simple market.

Brazil produces about 40 million tonnes a year and exports roughly 35 million of it. That is almost half of all sugar traded globally. The next largest exporter, Thailand, ships about 7 million tonnes. Everyone else is a rounding error or a net importer. When you trade sugar, you are trading Brazilian supply decisions.

But here is the thing that makes sugar different from most other agricultural commodities. The raw material that sugar is made from, sugarcane, can also be turned into ethanol. Brazilian mills make this choice every two weeks based on which product pays more. When ethanol pays better than sugar, mills divert cane away from sugar production and into fuel. When sugar pays better, they swing the other way. (Corn has a similar dynamic in the US, where ethanol competes with feed and food uses, but the mechanism is less direct because American ethanol production is structurally locked in by the Renewable Fuel Standard rather than driven by biweekly mill-level economics.)

This means sugar is not just a food commodity. Sugar is an energy commodity. Its price is linked to crude oil through a chain that runs: crude oil → gasoline price → ethanol competitiveness → mill allocation → sugar supply. When oil rises, ethanol becomes more valuable, mills make more ethanol, less sugar reaches the world market, and the price of sugar goes up.

Over the past twenty years, every major sugar rally has been driven by the same thing: a demand-supply gap caused either by weather disruptions (India, Thailand) or by Brazilian mills diverting cane to ethanol (2008-09, 2015-16, 2019-20). The correlation between year-over-year changes in crude oil and sugar prices over sixty years of data is striking.

The Consensus Was a Surplus

Going into this year, the consensus view on sugar was comfortable. Czarnikow, a London-based sugar trading and advisory firm whose forecasts the market trades off of, projected a surplus of about 3.4 million tonnes in February. By April they had revised that down to 1.1 million tonnes, but the story was still the same: more sugar than the world needed, prices drifting sideways or lower, nothing to see here.

Sugar #11, the global benchmark traded in New York, closed Friday at 14.95 cents per pound. It has barely moved.

We think the consensus is wrong. We think the surplus is already gone, and the market hasn’t counted yet.

The Mills Just Voted

Brazil’s sugarcane harvest started April 1. The cane gets crushed over about eight months, from April through November. Every two weeks, UNICA (Brazil’s sugarcane industry association) publishes how much of the crop went to sugar versus ethanol. That biweekly number is the single most important data point in the sugar market.

The first report just landed. In the first half of April, mills allocated 75% of their cane to ethanol and just 25% to sugar.

Same period last year was 56/44. The fifteen-year average opening split is roughly 53/47. The highest-ever opening prints, in the 2018 and 2019 super-ethanol seasons, were around 60%.

This opened fifteen points above those.

One more thing worth noting: on Friday, sugar closed up while crude oil closed down. That divergence is a tell. The market is starting to figure out that the supply squeeze is real, independent of the daily moves in Brent.

Why: The Price Stack

Why would mills choose ethanol so aggressively? Because sugar is not competitive with ethanol right now, and it is not close. To see why, you need to follow the full chain of fuel prices in Brazil, from what gasoline should cost all the way down to what mills actually get paid.

Here is the price stack, converted to dollars per gallon:

Every row is above the row below it. Mills are getting $2.34 per gallon for ethanol. They only need $1.48 to match sugar. That is a 58% premium for choosing ethanol. No rational mill makes sugar at these economics.

And this is with gasoline artificially suppressed. Petrobras, Brazil’s state oil company, has frozen refinery prices for nine months. The gap to international parity is roughly $1.70 per gallon. That freeze costs Petrobras about $20 billion a year in foregone earnings, or 45% of last year’s profits. It is mathematically unsustainable, and the pressure on the board to adjust prices is mounting. But Lula is unlikely to allow any change before October’s municipal elections. The 2018 truckers’ strike that brought down the last Petrobras CEO who tried to pass through international prices is the precedent the entire political class is structured around. So for now, the freeze holds, and Lula needs other levers.

His cleanest lever is the ethanol mandate. He cannot make gasoline cheaper, but he can stretch the barrel by blending more ethanol into it. On May 1, Brazil’s energy council (CNPE) voted to raise the mandatory ethanol blend in gasoline from 30% to 32%, known as E32. They also raised the biodiesel blend from 14% to 16%, which pulls soy oil out of the food complex. Energy Minister Silveira framed it explicitly: the goal is to “zerar importações de gasolina,” eliminate gasoline imports entirely.

The mills already knew. They started crushing two weeks before the vote like it had already happened. It had. They just got the formal piece of paper on May 1.

Here is the critical insight: the Petrobras freeze is not what makes this trade work. The Petrobras freeze is what is holding the trade back. Even at suppressed domestic gasoline, ethanol pays 58% more than sugar. If the freeze breaks after October, gasoline jumps to international parity, ethanol follows, and the crush pushes even harder. Sugar would need to rally to 23 cents per pound before mills become indifferent at current frozen prices. At international parity, it would need to hit 30 cents.

Sugar at 14.95 cents is not even in the conversation. The mills have no reason to make sugar at any gasoline price above $3.50 per gallon, and gasoline hasn’t been that cheap in Brazil since 2020.

What This Means for the Global Balance

Each percentage point of ethanol share above the consensus 52% removes about 850,000 tonnes of sugar from the world market. The opening fortnight print of 75% will fade as the season progresses and the sugar content of the cane rises with maturity, but even with the typical seasonal normalization, we estimate the full-year crush will come in around 58-62%. At 60%, that is roughly 7 million tonnes less sugar than what the consensus assumed.

The chart above shows the global demand-supply gap at different levels of Brazilian ethanol crush. The three lines represent different assumptions about the rest of the world: a scenario where India and Thailand have unexpectedly good weather, the current market consensus, and a scenario where the Indian monsoon disappoints.

The consensus heading into the year was a small surplus. At 60% crush, that flips to a deficit of 4 to 6 million tonnes depending on what happens elsewhere. At 55%, which is below the fifteen-year median and well below what the parity gap implies, the surplus is already gone.

The crush season runs April through November. Brent is $114 at spot and roughly $95 on the forward strip through the end of that window. At spot, our model implies 60-62% crush. At strip, 57-59%. Either way, well above the consensus 52%. The market is pricing a crude collapse it hasn’t bothered to check.

What Could Make This Bigger

Higher ethanol mandates. E32 just passed. The 180-day review window puts the next decision in November, right after the election. If Brent stays high and Lula wins, E35 is the obvious next step. E32 alone removes 1.5 to 2 million tonnes of sugar from the global balance. E35 removes another 1 to 2 million on top. The biodiesel raise to B16, passed the same day, tells you the political commitment is real across every biofuel lever Lula can reach.

Higher domestic energy prices. If Petrobras eventually unfreezes, domestic gasoline jumps, ethanol follows, and the crush intensifies. Paradoxically, higher energy prices in Brazil would also lift the sugar-equivalent floor: the price at which mills rotate back to sugar rises with gasoline. Sugar rallies, but not enough to restore supply. The equilibrium price moves up, it does not revert.

Indian monsoon. India is the second-largest producer and the swing exporter. India’s weather bureau (IMD) issued a below-normal monsoon forecast on April 13: 92% of the long-period average (the benchmark for total June-September rainfall, averaged from 1971-2020), with a 35% chance of deficient rainfall. That is the first below-normal April call since 2015, which preceded the El Niño that helped drive sugar from 11 to 23 cents in nine months.

India has two ways to tighten global sugar, and they work through completely different channels. The first is weather. Maharashtra and Karnataka together produce about 50% of Indian sugar, and both are heavily rain-fed. A bad monsoon physically reduces production by 3 to 5 million tonnes over the following year and triggers the government to ban exports. A monsoon miss flips India from putting 2-3 million tonnes on the world market to pulling 3 million tonnes off it. That is a 5-6 million tonne swing in global trade, and no other country can move the balance that fast.

The second channel is ethanol policy, and here the picture is more complicated. India has built roughly 9 billion liters of sugarcane-based ethanol capacity under its E20 program (a 20% mandatory ethanol blend in gasoline, similar in concept to Brazil’s mandate). The original government roadmap projected sugarcane contributing 55% of that ethanol. But under the current supply year, sugarcane has been allocated only 28% while grain-based ethanol got the rest. The government effectively chose grain over cane, which means more sugar stays in the domestic pool, not less. The ethanol channel is going the wrong direction for the sugar bull thesis right now. But that policy was set before Brent hit $114. India’s fuel import bill is the political pressure point, and the sugar industry now has a stronger argument at the Prime Minister’s office: sugarcane ethanol capacity exists, it’s sitting idle, and every liter reduces the crude import bill. An upgrade to E22 (a 22% blend) would divert another 1.5 to 2 million tonnes of sugar into fuel. That is upside we are not paying for, but it could arrive anytime between now and October.

I am not building the trade on India. But the tail is meaningful. The end-of-May IMD second-stage forecast, which adds the regional breakdown for the sugar belt, is the next checkpoint.

Speculative positioning. Agricultural commodity speculators have been reportedly turning long in grains, but our read is that sugar specs are still net short. A positioning squeeze on top of a fundamental deficit is the accelerant, not the cause, but it matters for timing.

What Could Make The Demand/Supply Gap Smaller

Much better weather everywhere. A normal monsoon in India, a full recovery in Thailand, no disruptions anywhere. Even in that scenario, our model shows a 58% crush wipes out the surplus. You need crush below 55% for the surplus to survive, and the parity gap says that is unlikely.

Peace. A rapid resolution at Hormuz drops Brent back to $70-75, compresses the parity gap, and the crush fades toward 53-55%. Sugar stays around 14-16 cents. This is the clean invalidation of the thesis. It requires the same thing that invalidates the broader Cascade.

The Historical Pattern

Historically, every sugar deficit of 5 million tonnes or greater has produced at least a 50% price rally. The one exception was 2018, which ran the largest deficit on record but barely moved price. The reason: it followed a record surplus and stocks-to-use was at multi-decade highs. The system had a cushion thick enough to absorb the shortfall without rationing through price.

Today, stocks-to-use sits below the historical average. Same deficit, thinner cushion.

At our base case of a 7 million tonne deficit, the historical analog points to sugar in the low-to-mid 20s. At the tail case, where Brazil crushes 60%+ and the Indian monsoon misses, the analog is 2009, when sugar went from 11 cents to 30 cents in eighteen months. That was the largest single-year stock draw on the 65-year chart and it produced the biggest sugar rally of the modern era.

The Trade

I am long sugar optionality in the 15-18 cent range. The downside is bounded by ethanol parity: sugar cannot fall far before mills mechanically cut ethanol production and restore sugar supply. The upside is open, with multiple catalysts stacked through the summer.

Forward call volatility on the March 2027 18-cent strike sits where it sat when Brent was $80. The market has not repriced the optionality for ~$110 crude.

The catalysts, in order:

• Late May: IMD second-stage monsoon forecast with regional breakdown for the sugar belt

• June through September: UNICA biweekly crush reports, each a real-time read on the mix

• June through August: Brazilian diesel stockout window, which forces further ethanol substitution

• October: ISMA first Indian crop estimate and Brazilian municipal elections

• November: CNPE E35 extension decision and Indian export policy for the new season

The crush season runs through November. Each of those events is independently bullish sugar. The convexity is in the conditional correlation: they are all driven by the same upstream variable, Hormuz keeping Brent elevated.

The mills have voted. The government confirmed the ballot on May 1. The market has not yet counted.

Disclaimers

This analysis was built over several days of active markets. Prices, curves, vol surfaces, and positioning data reflect the time they were pulled, not the time you’re reading this. Some of these levels will have moved by the time you see them. Do your own work. Check the chain before you trade. The price on your screen when you lift the offer may look nothing like the price on mine when I wrote this up.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

The Ramble does take positions in various assets we write about, and views are subject to change without notice.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure. Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION.

In my defense, we communicated this to paid subs in our portfolio review Monday. We are now long 10% crude through a combination of upside broken butterflys (we buy calls near the money, sell a couple out of the money, and then buy the deep tail back in case things get pretty wacky).

We did this for two reasons.

First, our UK Rates (SONIA) and gold positions had picked up way too much short oil beta. The Cascade trade of long Agricultural futures (Wheat, Corn, Sugar, and now a bit of Live Cattle) weren’t going up enough on days oil went up, and we were getting whacked on our long rates and gold positions. We needed to close the hole in the portfolio.

Second, the Venn diagram of potential agreement spaces between the US and Iran looked to be shrinking every day. The space an economist might call the “gains to trade area” was collapsing. Or as I put it in a cheeky tweet:

Now, for the record: I have no special info on predicting whether there will or won’t be conflict. I read the tapes like anyone else. But over the last couple of weeks, a few things became clear.

Iran had no intention of giving up its enrichment program. The government in Tehran was split between two blocs with different incentives: moderates looking to make a deal out of Parliament, and ideologues based out of the IRGC, many of whom just watched their fathers get killed by the US and Israel. And underneath all of it, Iran harbored deep ambitions for controlling Hormuz.

On the other side, the US pulled off something clever. By shifting the locus of conflict from “will the US invade, that’s too much blood and treasure” to “we’ll sit outside YOUR blockade and impose our own,” the US shifted the dynamic to one that is locally stable, even if it’s globally another notch of escalation with respect to the real long-term adversary, China.

In the meantime, US stocks ripped on the back of three things: the AI acceleration (which we are still long, though not as much as we were going in, so boo), American Energy Dominance (lower global energy production will inevitably lead to a boom in the folks who produce, pipeline, and export our abundant LNG to a world desperate for new supplies), and the fact that markets have not yet priced in substantial and sustained inflationary impacts from the conflict. Partly from the lid on oil, partly from expectations of a quick resolution.

Which adds cards to Trump’s hand and takes them away from Iran.

The Old Order Is Fracturing

Which brings us to the earth-shattering news: the UAE is leaving OPEC.

In retrospect, predictable.

For the past 5-10 years we’ve been hearing nonstop that the “old world order was dying” and “we need to prepare for a new global order.” The implication being, of course, that China would run that new era at the expense of Pax Americana. Ironically, 26 years into the 21st century, things don’t really seem to be playing out that way.

OPEC was born out of pan-Arab, pan-Islamic alliance against the West. The current war in the Persian Gulf has seen Iran explicitly targeting energy and civilian infrastructure belonging to its Arab Muslim neighbors. Not “we took out a terrorist in an apartment building” targeting. “We’re threatening to destroy your way of life by going after the desalination plants” targeting.

If you were the UAE, how would you feel about being in the same club as those guys?

This is what decoupling looks like. And decoupling isn’t just the US and China. It’s decoupling from the old world order entirely: freedom of navigation, access to markets, low barriers to trade, grow the pie rather than fight over it. The UAE looked at the organization, looked at who was firing missiles at them, and walked.

The UAE has been sitting on massive spare capacity for years, constrained by OPEC quotas that made less and less sense as their relationship with member Iran deteriorated. Their production capacity is set to reach 5 million bpd by 2027, while their OPEC+ quota held them well below that. A 1.4 million bpd gap between what they can pump and what they were allowed to pump. Now the leash is off.

And here’s the kicker: long term, they don’t even need the Strait.

The Habshan-Fujairah pipeline gives the UAE a direct route from its western oil fields to the port of Fujairah on the Gulf of Oman, bypassing Hormuz entirely. While the rest of the Gulf sweats over the Strait, the UAE is building optionality out of the chokepoint. Maybe a treasury dollar swaps line was involved, maybe not, we’ll likely never know, but the petrodollar bears now have some serious egg on their face.

Where Is the Oil Market Today

Let’s get into the numbers.

Production

The headline: Total OPEC production collapsed from 28.5 mb/d in Q3 2025 to 20.8 mb/d in March 2026. That is a nearly 8,000 tb/d decline. The biggest movers: Saudi Arabia down 2,314 in March alone, Kuwait down 1,369, Iran down 182. The war is physically removing barrels from the market.

Total Declaration of Cooperation (OPEC + allies) production fell from 42,969 in Q4 2025 to 35,055 in March. Down almost 8,000 tb/d from peak.

Global supply has fallen off a cliff. From roughly 103 mb/d in mid-2025 to below 98 and heading lower. The dotted line forward projection suggests further declines.

The Gulf producer stack tells the story in color. Saudi and Iran are the big drawdowns. UAE volumes also declining under the old OPEC framework, though that’s about to change now that they’ve walked. The March OMR trendline shows continued deterioration.

And don’t forget the supply side at home. If anything, oil rigs are down since the start of the crisis. US supply is not coming to the rescue anytime soon.

It doesn’t help that a lot of the profitability from US oil comes from natural gas and prices remain at the lows.

Shale isn’t coming to the rescue. Rig counts are down, drilled-but-uncompleted wells are depleted, and producers are prioritizing returns over growth. The US can increase production at the margin, but the days of shale as the global swing producer plugging every gap are behind us.

Inventories

The physical market is screaming.

Middle East floating storage has exploded. From near zero through most of 2025 to record highs. That’s crude that can’t find a home, bottled up, sitting on tankers because the onshore logistics through the Gulf are disrupted. This is what a chokepoint crisis looks like in the physical market.

Asia ex-China onshore stocks. In 2025, they tracked the seasonal average. In 2026, they’ve fallen off a cliff, dropping from ~600 mb in January to below 540 mb by April. That is a 60+ million barrel drawdown in three months, massively below both the 2025 path and the 2024-25 average. Refineries in Asia are running out of feedstock.

Let’s not forget the prior administration raiding the oil piggy bank.

The SPR question looms. The US has been drawing down strategic reserves to keep a lid on prices. How long can that last? What’s the actual flow rate, and what happens to the market when it stops?

Not to say I told y’all so, but...

Meanwhile, commercial inventories are starting to plummet.

Not to mention global.

And this tightness is just starting to flow through to distillate production.

When Do the Inventories Run Out?

Let’s do the math.

Asia ex-China onshore stocks are at ~540 mb and drawing at roughly 20 mb/month below the seasonal norm. The 2025 seasonal low was around 590 mb in October. They’re already 50 mb below that floor and it’s only April. At this pace they hit ~480 mb by July and ~420 mb by October. That’s 170 mb below seasonal. Refineries don’t wait until the tank is empty. They start cutting runs well before that. Call it Q3 for involuntary demand rationing in Asia.

The SPR sits at 398 mb after the largest weekly draw since October 2022. You can’t drain it to zero. The cavern infrastructure has a floor around 150 mb, below which you lose both flow rate and the ability to respond to an actual emergency. So you’ve got about 248 mb of usable runway. At 7.1 mb/week, that’s 35 weeks on paper. Call it 8 months. But no administration is going to drain the strategic reserve to fumes during a shooting war. Realistically you’ve got 4-5 months of meaningful releases before they blink and let prices clear.

Goldman’s scenario analysis tells the same story from the other direction. Even their benign scenario (Gulf exports normalize by mid-June) still shows global visible inventories hitting record lows before recovering. The baseline scenario hits record lows by mid-2026 and stays there. The severely adverse scenario (normalization by end-July) pushes inventories to depths we haven’t seen in the data series.

Here’s the math: commercial inventories are heading to record lows this summer regardless of scenario. The SPR buffer buys months, not quarters. And Asia hits refinery run cuts by Q3 if the draw continues.

The curve already knows this.

The Curve

The curve tells you everything. The curve is in steep backwardation. Prompt is nearly $110, but the back of the curve (2028-2029) is still trading in the mid-$60s. The market is paying a massive premium for barrels now and assuming things normalize later. In fact, the news that the UAE left OPEC actually led longer-dated oil to sell off this week, right as near-term oil went parabolic.

These are the kind of signs you want to see if you are willing to impose short-term pain on Iran in order to extract concessions. There will be short-term inflation as a result, for sure, but the market is telling you that in spite of the escalation this week, the long-term supply-demand picture for oil just got looser.

The Brent prompt spread tells you the urgency. From sub-$1 in November 2025 to $6.52 as of April 27. That is a move that screams physical tightness. The market isn’t gently adjusting. It’s scrambling for barrels.

The 3-2-1 crack spread, which measures the margin refiners earn turning crude into gasoline and distillate, has nearly doubled from around $28 in December to $52.80. Refiners are printing money because products are even tighter than crude. You’re going to feel this at the pump. And then you’re going to feel it in the CPI print.

European jet fuel at nearly four-year highs. The product markets are confirming what the curve is screaming.

China

The great unknown. China’s crude import data has been choppy and hard to read through the noise of their own economic slowdown. But the physical data from Asian inventories suggests they’re drawing faster than imports can replenish. They are claiming that inventories are constant. Not sure if I believe this as much as I believe the youth unemployment data. More than anyone, they want to put on a brave face right now.

More on this as the data comes in.

Where Will Oil Be Tomorrow

Here’s where it gets interesting. The world is starting to price in a global energy market in deep deficit for months, not weeks.

The Cascade: Oil → Fertilizer → Ags

This is where the Cascade trade reconnects. Natural gas and oil are key feedstocks for fertilizer production. As energy prices rise, fertilizer costs follow. And fertilizer is the single largest variable input cost for grain production globally. Update on this soon. Wheat already moving, corn catching up, and even sugar making some interesting moves today.

We’re seeing the early innings of what I think will be another face-ripping rally in agricultural prices. The Cascade trade (long wheat, corn, sugar, and now live cattle) is the expression of this thesis. Agricultural calendars are immutable clocks: if fertilizer is expensive during planting season, yields decline. That doesn’t show up in prices immediately. It shows up at harvest. The market hasn’t connected these yet.

The Inflation Question

Now let’s talk about what all of this means for inflation.

Pretty much since the early 2000s, OPEC played a role stabilizing the oil market, expanding production when the market was tight and taking oil offline when things got too loose. You can see in the chart how tightly oil prices track OPEC production changes. This time is no different in mechanics, just in magnitude. OPEC production isn’t declining because of voluntary cuts to manage price. It’s declining because the cartel is fracturing and its members are at war.

We sold some nominal bonds yesterday, and need to look at breakeven inflation. Not sure it’s pricing in enough.

Warsh is coming in at a very difficult time, and even ex-oil, the economy looks to be heating up.

Who Benefits

So who benefits from all of this? Look at the table.

The US is now producing more total liquids than OPEC crude output. That’s not a typo. US total liquids production sits at ~23.4 mb/d, essentially flat year-over-year. OPEC total liquids production has fallen to ~23.5 mb/d, down 10.3 mb/d from 2025. American energy dominance used to be a slogan. Now it’s in the data.

Needless to say, you want to be long folks that are levered to the gap between European/Japanese and American gas prices, at decent valuations, with a lot of upside. For paid subs, we’ll put the individual names we think are the best bets in the Appendix below, along with a link to a dashboard which provides more depth.

And here’s the longer-term structural point. The UAE has massive spare capacity that they are going to bring online. As soon as they can export more, they will. And they’re going to work to bypass the Strait doing it. That’s bearish for OPEC discipline and bearish for the narrative that supply cuts are sustainable. But it’s bullish for total global supply eventually, and it reshapes who the winners and losers are in the energy complex.

The world keeps trying to price this conflict like it’s a two-week event. The production data says otherwise. The inventory data says otherwise. The curve says otherwise. And now OPEC itself is fracturing.

I’ve avoided trading oil for the last couple of years for reasons I won’t get into here. But as the Dread Pirate Roberts once said:

I am not left-handed either.

It’s time to put the sword back in my fighting hand.

Disclaimers

This analysis was built over several days of active markets. Prices, curves, vol surfaces, and positioning data reflect the time they were pulled, not the time you’re reading this. Some of these levels will have moved by the time you see them. Do your own work. Check the chain before you trade. The price on your screen when you lift the offer may look nothing like the price on mine when I wrote this up.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counterparties.

The Ramble does take positions in various assets we write about, and views are subject to change without notice.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure. Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION.

Appendix: US Energy Dominance (Paid Subscribers)

We built this dashboard to screen the US gas and LNG equity universe across value, growth, spread beta, quality, and what we call “setup” (how much of the move is already priced in). A few names jump out.

The “Sweet Spot” ranking surfaces names with the best combination of LNG torque, reasonable valuation, and balance sheet quality. VG (Venture Global) sits at the top of the core bucket: maximum LNG torque with high leverage to the legal/contract overhang resolution. CRK (Comstock) leads the small/levered bucket: Haynesville gas producer near the LNG corridor with a clean balance sheet and solid hedge matters.

LINK

(work in progress, data may be lagged since we pulled this together over the weekend)

Back to flat on the year. Oof.

Not a disaster. But the reason matters.

The market priced resolution. I’m positioned for recovery amidst chaos. Those are not the same thing. This is why you take down risk when you feel like you are early, so you can save your bullets for when the world starts to price into your conviction.

Quick framing for anyone new. This is my alpha/convexity book, not my absolute return portfolio. As discussed in our last piece, the absolute return portfolio layers this on top of a beta portfolio that’s still under construction. The goal is to provide convexity on top of my diversified beta, with a lean-short bias given I have a lot of illiquid startup equity sitting off-balance-sheet. By summer I’ll tie the two together into something that looks more like a traditional investable portfolio. For now, think of this as half the book sitting against a good diversified set of “all beta.”

The Mistake

I got the macro right and the sequence wrong.

Oil shocks don’t hit all at once. They propagate. Energy to fertilizer to crops to CPI to rates. The market is still trading step one. I’m positioned across steps two through four. That timing mismatch cost me.

Here’s the honest part: as I try to trade at higher frequency and complexity (in part wanting to show process for you kind folks), I’m taking the kind of positions I used to take at Lehman where I would watch them all day. Instead I’m making mistakes in risk, ops, and implementation that account for about half of the drawdown. You could argue these are things that wouldn’t be an issue if I were doing this full time or had help. I hope that’s true, but I need to demonstrate it. Makes me think harder about the timeline to turn back to the Ramble and the portfolio full time. Automation is a sexy buzzword, but right now it’s costing me 10 hours for every hour I save.

The View

I don’t see a Venn diagram where the two sides meet. Trump seems to be getting comfortable either waiting Iran out and doing the rimland strangle (hoping eventually China forces them to the table) or waiting for the Iranians to trip another wire of escalation. The one thing that seems clear is that US stocks want to shrug it off. That was the wrong implementation of my view. Oil and the knock-on effects of Hormuz closing cannot shrug it off.

The US economy looks to be reflating, as do others, though it’s hard to disentangle the inflation component in oil-importing countries (UK for example). That led me to add some cable shorts and a bit of oil upside Sunday night. Given the size of my commods and duration exposure, I’m getting pretty exposed to higher rates and a strong dollar. Bought some cable puts defensively since it covers some dollar exposure and captures the tail of the UK losing control of their rates market, which would pull SONIA up but crush GBP.

The curve is pricing resolution. It’s not pricing recovery. If stocks continue to rip while Hormuz stays closed, eventually bonds will need to sell off as folks price in an inflationary risk on. This would in turn crush many of the hot AI plays which we had on in size before the war. Which is why we remain under allocated and paying theta.

The Book

We’ve taken VaR down to ~1% (around 16% expected annual volatility, down from ~40 in January) along with the increase in uncertainty. This understates our upside volatility though, as we have a lot of embedded convexity in the portfolio.

Read more

I think they don’t know what beta they own.

You see it both ways. Institutions paying for “alpha” that’s just a factor with fees on top. Individuals concentrating into a few names and calling it conviction when it’s really just leverage.

Different paths, same outcome: a portfolio that looks diversified until it isn’t.

That’s the actual problem.

Eagle-eyed readers have reached out to complain that I have a tendency to start series and never finish them. Which, if you're still waiting for Systems 103 or the Campbell builds the AI-investing-machine-god piece from October, is a justified concern. But today we're going deep into the archives to continue a series from 2017 called "Whatever Happened to Alpha?" Partially because we ran a reader poll and portfolio construction won. Partially because I'm thinking about starting my own fund. And partially because after the last couple of weeks — having just missed out on a 10% rally, eye roll — I wanted to spend some time on my beta portfolio.

We tend to slant the alpha book away from having too much levered beta. Call it a hangover from trying to sell institutions a portfolio that would diversify them from their growth-stage tech exposure, or from building my own alpha book to hedge against too much startup equity. Either way, the bias shows up the same way: real diversifying alpha is the most valuable thing in the investable world, and most of what gets sold as alpha isn't that. But for most people, the bulk of the portfolio should be beta anyway. Which is what this piece is actually about.

We’ve talked about beta vs alpha before, but as a refresher: beta is the risk premium you get from holding the market portfolio. Alpha is the return you get by deviating from that portfolio. This piece gets technical, so for those without a lot of time:

A lot of what people call alpha is levered beta. Institutions invest in funds that dress up levered beta as alpha. Individuals, facing leverage constraints, concentrate into high-beta stocks that carry implicit leverage in their business model or balance sheet.

The former is expensive. The latter is dangerous. Retail concentration introduces idiosyncratic risk that increases your road to ruin.

Both would be better served building a basket of diversified betas. What makes up that basket depends on liquidity, risk tolerance, instrument access, and existing exposure. Tech founders should note how much tech is already in their "passive" retirement portfolio. Same principle as oil-state sovereign wealth funds not holding oil.

The funny thing about spending two years building an alpha book is realizing it was defensively shaped by the beta you didn’t know you had.

For retail readers looking for a practical framework: if you're asking this question, you are probably overallocated to alpha. Reorganize to 80-90% beta and 10-20% alpha, at least until you've demonstrated a 0.5 Sharpe (meaning if you're running 16% vol, your excess return over the risk-free rate averages 8%). Below that, put almost everything into beta and focus on earning more to invest.

Three questions to answer about your beta: what volatility can you stomach, what drawdown is unacceptable, and what access do you have to leverage. If your answers are something like 10-12% vol, 25% max drawdown, and not much leverage, you're probably looking at something like Bridgewater's All Weather (ALLW, 85bps) or Advanced Research Risk Parity (RPAR, 50bps). Full disclosure, I used to work at Bridgewater and many of my friends still do. Neither is cheap, but they provide 1.5-2x gross leverage in a set-it-and-forget-it wrapper — which is the part you can't easily replicate yourself at retail scale.

Now for the fun stuff.

You Think You’re Diversified

If you own 40 stocks, you are not diversified. You are long one trade, in 40 slightly different wrappers.

As you add stocks, the idiosyncratic piece dies. What survives is the systematic part. With normal equity parameters (σ ≈ 20%, ρ ≈ 0.3), the marginal benefit of adding a stock drops below noise by about 12-14 names. By 30 it’s a flatline.

The illusion holds when markets are calm. Tech up, energy up, consumer up, each by a different amount, and it looks like you have a bunch of different bets. Then something breaks. Correlations spike. You remember what you were always holding.

What you own, after the math is done, is one thing. Equity beta.

Return vs. Risk

This chart tells you most of what you need to know. Retail portfolios cluster in the bottom left: high volatility, moderate returns, low Sharpe. Institutional portfolios sit further up and to the right: better diversification, better risk-adjusted returns.

Both want to move up and to the right. They get there through different doors.

The Institutional Tradeoff

Institutions have leverage. They don’t need to pay someone to provide it. So the question isn’t “how do I get more exposure?” It’s “how do I get a better Sharpe?”

Hedge fund alpha is real. When we compare the hedge fund index against our “All Beta” portfolio (which just equal weights 100+ macro betas from equity, bond, currency and commodity markets), we find two things, the average fund is around ~50% correlated to their relevant beta (sometimes more) and funds in aggregate have roughly 5% gross alpha (though note the diversification benefit from grouping them into a single portfolio.

But at 2/20, the manager takes about half. Net alpha to the investor is maybe 2.5%. And if that manager is running with a correlation (ρ) = 0.7 to your existing equity book, the marginal contribution to your portfolio’s Sharpe is close to zero. You’re paying for levered beta in a nicer wrapper.

Which brings up an interesting point, for an institution, the manager worth paying for is the one with genuine low correlation. A 0.5 Sharpe at ρ = 0.1 adds more to your portfolio than a 1.0 Sharpe at ρ = 0.8. The marginal contribution scales with correlation as much as with standalone performance.

Which means institutions should actually be more willing to back less proven managers with differentiated books than to write another $250M check to a platform hedge fund running the same equity tilt they already own.

The Retail Tradeoff

Retail doesn’t have leverage. So the entire game changes.

Most retail investors reaching for returns aren’t chasing alpha. They’re trying to synthesize leverage. And they’re doing it through the worst possible door: concentrated high-beta tech, 3x daily-reset ETFs, meme stocks. Instruments that are 90%+ correlated to equity beta, path-dependent, fee-heavy, and designed to decay in exactly the environments they’re supposed to work in.

The right move is counterintuitive. Step one: diversify. Going from 40 correlated stocks to a diversified beta book and your vol drops from 18% to 8%. Your return barely changes. Your Sharpe roughly doubles.

This is the key. This is why folks call diversification ‘the only free lunch in investing.'“

Step two: lever. You end up back at 16% vol, but now you’re on the 0.65 Sharpe line instead of the 0.3. Same risk. Fifty percent more expected return.

Now, this also explains WHY retail does what they do. Usually, they lack good sources of leverage. Be it from restrictive / protective brokers, or retirement accounts where it is verboten. Pushing people into more esoteric, difficult to manage instruments like futures, and options.

This is why 3x levered ETFs exist, this why folks blow themselves up on meme stocks. Because the leverage constraints forces upon retail push them into more and more dangerous, confusing or expensive ways to play with leverage.

The All Beta Portfolio

Here’s where this gets concrete. Back in 2017, we built a macro beta portfolio. Twelve asset classes, full correlation matrix, equal-risk-contribution weights. We’ve kept it updated since.

A “beta for dummies” (like me) ETF version tracks the futures version closely. Which means for most people, you can build something very close to this with a handful of cheap ETFs.

Going from a bad beta portfolio to a decent one is where most of the value lives. Going from decent to perfect is small, uncertain, and requires judgment you should probably deploy elsewhere.

The Regime Problem

One caveat. The All Beta portfolio, like all risk-parity type strategies, has benefited enormously from the 40-year decline in interest rates. Bonds diversify equities in a disinflationary regime. We’ve argued since 2020 that this regime is shifting. Conflict is inflationary. In that world, bonds and equities correlate positively (2022 was the preview), and the traditional bond allocation has to shrink. Gold and commodities become the diversifiers.

It also means if you tried using the relationship between the blue line (bonds vs trend) and the red line (all beta returns) above to trade beta returns, you would have had a really good time up until we saw a resurgence in inflation in 2021.

Your beta portfolio is a function of the regime you’re in. Which is the first moment where what you call “beta” starts quietly becoming something else. Which calls back to one of the first conversations I had with the big man, way back in 2012. “If everyone listens to you about your optimal beta portfolio, doesn’t that to a reallocation in assets that pushes up your portfolio? Does that give your beta portfolio alpha exposure?”

Anyway, five years layer, that chart above was the start of the work that led us to start increasing our gold allocation to dangerous levels, and to advocating in 2017 to both individuals and institutions to try “stocks and gold” vs “stocks and bonds.”

That being said, Gold is the position I’m most conflicted about right now. Strategic view says this regime wants more gold, not less. Tactical read right now says the opposite: flows are crowded, gold is trading more like a bond than a diversifier, the negative correlation to oil keeps surprising me. I’m running gold heavier than the ETF book below would suggest, but I’m watching it closely. More on this in a later piece when we revisit our gold view.

How Much Beta Is in the Alpha?

Scale this up and most of the “alternative” universe collapses. PE is levered equity with a smoothing lag. Most hedge fund strategies are levered factor tilts available in an ETF for 15bps. Private credit is levered corporate bond beta with an illiquidity premium.

Standard error on a Sharpe ratio is roughly 1/√T. A one-year track with an observed Sharpe of 1.0 has a 95% confidence interval from zero to 2.0. So, you really need quite a bit of time to really understand if someone is good at this whole money management thing.

Where This Leaves You

If you’re running a meaningful chunk of your book in active picks, there’s an implicit claim in that decision. You’re claiming the residual after you regress against your beta book is real and persistent.

If you haven’t measured it, you don’t know what it is.

I spend most of my time here writing about the alpha book. The part of the portfolio that’s actually working right now is the beta book. The alpha book was shaped, for two years, by whatever the beta book was doing and the exposures I had in my life.

Most portfolios don’t fail because they were wrong. They fail because they were one bet pretending to be many.

Next time: the zany math and the professional-grade implementation. Residual Sharpe decomposition, Bayesian shrinkage on short track records, the leverage multiplier that makes Sharpe the only variable that matters, and a head-to-head backtest of four beta constructions — pure equal-risk-contribution, synthetic All Weather, leverage-constrained retail (RSSB-style), and futures-overlay institutional — with realized Sharpes, 2022 drawdowns, and the real cost difference between the retail and institutional paths. Bring a calculator. Subscribers get it first.

Building the Book

For those who want an actionable implementation in the meantime. With the huge caveat that I don’t know, man. I’m spending a lot of time in my alpha book trying to hedge or short stocks and I’m this close to putting on a big steepener (selling inflation-sensitive long-term bonds to buy growth-sensitive short-term bonds). So you probably want to wait until we see another 10% drawdown here before yoloing in. Yes, you can be mad at me for not telling you this a couple of weeks ago, and that would be fair, but this is how we do that. It’s exactly that gap that reminded me to write this piece.

The ETF Weights

One frame before the table. This isn’t a pure All Beta book. A proper All Beta is equal-risk-contribution across twelve macro sleeves using the full correlation matrix, and it’s the thing I’ll build out properly next time. What’s below is a practical half-step — closer to a simplified All Weather than a pure factor-neutral construction — and it’s the version that’s implementable in any brokerage account without a covariance matrix and a spreadsheet. The proper version comes with the backtest, in Part 5.

Blended expense ratio: roughly 12bps. No leverage. Runs about 7-8% vol unlevered.

A few things to notice. TLT at 20% is the biggest single weight because it provides the most diversification per unit of risk in the historical sample. That’s the weight that should make you nervous in the current regime. If you believe (as I do) that we’re in a conflict-inflation shift, you’d want to trim TLT and add to GLD and DBC. Which is already an alpha call on top of the beta book. See how fast the line blurs.

How to Lever It

If you have futures access. Don't sell anything. Hold the ETF portfolio as collateral. Overlay 3-4 micro E-mini contracts (/MES, roughly $26K notional each) for the growth sleeve, 1-2 /ZN contracts for duration, and 1 /MGC micro gold contract (roughly $34K notional) for the gold sleeve. The ETF portfolio stays intact. Effective gross goes to about 200% on $100K equity. Vol goes from 8% to 15-16%. No daily reset, no drag. You roll quarterly. Implicit financing cost is close to the risk-free rate, which you're already earning on undeployed cash. Available at IBKR with futures approval. One note: this three-contract overlay levers equity, duration, and gold but leaves TIPS, commodities, REITs, and EM un-levered — so scale up those ETF allocations proportionally before adding the futures, so the un-overlay sleeves don't get diluted. (If you want to go further, /MCL micro crude and /MHG micro copper exist for the commodity sleeve, but they're less liquid and you need a bigger book so you don’t get unbalanced.)

If you don’t have futures access. Return-stacked ETFs. NTSX gives you roughly 1.5x through 90% stocks + 60% intermediate treasuries at 20bps. RSSB pushes to 2x through 100% stocks + 100% bonds at 40bps. Build around RSSB as your core equity/bond exposure and add the commodity/gold/EM sleeves around it.

If you have $125K+ at a decent broker. Portfolio margin lets you gross up to about 2x on a diversified book at a financing cost of a few percent. Cleaner than daily-reset ETFs. The carry still eats returns in a way futures don’t.

Never. 2x and 3x daily-reset ETFs. TQQQ, SPXL, SSO. The daily rebalancing forces the fund to buy highs and sell lows every trading day, compounding a drag that eats 3-8% annually depending on realized vol. You’re paying for leverage in decay instead of in financing cost. Decay is always more expensive.

As you can tell there’s a lot of ways to skin this cat, and again, sometimes the easiest thing to do it just plop your beta in an ETF like ALLW or RPAR and pay the vig. Anyone who has messed around with futures knows you are always likely to miss a roll when you most need it.

Rebalancing

Monthly or quarterly. Don’t overthink this. The point of risk parity is that the weights are set by volatility contribution, which moves slowly. If something blows out (say TLT drops 20% in a quarter), that’s when you rebalance into it, not out of it. Rebalancing is the mechanism by which you harvest the diversification premium, buying what just got cheap relative to risk.

What I’d Change for the Current Regime

More gold, more commodities, less duration. I’d take TLT from 20% to 15%, move 2.5% to GLD and 2.5% to DBC. Which gets you gold at 17.5%, commodities at 12.5%, and long treasuries at 15%. That’s a meaningful regime bet, and it’s already an alpha call layered on top of the beta. Which is, again, how fast the line blurs.

I’m not telling you to do this. I’m telling you what I’m doing. And the reason I’m writing it down is so that when I’m wrong, I can point to this and trace where the mistake was.

The proper version — a pure equal-risk-contribution build without the regime tilt, backtested against All Weather, AQR’s risk parity with the leverage paths compared head-to-head — is for the next time we hit this space. Bring your calculator and a couple of data frames.

Disclaimers

This one is reaaaally not investment advice, it’s intended to point you in the right direction to understand how to get access to good betas. Your risk tolerance, liquidity needs, and leverage situation will vary and inevitably make this bad advice relative to your actual situation. That’s why investment advisors get paid. This is a newsletter. Don’t trust machines, don’t trust a backtest.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counter parties.

The Ramble does take positions in various assets we write about, and views are subject to change without notice.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure. Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

Whether or not it’s working is up in the air. Crude is up 40% since the strait closed, but down ~5% since Trump’s blockade was announced. In response Iran threatened to close Bab el-Mandeb alongside Hormuz. Both chokepoints now at risk. VLCCs too large for Suez to carry oil at full weight. And yet, rumours that Iran was willing to bench on uranium enrichment, the key issue blocking any medium term solution. Stocks rallied. The market has priced the first order impact on energy and growth, whether or not appropriately, the assets move with the headlines.

What the market has not priced in is the second node in the chain. Or the third.

This article is about those nodes.

There’s this joke I post on Twitter every now and then, when one of my deep out-of-the-money calls hits.

“My hyperstitions usually take 6-18 months.”

Hyperstition being one of those SF/AI-speak words for NYC/Soros’s reflexivity. A prediction which becomes more true in the act of making it. The idea that making a call publicly sometimes influences the outcome. Markets work this way directly. You lift an offer, the price moves, other people react. Classic examples being Soros breaking the Bank of England or the market ganging up on LTCM.

I’m bringing this up because when I look at my best trades/hyperstitions, they come in two forms. The first is fast synthesis under pressure. Over the past few weeks there have been moments where a speech or news item hit the tape and I made 20-50bps by synthesizing quicker than the market. Trump’s speech where he opened with the length of WWI and Vietnam and then said “32 days”, he wasn’t about to say “and I’m OUT.” That kind of punting is real but it’s not scalable. You end up hitting 2/3rds, but the opportunities come almost at random, and you have to be glued to the screens to capture them. Plus, there are only so many Druckenmillers, and markets are more efficient than in his day so it’s hard work and not something you are going to retire on.

It also didn’t quite cover my theta bill, which is growing.

The second form is the one I hang my hat on. Looking at a chaotic, ambiguous system and identifying what idea or event is just outside the Overton window and about to come into frame. Which then results in a radical repricing. Gold in China from 2018. Silver from 2023. AI infrastructure from 2023. Rare earths in ‘24. All of them looked picked over or fundamentally unpopular when we started building positions. All of them took 6-18 months to reprice.

The cascade trade is this second form. The damage is locking in now, on biological and chemical timelines the market doesn’t track. The repricing comes later. The longer the heart attack in Hormuz lasts, the greater the repricing.

Quick update on the book before we get to the meat.

Credit short has been a waste of premium. We had an opportunity to capture gamma when it dipped below our strike, but late on the trigger meant we missed the move before headlines ripped through the top strike (I really need a trader to run all this stuff while I’m in meetings and traveling but again, no crying in the casino Campbell).

Lesson learned: when you’re trading credit via puts on actual bonds (HYG), your duration hedge gets messy. As the corporate bonds fall through your first strike, you pick up a lot of delta (which also means duration), and then if the risk-free rate compresses on growth fears your puts lose even if spreads are unchanged. I rolled the short exposure into May SPX put spreads, 1.5-4% out of the money. CDS remain the right instrument for this leg. We’ll get there, someday. Even if it means building an ISDA scale book brick by brick with newsletter subs (kidding).

VIX calls from Friday worked, for a little while and ended down on the day . I bought them expecting the fragile peace to break. What I didn’t expect was Trump taking my advice and counter-escalating to a full naval blockade so effectively that Iran came back to the table and began caving on enrichment. To me this is a real line in the sand that I am looking for to shift my peace balance. It’s also why stocks are rallying into what could be a radical escalation. I took off around 1% of crude call spreads after the news, down 15% from the daily highs, but still up on the trade.

Couple surprises in one day and the book still had a fine session, mostly bailed out by the rates and equity book (long AI equities, critical minerals, and energy exporters). The US equity market is either a monster or entirely delusional. Probably both.

Gold is trading terribly. Correlated with stocks, underperforming on the squeeze. Heavy. The Turkey/Middle East deleveraging from last month is still spooking people. But longer term, if Iran is demanding toll payments in yuan and crypto, this has to be constructive for gold. I’m getting more bullish, not less. Slowly moving out of delta one and into long-dated call spreads and flies. If we do get peace here, I’ll probably stick 5% of the book in long-dated broken butterflies and just stop watching it. IF you have friends looking to launch a gold-backed stablecoin, I’m in.

Which brings us to our ‘cascade positions’: our calls are underwater, down about 1% combined across a set of strikes and contracts. We’re early. And we added winter wheat futures after doing a ton of vol work which convinced me to puke the vol and just buy delta. Seems to use the wheat wheat market is looking at a chart that includes actual supply destruction from Ukraine in 2022 and pricing implied vol /skew way too high relative to what the current setup justifies. More on this below (maybe more than you want).

In terms of the outlook.

We continue to stand at the edge of war and peace. Today brought moves that suggest we may actually be close to some sort of resolution. We still want a portfolio that is balanced to both outcomes.

If Iran folds on enrichment, there are real odds Trump walks, claims victory, and pivots to stimulating the economy ahead of midterms. Venezuelan regime change, Iranian nuclear deal, Indonesian security deal, revitalized the Abraham Accords. Forcing the difficult conversation on NATO and heading into the USMCA review this summer with a very aggressive posture (watch out Canada, it’s going to hurt). There’s an argument to American voters that the sacrifices were worth it. The counter-blockade could be another “Leave the Bride at the Altar” (Stage 5 in the maximum pressure playbook) designed to extract last-minute concessions.

On the other hand, military assets are still moving into the region. Lebanon is unresolved. We don’t know who’s actually in charge in Tehran or what they’re willing to accept. China would clearly prefer the US get tied up in the heartland, but maybe by subjecting them to the same pain as Europe and the rest of Asia (and deepening their economic pain from property which appears never-ending), we can bring them to the table. Remember where all this military kit is coming from after.

Which brings us to the trade structure. We want a portfolio that is balance to peace and war.

For peace: UK SONIA short rates. The UK is the most energy-import-sensitive major economy. If this resolves, oil falls back to $70, inflation expectations collapse, and the Bank of England cuts aggressively. SONIA Mar'27 futures are pricing that world at maybe 30% probability. We think peace odds are higher than that. The position has already paid us on the ceasefire rally, and if Iran folds on enrichment it likely pays enough to cover the grain book's losses in full. That's the portfolio construction: SONIA profits on peace, grains profit on war, and the calendar grinds in our favor either way. See our prior write up and Jason’s substack for this idea in depth. A lot of people will just yolo stonks here (and clearly already have - stocks are back at the highs!) but we want an expression with some more convexity to the upside.

For war: enter the grains. Add if things deteriorate. Every day the strait is closed and energy remains elevated, the combination of supply destruction and demand pressure is building, currently hidden by markets that are in supply glut for the short term. Exactly the kind of trades I specialize in. Early, but explosive.

The cascade thesis doesn’t require the worst case. It requires the calendar to keep running while the market prices the first node and ignores the second. Even in the most optimistic scenario, Hormuz will take time to normalize. Energy prices will remain elevated for months, not weeks. The flow-through to fertilizer prices has already happened and won’t reverse on a handshake. High diesel at $5.98/gallon leads to abandoned marginal fields. Ethanol economics pull corn off the food market to supplement refined products. That’s the inherent convexity here, and why I probably want to continue to own these even if we do get peace in the next couple of weeks (after our gains in SONIA offset the likely losses in these positions).

The Three Commodities

Three crops, three mechanisms, one blockade. Each has a gate. A calendar window where the damage locks in permanently. Sugar's gate is the ‘crush allocation’ (choice between sugar for your car vs sugar for your tea), opening now. Wheat's gate is the middle east demand bid, already open and kicking in earnest in the next month. Corn's gates are stacked: acreage, pre-plant, sidedress, each closing in sequence through June.

Sugar is the most direct link. Sugarcane is one crop that becomes two products: sugar crystals or ethanol. Brazilian mills decide at crush time. The Centre-South crush started April 1. At $110+ Brent, ethanol is more profitable than sugar at the mill gate. Allocation shifts. The 11.2 million metric tonne global surplus starts shrinking in real time. UNICA fortnightly crush data starting reporting in late April and gives us near-real-time confirmation. Swap dealers are net long 157,000 contracts, near the highest since 2008. Managed money is pressing short at 56,000 contracts against them. The physical end users see something the specs don’t. The self-correcting feedback (mills shift back if sugar rallies above 18-19 cents) caps the move, which means the trade is 14 to 20, not 14 to 30. We remain long sugar call spreads.

Wheat is the cleanest setup because it doesn’t need a perpetual blockade to work. Winter wheat opened the season at 35% good-to-excellent, worst since 2022. Oklahoma 54% poor-to-very-poor. Planted acres are the smallest since 1919. That’s a domestic supply problem that exists if the strait reopens tomorrow. Hormuz adds the demand side: MENA food security buying is a political imperative, not a discretionary purchase. Bread prices are regime risk. Egypt, Algeria, Iraq don’t cut wheat imports because prices rose 20%. They panic-buy. KC HRW implied vol at 25.6% against 24.7% realized is basically fair, though skew is intense. The market is not pricing a panic bid into the smallest US wheat crop in a century during a naval blockade. We added winter wheat futures (after rolling out of our calls last week, which I’ll walk through in the vol section below).

Corn has the biggest buffer and the longest fuse. Record production at 17 billion bushels, ending stocks at 2,127 million. The cascade changes the math through farmer decisions at specific calendar gates, each one permanent once the window shuts. Pre-plant nitrogen (happening now, 40% buying at $924 spot). Acreage allocation (June 30 reveal). Sidedress nitrogen (the binary gate in June, where 15-20 million acres either get their top-up or don’t, forever, after V8 around June 22). The corn trade needs the blockade to persist through sidedress AND the July WASDE reports to revise yields. Three gates, sequential, uncertain. We hold longer term (fall/winter) calls but the catalysts are closer and the mechanisms more direct in the other two.

Over 21 million tonnes per year of nitrogen capacity is now physically offline, destroyed, or export-blocked. QAFCO in Qatar, 5.6 million tonnes, hit by strikes, 3-5 year repair timeline. Yara Pilbara in Australia dark until late May. Nutrien Trinidad idle since October. India shuttering plants as Qatar’s LNG stops flowing. China blocking exports. And Nutrien’s Lima, Ohio plant about to go offline for a 72-day turnaround in August through October.

The supply isn’t delayed. It’s gone. The calendar keeps running.

What follows is the deep dive. Commodity by commodity, the supply and demand picture, scenario models across three blockade durations (one month, three months, twelve months), the vol work, the positioning, and the specific trades.

We start with sugar because the crush is happening now. Then wheat because it’s harvesting now and the demand story is already live. Then corn because the gates are still closing and the biggest catalyst (sidedress) is two months away.

Sugar: The Ethanol Diversion

Sugar is the most direct link in the cascade because it skips the fertilizer node entirely. The transmission runs through oil price.

Sugarcane is one crop that becomes two products. Brazilian mills crush the cane, extract the juice, and decide at that moment whether to crystallize it into sugar or ferment it into ethanol. The Centre-South crush started April 1. At $110+ Brent, ethanol is more profitable than sugar at the mill gate. Every $10 move in crude shifts approximately 1.5% of Brazil’s sugar-to-ethanol allocation. That sounds small until you remember that Brazil produces 45 million metric tonnes of the world’s 189 million. A 4-point shift in allocation is 3.5 MMT of sugar that simply ceases to exist. The cane is still growing, still getting harvested, still getting crushed. The mill just pointed the juice at a different tank.

This mechanism became structural in 2003 when Brazil launched flex-fuel vehicles. Before that date, sugar squeezes were pure supply shocks: Soviet crop failures, Cuban embargoes, Indian monsoon collapses. After 2003, every major sugar rally has involved the ethanol channel as either the primary driver or a major accelerant.

The lag between oil moving and sugar following has historically been 2-6 months. The lag is the crush calendar. Mills can’t reallocate until the new season begins. The 2026/27 season started twelve days ago. If the pattern holds, the red line should start moving by May/June.

The balance sheet

World sugar production sits at 189.3 MMT (million metric tons) against consumption of 178.1 MMT, a surplus of 11.2 MMT. That’s the most comfortable balance since 2017. ICE #11 at 13.75 cents reflects this reality accurately. The market is not mispriced for the world as it stands today.

Brazil’s Centre-South produced a record 45 MMT last season at a 50.7% sugar allocation. India came in at 28.3 MMT, below domestic consumption of 28.5 for the second consecutive year. The government approved a 2 MMT export quota but only 300,000 tonnes have shipped because Indian breakeven costs sit around 18.5 cents, well above the world price. Thailand is steady at 10.3 MMT. EU beet sugar contracted to 14.8-15.5 MMT as collapsed white sugar prices drove an 8-10% reduction in harvested area.

The surplus is real. The question is how much of it survives the crush.

Where the stocks actually sit

The headline number, 44.5 MMT of global ending stocks, is misleading for trade purposes. Most of it is locked up in countries that will never export it.

Thailand holds 10 MMT. That’s accumulated unsold inventory from exports running below expectations. It’s available to the world market but Thai mills need a price signal to release it. India holds 8.5 MMT, almost entirely domestic buffer against their own production shortfall. China’s 2.4 MMT is strategic reserve. Pakistan’s 2.3 MMT is domestic. US 1.7 MMT is managed by Commerce Department quota programs targeting a 13.5% stocks-to-use ratio. Out of 44.5 MMT of global stocks, the truly exportable, available supply is perhaps 12-15 MMT. If Brazil diverts 4-7 MMT from sugar to ethanol, it’s eating into the available surplus, not the headline number.

Three scenarios

Closed one month. The crush is only 6% complete by early May. Allocation slips to 49.5%. That’s 0.5 MMT diverted. The surplus is unchanged at rounding level. Sugar moves to maybe 14.80. This is noise.

Closed three months. If crude sustains above $110, this is where things start to really move. São Paulo hydrous ethanol parity sits above sugar at the mill gate. Allocation drops to 46.7%. Brazil CS sugar production falls to 42 MMT, down 3 from baseline. India produces 28 MMT, again below domestic consumption. EU beet sugar contracts further to 14.8 MMT. Total ethanol diversion: 3 MMT. The surplus shrinks from 11.2 to roughly 9.5 MMT. ICE #11 moves to 17.50. The 56,000 contracts of managed money net short starts covering.

Closed twelve months. Crude at $120. Allocation collapses to 42.2%. Brazil CS sugar falls to 38 MMT, down 7 from baseline. India slips to 26.5 MMT as the El Niño building at 62% probability disrupts the monsoon. EU contracts to 13.5 MMT. Thailand loses 300,000 tonnes. World production drops to 185 MMT. The surplus narrows to 6 MMT. Still positive, but the trajectory points toward a 2027/28 deficit. ICE #11 reaches 22 cents.

The self-correcting feedback caps the upside. If sugar rallies above ethanol parity at 18-19 cents, mills shift allocation back mid-crush. The supply response is measured in weeks, unlike cattle (years) or corn (one growing season). This means the trade is 14 to 20. Not 14 to 30. Every sugar squeeze since 2003 has had this governor. The ceiling is real. Size accordingly.

The cascade flow

The supply path through fertilizer and diesel barely matters for sugar. Sugarcane is a perennial ratoon crop with lower nitrogen intensity than corn, and Brazilian costs are denominated in reais, which cushions dollar-denominated input spikes. The demand path through crude oil into ethanol economics is 85%+ of the story.

Positioning

This is where the trade got more complicated than I expected.

The legacy COT report shows commercials net long 67k contracts against non-commercials net short 71k. That’s the number everyone cites. The most extreme divergence in 18 years. And it’s real, but it hides the mechanical detail that matters.

Pull up the disaggregated report and the picture sharpens. Swap dealers are net long 157k contracts, near the highest reading since June 2008 (which preceded the 13-to-36-cent rally). They added 8k contracts this week. Still accumulating. Seventeen firms holding $3.5 billion notional of physical hedging demand. These are the intermediaries standing between Coca-Cola, Nestlé, and every food company on earth and the futures market. When a swap dealer is long 157k contracts on the exchange, it means their OTC clients have locked in supply at an unprecedented rate. Physical end users see something.

Managed money sits at -56k net short, and they pressed 15k contracts further short this week. They’re actively betting against the physical market. But here’s the update that changed my Bayesian prior on this trade: managed money has already covered a substantial amount of their short position over the past two months. They were likely -150k+ in mid-March based on the CFTC disaggregated data I cross-checked. The move from deep net short to -56k, and price barely twitched. Went from 13.30 to 14.50. The physical surplus absorbed the buying flow. Producers sold into the covering.

So the simple “specs cover and price rips” narrative is weakened. A lot of the covering already happened, and it moved the market one cent. What hasn’t happened is the physical catalyst that removes sugar from the world market. That’s the UNICA allocation data. When managed money’s remaining 56k shorts try to cover into a market where the physical side is bid and the sugar supply is literally becoming ethanol, the absorption capacity of producers disappears. That’s the difference between a one-cent grind and a three-cent snap.

Volatility

ATM implied vol on Jul ‘26 sugar sits at 30.5%. The call skew is steep. The vol surface is pricing the upside tail reasonably but not aggressively.

Since 2003, every oil spike has been followed by a sugar rally within 2-6 months. The lag equals the crush calendar. The vol surface doesn’t price this historical pattern because it’s pricing the current surplus, not the conditional path where oil stays elevated through the crush. If Brent holds above $100 through June, realized vol on sugar will exceed current implied.

In order to pick our strikes, and size our confidence. We took the liberty of pulling out our secular timeseries of sugar prices to do a little math here. Over 160 years of sugar price data, a 10% rally within any two-month window has occurred roughly 12% of the time. That 12% is unconditional. It includes decades of quiet markets with no catalyst. Conditional on the current setup (oil above $100, crush underway, positioning at 18-year extremes), the historical analogs compress to a handful of episodes and the hit rate is closer to 80%. The vol surface at 30.5% ATM is pricing a 15-18% probability of that move. The conditional rate says much higher.

The trade

We remain long sugar call spreads, Jun ‘26 expiry. Max gain per contract/spread is approximately $1,454 per spread, max loss $213. That’s a 6.8:1 payoff ratio at expiry which becomes 3:1 pretty quickly. We will roll them once the price starts moving and we get to the middle of the spread, since this is a relatively short term option. Size to your preference.

The defined-risk structure is deliberate. This trade needs the physical catalyst (UNICA allocation confirmation) to work, and we don’t know when or whether it arrives. If crude drops back to $80 on a ceasefire and UNICA shows 50%+ to sugar, these likely expire worthless. That’s the $213 per contract outcome. If the three-month scenario plays out, the spread captures the bulk of the move. The June expiry gives us the first two months of UNICA data plus the peak crush volume window.

Sizing: this is a 1-2% of book position. The conviction is high on the mechanism but uncertain on timing. The call spread structure means maximum loss is known and bounded. If the first UNICA reports confirm the allocation shift, I’ll add.

What to watch

Late April: First UNICA fortnightly crush data with real volume. This is Gate 1. If the opening allocation reads below 49% to sugar, the surplus narrative starts cracking. The data publishes on UNICA’s website, typically the second and fourth weeks of the month during crush season.

Crude oil price. This is the direct input. Not fertilizer, not weather, not acreage. If Brent falls below $90, the ethanol parity math flips back toward sugar and the thesis weakens. If Brent stays $100+, the allocation shift accelerates with every fortnightly report.

June-August UNICA reports. The crush peaks in these months. The cumulative allocation math becomes undeniable by the third or fourth report. This is where the surplus erosion chart (four scenario lines above) diverges meaningfully.

Managed money short covering pace. Track the weekly COT. The -56k is fuel. When it starts burning, price will move faster than the one-cent grind we saw from March covering. The difference is whether producers have physical sugar to sell into the bid. If UNICA confirms the ethanol diversion, they don’t.

Brazil corn ethanol capacity. The one structural risk to the thesis. Rabobank flagged 3+ billion liters of new corn ethanol capacity in 2026. If corn-based ethanol absorbs the E30 mandate increase without needing cane ethanol to flex, mills can keep 50%+ to sugar even with high oil. Check the Rabobank estimate against actual capacity additions.

Historical context

Six sugar squeezes since 1963. Every one shared three ingredients: a supply disruption or deficit, a positioning extreme, and a catalyst that forced shorts to cover before new supply could arrive. The current setup has two of three confirmed. The positioning is still extended. The catalyst (Hormuz into oil into ethanol) is live and observable. The confirmed deficit is what’s missing. UNICA data over the next eight weeks determines whether the third ingredient arrives.

The base rate from COT extremes of this severity: +20-30% within 90 days, in every case. The risk is that this time the surplus really is large enough to absorb the ethanol diversion without flipping to deficit, which would make the move a positioning-driven 14-to-17 trade rather than a fundamental 14-to-22 trade. Both are profitable on our structure. The question is whether we add.

Wheat: The Demand Cascade

Wheat is the opposite setup from corn. As we’ll see, corn is supply destruction through the fertilizer gate. Wheat is demand explosion with supply already damaged.

The crop was broken before a single ship was blocked.

Winter wheat opened the season at 35% good-to-excellent. Worst since 2022. Oklahoma is 54% poor-to-very-poor. Texas 51%. Total planted acres are the smallest since 1919. That’s a lot. One hundred and seven years. The crop was stressed by drought through the Southern Plains growing season, and the acreage just isn’t there because wheat economics have been terrible for years. Farmers have been rotating out steadily.

Hormuz adds the demand side on top.

Mid East food security buying is a political imperative, not a discretionary purchase. Egypt, Algeria, Iraq don’t cut wheat imports because prices rose 20%. They panic-buy. Bread prices in Cairo and Algiers are a pillar that keeps the regime standing. The Arab Spring started with bread. Every MENA procurement officer knows this. When the strait closed, the first thing they did was to pull forward six months of tenders.

The KC-Chi spread is an interesting real-time monitor. KC is hard red winter — protein wheat, what you bake bread with. Chicago SRW is feed wheat. When MENA bids, they bid KC, not Chicago. The spread widens. Right now it’s about 35 cents, which is where it sits in quiet markets. In 2022 during the Ukraine spike, the ratio hit 35-40% premium. In 2007-08 it was 15-20%. We’re at 5-10% today. The MENA protein bid has barely started showing up.

As long as the spread is widening, the bid is accelerating. If it compresses, the panic is fading. One number to watch. Could be a nice way to play this, come to think of it.

The balance sheet

Baseline (pre-crisis). Production 1,803 million bushels at 51.5 bu/ac on 35 million harvested acres. Beginning stocks 794M. Ending stocks 792M at 41.1% stocks-to-use. Imports 120M. Exports 850M. Feed and residual 100M. The balance sheet looks comfortable at first glance. 41% stocks-to-use is a lot of wheat. But that comfort is deceptive because the demand side can move faster than the supply side can respond. Winter wheat is in the ground. Spring wheat is planted. You can’t grow more acres in July because Egypt is buying.

Closed 1 month. Supply unchanged. Winter wheat in the ground, spring wheat seeded, nothing moves on a 30-day timeline. Exports tick up 20M bushels as MENA buyers start front-running their tenders. Russia loses half a million metric tonnes of export share, not because Russian wheat is physically blocked (it ships through the Black Sea and Turkish Straits, not Hormuz) but because importers start diversifying sourcing for reliability. Insurance costs on Black Sea cargoes rise. Payment friction increases. Ending stocks: 772M. Stocks-to-use: 39.7%. SRW to $6.00, KC HRW to $6.40. KC-Chi spread widens from 35 to 40 cents. Noise.

Closed 3 months. Both sides break open simultaneously.

Supply: yield drops to 50 bu/ac as fertilizer costs hit spring wheat top-dress applications. Production falls 53M bushels. Not catastrophic. The yield hit is smaller than corn because winter wheat was already fertilized before the blockade. The damage here is concentrated in spring wheat regions.

Demand surges 110M bushels. MENA food security buying goes from front-running to panic. Russia loses 3 MMT of export share as insurance costs and payment friction compound. Feed substitution kicks in, as corn gets expensive, livestock operations blend more wheat into rations, adding 10M bushels of feed demand that wasn’t in anyone’s baseline.

Ending stocks: 629M. Stocks-to-use: 30.9%. SRW to $6.80, KC HRW to $7.30. KC-Chi spread widens to 50 cents. The spread widening is the signal that the market is differentiating between feed wheat and food wheat. The MENA bid concentrates in hard red winter because they need protein wheat for bread.

Closed 12 months. Yield drops to 47 bu/ac. Acres decline 1.5M as spring wheat farmers pivot to oilseeds (canola and soybeans fix their own nitrogen). Imports fall, Canada and Australia prioritize their own balance sheets. Production: 1,575M bushels, down 229M from baseline.

Demand: 2,160M bushels. Russia has lost 8 MMT of export share. MENA pulls an additional 200M bushels. Feed substitution at 30M bushels. Ending stocks: 309M. Stocks-to-use: 14.3%.

That number matters. 14.3% is below the 2007-08 level that preceded the Arab Spring.

KC HRW to $9.00. KC-Chi spread at 80 cents. Protein wheat rationed by price.

Positioning

Managed money is net long 19k contracts and declining, they reduced by 5k this week. Other reportables are pressing short at -15k, also adding. Swap dealers are steady at +71k.

This is not a squeeze setup. Managed money isn’t extreme long, specs aren’t piling in but have bought recently. The trade hasn’t been front-run. That’s the kind of setup where a fundamental repricing catches people on the wrong side gradually rather than in a spectacular short squeeze.

The vol work

This is where we spent the most time, and it changed the trade.

KC HRW Dec-26 ATM implied vol is 26.9%. Three-month realized is 27.3%. The premium is 31 basis points. Basically nothing. ATM options are fairly priced.

But the call skew is steep. 25-delta calls at 28%. 10-delta at 34%. 5-delta at 40%+. The further out of the money you go, the more the market charges. By the time you’re buying the kind of OTM call that gives you 5:1 or 10:1 payoff, you’re paying 6-12 vol points above ATM. That premium eats the convexity you’re buying.

We pulled the Rose AI 220-year wheat return history to check the base rate. A 17% rally in 4 months (what the Aug/Sep 700 calls need) happens about 14% of the time unconditionally across two centuries of data. The Aug/Sep vol surface at the 700 strike is pricing that probability at 32%. That’s 2.3 times the historical base rate. Over 7 months (the Nov/Dec timeframe), the historical base rate for a 17% move rises to 21%. The Nov/Dec surface prices it at 28%. Still 1.3 times above history, but less extreme.

The pattern across all three expiries:

Jun/Jul (75 days): 3.2x historical markup at the 17% move threshold. Aug/Sep (131 days): 1.8x. Nov/Dec (222 days): 1.3x.

The market is frontloading the event premium into the near-dated options. Which makes sense, the catalysts are front-loaded, so near-term vol should be bid. But it means every OTM call on the board is expensive relative to where the historical distribution says it should be priced. The market already sees the blockade, the crop stress, the MENA bid. There’s no free lunch here.

The contango adds another layer. Dec futures at 647 vs front month at 601. That’s 7.7% of carry already baked in. A Dec 660 call that looks 2% out of the money from December futures is really 10% above spot. If the blockade resolves, Dec doesn’t just drop the crisis premium, the contango collapses too. You’d be paying for peace risk embedded in the forward curve on top of the skew premium embedded in the vol surface.

So we looked at it from every angle. Outrights at each expiry, call spreads across the skew, the capital efficiency trade-off of cheap OTM versus expensive vol. In every case the vol surface was pricing the tail we wanted to own at or above fair value. ATM was the only part of the surface where implied was at or below realized, and an ATM call at these levels doesn't give the risk/reward ratio we look for on the book.

We sold our wheat options.

The decision is simple once you frame it correctly. There’s enough convexity in the underlying itself. Wheat at $6.00 going to $7.30-9.00 in the scenarios is 20-50% upside. You don’t need option convexity to capture that. Futures give you full delta participation without bleeding theta or overpaying for skew.

The trade

Winter wheat futures. KC HRW front month.

No calls. The vol surface was telling us not to pay, and we listened.

The honest trade-off: we gave up defined risk and gained full delta at zero vol premium. If the thesis is wrong, futures hurt more than expired options. If it’s right, we capture the full move without paying the market’s inflated fear premium. Size accordingly. Risk management on futures is about stop discipline, not defined payoff.

What to watch

April 14: Crop Progress. First condition ratings since the blockade. Already bad, but watch whether the Southern Plains deteriorated further during March. Any drop below 30% G/E is a trigger for MENA acceleration.

May 9: WASDE publishes the first 2026/27 new-crop balance sheet. USDA will likely use 95.3M acres and trend yield. The balance sheet will look comfortable. If they print 51.5 bu/ac yield on a 35% G/E crop while the strait is closed, the gap between their projection and reality is the trade. WASDE creates the disconnect. The market corrects it.

June 30: Acreage Report. Actual planted acres versus March intentions. In 2022, actual came in 4M acres below March intentions. If something similar happens here, the supply story cracks open.

July WASDE: First yield revision from actual crop condition data. This is the big one. The market discovers what was locked in biologically during April and May. Every week of poor conditions between now and July compounds, irreversibly.

KC-Chi spread: The confirmation signal. Widening = MENA bid accelerating. Compression = panic fading. Watch this number more than the outright price.

Vol surface: 28% implied today, 50%+ is where vol goes during actual supply shocks. If the thesis transmits and you want to add option convexity on top of the delta position, wait for the confirmation and pay the elevated vol then, when the move is underway. Buying vol before the catalyst fires is what the market wants you to do. Don’t.

Corn: Five Gates, One Calendar

Corn has the biggest buffer and the longest fuse. Record production at 17 billion bushels, ending stocks at 2,127 million, quarterly grain stocks at a record 9 billion. The market looks at this and sees comfort. The market is correct, for now.

The cascade changes the math through farmer decisions at specific calendar gates, each irreversible once closed. Pre-plant nitrogen is happening now. Acreage allocation closes this month. Sidedress nitrogen, the binary gate, arrives in June. The corn trade needs the blockade to persist through sidedress AND the July WASDE to revise yields down. Three gates, sequential, uncertain. We hold the calls but conviction has migrated to wheat and sugar, where the catalysts are closer and the mechanisms more direct.

That said, corn is where the deepest analytical work lives, and where the optionality is cheapest relative to what the S/D model produces if you walk the gates forward honestly.

The transmission mechanism

Corn is the most nitrogen-intensive major crop on earth. A bushel of corn requires roughly 1.0 lb of applied nitrogen. At 186 bu/acre trend yield across 95 million acres, that is approximately 17.7 billion pounds of nitrogen, or about 4.4 million tons of anhydrous ammonia equivalent, purchased and applied every spring.

The blockade transmits to corn through a chain of five farmer decisions, not through a single price shock. This matters because each decision has a different sensitivity to nitrogen price, a different calendar window, and a different reversibility profile. The market prices corn as if fertilizer cost is a smooth input. In practice, the transmission is lumpy, sequential, and path-dependent.

The critical variable is the nitrogen-to-corn price ratio, not the absolute price of ammonia. In 2022, ammonia hit $1,600/ton but corn was at $7.50, so the ratio peaked around 0.14-0.16. Farmers adapted. Yields came in 3.5% below trend, mostly attributable to drought. Today ammonia is only $924 but corn is $4.50. The ratio has hit 0.20 for UAN, the highest in five years. The denominator is what makes this crisis worse than 2022 on a farm-economics basis. Farmdoc at Illinois published updated MRTN guidance on March 31 specifically referencing the Iran conflict and recommending rate reductions of 6-8 lbs/acre from fall levels.

What is “MRTN”? Maximum Return to Nitrogen. It’s a calculator built by Midwest land-grant universities that finds the economically optimal nitrogen application rate given two inputs: the price of nitrogen and the price of corn. The yield response to nitrogen follows a diminishing-returns curve. The first 100 lbs of N per acre produces most of the yield gains. The last 30 lbs near maximum yield adds maybe 2-3 bu/acre. The MRTN finds the point where the marginal cost of the next pound of nitrogen exactly equals the marginal revenue from the additional corn it produces.

At today’s ratio of 0.20, the MRTN recommends 172 lbs/acre for corn-after-soybeans (down from 178 last fall) and 186 for continuous corn (down from 194). Those 6-8 lb/acre reductions sound trivial. Across 90+ million acres they compound into hundreds of millions of bushels of yield risk.

If ammonia goes to $1,200-1,500 in the 3-12 month scenarios, the ratio pushes past 0.25-0.30 and recommended rates drop 15-25 lbs/acre. That is where yield losses compound from a rounding error into a balance sheet event.

The baseline

US corn production set a record at 17.02 billion bushels on 98.8 million planted acres at 186.5 bu/acre. Ending stocks at 2,127 million bushels are up 37% year-over-year from 1,551 million. The stocks-to-use ratio is approximately 12.9%. Exports are running at a record pace of 3.3 billion bushels with inspections 42% ahead of last year. Feed and residual use is projected at 6.2 billion, ethanol at 5.5 billion. The season-average farm price sits at $4.10.

The global picture is less comfortable. World ending stocks declined to 293 MMT from 296 MMT. Excluding US and Chinese reserves, global stocks hover near 11-year lows at roughly 82 MMT. The IGC’s initial 2026/27 outlook projects world maize production declining from this year’s record, with cumulative grain stocks falling 23 MMT. The US surplus is real but it masks a tightening rest-of-world picture that leaves less margin for error if production disappoints.

The forward curve confirms the surplus. May at 454 (as of writing), July 465, December 481, March ‘27 at 492. Classic contango reflecting ample nearby supply with carry charges and new-crop uncertainty in deferred months. The contango costs us to hold long positions, which is one reason conviction has migrated to wheat (where we went delta, not calls) and sugar (where the contango is shallower).

Prospective Plantings on March 31 showed corn area intentions at 95.3 million acres, already down 3.5 million from 2025. This was surveyed in early-to-mid March, roughly two weeks into the blockade but before the worst of the fertilizer spike hit retail. Ammonia was maybe $750-800 at survey time. The actual number will be reported June 30. In 2022, actual corn planted came in 4 million acres below March intentions.

The five gates

The cascade doesn’t hit corn production all at once. It passes through five sequential decision points, each with its own price sensitivity and calendar window. Understanding which gates are locked, which are closing, and which remain open is the entire analytical challenge.

Gate 1: Crop choice (March-April, closing now). At a 0.20 N:corn ratio, corn is unprofitable on marginal land. Breakeven on average soil runs $4.90/bu before the fertilizer spike, closer to $5.10-5.50 at current input costs. Fields yielding 140 bu/acre or less cannot make the math work. The soybean pivot is already underway. Prospective Plantings captured some of this, but nitrogen prices have risen another 15-20% since the survey. We estimate 5-8 million acres exit corn beyond the 3.5M already reflected in intentions. At roughly 140 bu/acre average on those marginal fields, that is 700 to 1,120 million bushels of production removed from acreage alone.

The mechanism here is not uniform rate reduction across all fields. Farmers on good land (200 bu/acre potential) still plant corn because the economics work even at $924 ammonia. Farmers on marginal land (140 bu/acre or less) exit entirely to soybeans (which fix their own nitrogen) or fallow. The bad fields disappear. The good fields stay but trim rates. The weighted average yield drops because the bottom of the distribution gets removed.

Gate 2: Pre-plant nitrogen rate (April-May, happening now). Roughly 60% of US corn nitrogen was purchased last fall at $600-680/ton ammonia. Those farmers are unaffected. The remaining 40%, about 35 million acres, are buying spring pre-plant at $924. The MRTN calculator tells them to cut 6-8 lbs/acre from the recommended rate. Some will cut more. At 3-8 bu/acre yield loss across those 35 million acres, that is another 105-280 million bushels of production risk from rate reduction.

The corn planting season unfolds as three sequential S-curves, and the timing matters for understanding which decisions are still open.

Pre-plant nitrogen (amber) is already 40% complete as of early April and will be 85% done by May 1. That horse is leaving the barn. Corn planting (blue) is just starting at 3% and ramps through late May. Sidedress (green) doesn’t move until June and doesn’t peak until mid-June.

In physical quantities, the magnitudes make the sidedress vulnerability concrete. Pre-plant is 4.4 million tons of NH₃, mostly deployed. Corn planting covers 93 million acres. Sidedress is 5.4 million tons of UAN solution that needs to be purchased and applied between late May and early July. At current UAN prices of $464/ton, that is $2.5 billion of farmer purchasing decisions sitting entirely inside the blockade window. Zero of it has been bought yet.

Gate 3: Sidedress, the critical fork (June). This is where the trade lives.

About 25-30% of Corn Belt acres rely on split nitrogen application, putting 30-40% of their total N down as sidedress at V6-V8 stage, typically early-to-mid June. Most sidedress is UAN solution, purchased from dealers in late May for application in the first three weeks of June. The purchase window is May 15 to June 10. The ceasefire expires April 22. May UAN price is the single most important data point for the trade.

If UAN is below $500/ton in late May (ceasefire holds, logistics normalizing), most farmers sidedress normally. Yield potential preserved. The cascade stops at gates 1 and 2. Corn goes to $4.80-5.00 on acreage loss and modest rate cuts. Our Dec 525C probably expires near zero.

If UAN is above $600/ton (blockade persists, logistics disrupted), 60%+ of the sidedress-dependent acres skip the top-up. Twenty-five million acres lose 10-15 bu/acre. That is 250 to 375 million bushels of additional production loss, locked in permanently after V8 around June 20. The corn plant has set its ear size by then. No amount of money or rain changes the outcome after that date.

This is the binary. Everything before it is gradual. Everything after it is biology.

To see how fertilizer cost bends each of the three curves differently, look at what happens to acres, pre-plant N, and sidedress N at three price levels: normal ($500 NH₃), today ($924), and 12-month blockade ($1,400).

The top panel (acres) barely spreads. Even at $1,400 ammonia, 80 million acres still plant corn because the good land is still profitable. You lose 13 million acres off the bottom. The middle panel (pre-plant N) spreads modestly because most of it was purchased before the blockade. The bottom panel is where it blows apart. At normal prices the sidedress line climbs to 5.4 million tons. At today’s prices it reaches 3.5 million. At $1,400 it collapses to 1.2 million. That is 4.2 million tons of UAN that never gets applied. That is the corn that turns pale green in July.

A note on what happens after June. Once sidedress closes, the nitrogen price ceases to matter for the current US crop. July is pollination, which is weather-driven. August through October is crop condition reports revealing damage that was locked in during April-June. The market discovers what biology already determined. The June-September ramp in global production loss is driven primarily by other regions (India’s kharif corn begins planting in June, SSA continues bleeding, Argentina’s next crop gets exposed at 12 months), plus the revelation of US damage through WASDE revisions. The US corn line actually flattens at -28 to -30 MMT from August onward. The continued global ramp is everyone else.

Gate 4: Harvest or abandon (September-November). Diesel at $6-7/gallon. Under-fertilized fields that yielded 120 bu/acre on marginal land may not be worth combining. Harvest costs run $35-45/acre at normal diesel, $50-65/acre at current levels. We estimate 2-5 million acres abandoned at the margins, though this gate has a self-correcting feedback: if corn prices have risen to $5.50+ by harvest, even weak fields become worth cutting. The price signal partially rescues this gate. Impact: 240-600 million bushels, partially offset by higher corn prices.

Gate 5: Ethanol demand pull (ongoing). This is the demand-side accelerant that makes the supply-side cascade worse. At $95+ crude, ethanol margins blow out. US plants running at 94% utilization push toward 98%. At full capacity, that is roughly 300 million additional bushels of corn demand. If E15 goes nationwide under political pressure from high gasoline prices, the ceiling is 500 million bushels. The EPA’s record-high 2026 RVO of 26.81 billion gallons creates a floor that does not flex with price.

The ethanol bid does not create corn scarcity by itself. But it tightens the balance at exactly the moment supply is contracting from gates 1-4.

Historical precedent

The yield response to fertilizer withdrawal follows a consistent hierarchy based on where an agricultural system starts on the nitrogen response curve. The US and EU sit on the flat part of the curve at 150+ kg N/ha. They can trim 20-30 lbs per acre and lose 3-8 bu/ac. Sub-Saharan Africa sits on the steep part at under 20 kg N/ha. Any reduction comes straight off yield because every pound of N they apply is still on the high-return portion of the curve.

In 2022, the global fertilizer spike produced a 16% cereal production decline in East Africa but only 3.5% below-trend corn yield in the US. In Kenya, a 100% increase in fertilizer prices produced a 37-38% decline in maize production. Sri Lanka’s complete fertilizer ban in 2021 collapsed rice yields 20-32% and tea yields 27% in a single season, toppling the government. Zimbabwe’s currency collapse through the 2000s cut agricultural output 51% and tobacco 79% as farmers lost access to imported fertilizer.

The key difference between 2022 and today: in 2022, corn was at $7.50, so the N:corn ratio stayed at 0.14-0.16 despite ammonia at $1,600. Farmers could afford the nitrogen. Today corn is at $4.50 with ammonia at $924. The ratio is worse at a lower absolute nitrogen price.

The global asymmetry is why the food security cascade generates MENA wheat panic buying. Developing countries cannot absorb the nitrogen shock. They bid for imports. That tightens global stocks and lifts prices for everyone. The US loses 7% of cereal yield. SSA loses 35%. Same shock, different position on the curve.

Scenario walk

Closed one month (through mid-May). Pre-plant nitrogen largely applied at elevated but manageable prices. Acreage loss limited to the 3.5-5M acres already in the Prospective Plantings data plus perhaps another 1-2M that shifted after the survey. Sidedress window still open. If ceasefire holds and nitrogen falls below $800 by late May, sidedress proceeds normally.

Production impact: 16.3-16.5 billion bushels (vs 17.0 baseline). The loss is 500-700M bushels, almost entirely from acreage exit at the margins. Ending stocks: 1,700-1,900 million bushels. Still above the 1,551M that supported $4.20 last year. Corn: $4.60-4.80. Our calls bleed theta.

Closed three months (through mid-July). The sidedress fork goes the wrong way. UAN above $600 in June. Fifteen to twenty million acres under-fertilized at sidedress. Pre-plant rate cuts locked in across 35M acres. Total acreage down 5-8M from baseline.

Production impact: 14.8-15.5 billion bushels. The yield hit reveals itself through crop condition reports starting late May, then WASDE revises in July. Ending stocks: 1,200-1,700 million bushels. At the low end, that is approaching the 2020/21 stocks-to-use of ~8% that supported $5.50+ corn. The contango flattens, potentially inverts if WASDE revises below 1,400M.

Ethanol adds 150-300 million bushels of incremental demand. Corn: $5.00-5.50 by September.

Closed twelve months (through April 2027). Full cascade plus second-order effects. Total corn acreage down 8-13M from baseline as 2027 planting decisions are made under persistent scarcity. Yield down 5-10% from trend on the acres that remain. Ethanol demand up 300-500M bushels at sustained $95+ oil.

Production impact: 13.0-14.5 billion bushels. Ending stocks: 600-1,000 million bushels. At 880M, that is a stocks-to-use ratio under 6%. The last time corn stocks were this tight relative to demand was 2012, when prices exceeded $8.00/bu.

This is the tail. The options we own are buying this tail.

The global picture compounds the US story. At three months, total world corn production loss reaches roughly 18 MMT against 293 MMT of global ending stocks. Manageable. By six months, the loss reaches 65 MMT as the US damage ramps through pollination and harvest, SSA continues bleeding, and India’s kharif season is fully exposed. By twelve months, 112 MMT is gone. Global ending stocks drop from 293 to 181 MMT. For context, the 200 MMT level on the chart below is where global stocks sat during the 2012 drought crisis.

Positioning and vol

Speculative funds are net short corn. This is the consensus position for a record-supply environment and the market is not wrong about the current balance. What the positioning tells us is that the repricing, when it comes, will be violent. Spec shorts covering into a market that is simultaneously receiving bearish-to-bullish WASDE revisions is how you get 20-30% moves in weeks rather than months.

Corn ATM implied vol sits around 24.4%, roughly in line with 3-month realized. There is no premium being charged for the blockade scenario. Compare this to wheat where IV has expanded to 26-30% across the surface. The options market has decided corn’s surplus protects it. We disagree for the reasons above, specifically around gates 3-5 which have not been priced.

Dec corn (Z6) at 481 implies the market expects modest new-crop tightening but nothing dramatic. The 525 strike sits about 9% out of the money with 8 months to expiry. In a normal year this is a lottery ticket. In a year where 21 million tonnes of nitrogen capacity is physically offline and the sidedress window opens in two months, it is priced too cheap for the scenario it captures.

The vol work tells us something else. Corn realized vol has historically spiked to 35-50% during genuine supply shocks (2012 drought, 2020 demand shock, 2022 Ukraine). The current 24% implied assumes none of that. If the WASDE revises yield down 5+ bu/acre in July and realized vol expands to 30-35%, the Dec 525C reprices significantly on vega alone before delta kicks in.

We have half the corn allocation in vol (straddles and strangles closer to the money) and half in upside (the Dec 525C and some Dec 550C further out). The vol position profits if corn moves in either direction. The upside position needs gates 3-5 to trigger.

The trade

We hold Dec corn 525 calls. They are underwater. We are early. The thesis requires patience through gates 1-2 (where the market sees the acreage data but not the yield implications) and then confirmation at gate 3 (sidedress).

At current IV, the Dec 525C costs roughly 12-15 cents (as of writing, it has probably moved!). If the full cascade triggers and corn reaches $5.80-6.20, the 525C is worth $55-95 in intrinsic value alone. The risk/reward is asymmetric by construction.

What could go wrong: if the ceasefire holds and nitrogen prices normalize by June, sidedress proceeds, yield loss is limited to 3-5 bu/acre from pre-plant rate cuts, and ending stocks land at 1,700M. Corn trades to $4.80-5.00. Our calls expire worthless but the loss is capped at premium paid, and the SONIA position should offset on the peace scenario.

This is the inherent convexity of the cascade structure. Even in the most optimistic scenario, Hormuz will take time to normalize. Energy prices will remain elevated for months, not weeks. The flow-through to fertilizer prices has already happened and will not reverse on a handshake. High diesel at $5.98 leads to abandoned marginal fields. Ethanol economics pull corn off the food market to supplement refined products. We want to continue owning these even if we get peace in the next couple of weeks.

What to watch

Today, April 14: Crop Progress report (4pm ET). Corn 3% planted last week. First reading with a full week of planting activity since the blockade announcement. Watch planting pace vs five-year average and any early state-level divergence.

April 22: Ceasefire expiration. The most important binary for the book. If the blockade continues, the probability of UAN above $600 in May rises sharply and gate 3 tilts toward skip.

May 9: May WASDE. First report with USDA’s official 2026/27 new-crop balance sheet. They will use Prospective Plantings acres (95.3M) times trend yield (~181). The market will compare that to its own acreage estimates. If traders think actual acres are 91-93M, there is an immediate disconnect to trade against.

Weekly through May: USDA AMS fertilizer prices (Thursdays). Track UAN28 retail specifically. The sidedress purchase window opens in five weeks. The trend line from now to then determines gate 3.

May 19: Crop Progress. By this date, 55-70% of corn should be planted in a normal year. If the pace runs 5+ points behind the five-year average, acreage loss is real and the June 30 Acreage Report becomes a potential catalyst.

Late May: UAN spot price. The trigger. If UAN is below $500, sidedress happens. If above $600, the skip probability jumps to 60%+.

June 20: V8 cutoff. After this date, sidedress nitrogen does nothing. The yield is biologically locked. Whatever was applied is what the crop has to work with through pollination.

July WASDE: First yield revision. This is when the market discovers what was locked in during April-June. If USDA revises yield below 180 bu/acre (vs 186.5 trend), the repricing is immediate and the Dec 525C starts working.

The damage is locking in now, on biological and chemical timelines the market does not track. The repricing comes later.

Conclusion

The cascade runs on calendars and chemistry. The rimland strategy handles crude and the dollar. It does not handle what crude feeds, what fertilizer grows, or what happens when 21 million tonnes of nitrogen capacity disappears during planting season.

Sugar is ticking now. The crush started April 1. Every UNICA report either confirms or denies the ethanol diversion. We hold call spreads with 6.8:1 payoff and $213 max loss. The data arrives in weeks.

Wheat was ticking before the blockade started. Smallest crop since 1919, worst conditions since 2022, and a MENA demand bid that compounds with every week the strait stays closed. We bought futures because the vol surface was charging too much for options. The KC-Chi spread tells us whether the bid is accelerating.

Corn is the longest fuse and the biggest payload. The surplus buffer is real. The options are cheap because the market believes the buffer protects it. We think the sidedress gate in June determines whether the buffer holds. May UAN spot is the single most important data point in the cascade.

The ceasefire expires April 22. The WASDE publishes May 9. The sidedress window opens in June. The calendar does not wait for diplomacy.

We remain long the cascade.

Disclaimers

This analysis was built over several days of active markets. Prices, curves, vol surfaces, and positioning data reflect the time they were pulled, not the time you're reading this. Some of these levels will have moved by the time you see them. Do your own work. Check the chain before you trade. The price on your screen when you lift the offer may look nothing like the price on mine when I wrote this up.

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counter parties.

The Ramble does take positions in various assets we write about, and views are subject to change without notice.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure. Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

The tension we laid out last Wednesday has proved irreconcilable.

Iran wants nukes and control over the Strait. Trump cannot allow either. The two circles of the Venn diagram sit far enough apart that Israel’s war on Lebanon didn’t even make the agenda.

I won’t claim I called this perfectly, but we may finally be in the mid-game. This isn’t something that gets called off in an afternoon. The question is fundamental: who controls the most important waterway in the world, and does Iran’s willingness to threaten its neighbors give it enough leverage to negotiate its way to a nuclear weapon?

That’s the crux.

What’s becoming clear is the strategy. And for those who’ve been reading along, from “War for the Dollar“ through “Don’t Take the Bait“ through “Waking the Hegemon“ and “The Fragile Peace,” the pattern should be recognizable by now.

Trump is running the Rimland playbook.

Interdict the ships. Threaten 50% tariffs on anyone supplying weapons to Iran. Bring China into the conversation not by invading the heartland, but by controlling the water their energy flows through. For every ship they mine, for every tanker they hit, seize a ten of theirs. Park the carriers, commandeer the tankers, sell their crude. In dollars.

And then the Abraham Accords. Saudi oil flowing through Jordan to Haifa. The Tapline reborn. A physical corridor connecting the coastal states into an energy network that bypasses the heartland entirely. Rimland coalition infrastructure built in pipe and steel. By my account, the reason we are in this mess (Iran/China acting via Hamas on October 7th to trigger Israel and derail the normalization of relations which would then provide alternative routes for trade outside of Hormuz/BRI).

This helps explain the gap between Washington and Brussels. The US feels the weight of responsibility. Europe appears to believe it can cut side deals for its energy while the younger brother fights its battles. France blocked the UNSC resolution, negotiated bilateral passage through the Strait, and called for a “coalition of independents.” Classic heartland thinking: cut a deal with the interior, avoid the fight, pretend the sea lanes maintain themselves.

Trump just closed that loophole. And in doing so, made America’s problem the world’s.

As of writing, crude is back up 6%+, stocks are down 1%, and the ceasefire gains from last week look extremely likely to unravel. I bot some VIX calls late last week, so call me biased.

How much depends on a series of first-order questions about how far this escalation goes:

• Can the ceasefire hold for another week, or does backward induction lead to resumption?

• Trump has promised to interdict ships that paid Iranian tolls. Does that include Chinese vessels? What happens when they try to pick up crude from Kharg?

• He’s also reiterated the 50% tariff threat on any country supplying weapons to Iran. Is the trade war back on the menu?

Then there’s the counter-move. Iran activates the Houthis, who still appear capable of making Bab el-Mandeb impassable. Worth noting: the majority of tankers lifting Saudi barrels from the East-West pipeline are VLCCs too large for Suez. A Houthi escalation doesn’t just threaten Red Sea traffic. It forces the long route on the biggest ships carrying the most critical barrels.

The through-line is that this conflict continues to grow in scale and scope. By widening to a total interdiction of toll-paying ships and reiterating tariffs, Trump has brought China squarely into the conversation. Beijing has been stockpiling crude for years for exactly this scenario. But given the economic drag from the property hangover, how long can Chinese markets stay sanguine? How likely are they to escalate to defend their energy supplies? The sequencing — Venezuela, then Iran — looks increasingly strategic.

Rimland returns.

The second-order market questions follow directly:

• How bad is Monday morning? The first leg down was fast money and retail put buying. At what point do real money investors decide the vol is unmanageable and puke, or hit risk limits?

• Last week saw hedge fund pods rush back into long AI hardware, short software. As oil rallies, bonds sell off, tightening liquidity. Combined with supply chain risks around Gulf helium, a critical input in chip fabrication. Is this enough to force a reprice of the AI acceleration timeline?

• US growth was basically flat in Q1 going into the conflict. As energy prices surge, disposable income gets pulled into gasoline, heating, jet fuel. Will households cut spending or lever up?

• The Fed minutes revealed the board is already discussing tightening to contain the inflationary impact of higher energy. Expect a fresh round of debate about the appropriate response to a negative supply shock. Can the Fed “look through” an energy shock of this magnitude?

All of which brings us to the cascade. The rimland strategy addresses energy and the dollar. It does not address what energy feeds. The market is pricing the first node. It has not yet connected the second. Oil reprices on headlines. Agricultural calendars don’t. Urea is still at $700. The USDA’s smallest planned wheat crop since 1919 doesn’t reverse because two diplomats shook hands. Farmers who couldn’t afford fertilizer in March aren’t retroactively applying it in April.

Tomorrow we go deep on ags. The supply and demand picture, what’s already baked in, and where the market is still mispricing the cascade.

Disclaimers

Charts and graphs included in these materials are intended for educational purposes only and should not function as the sole basis for any investment decision.

These opinions are mine and mine alone and do not represent the views of any of Rose’s clients or counter parties.

The Ramble does take positions in various assets we write about, and views are subject to change without notice.

This letter does not constitute an offer to sell, a solicitation of an offer to buy, or a recommendation of any security or any other product or service by Rose, Campbell or any other third party regardless of whether such security, product or service is referenced in this brochure. Furthermore, nothing in this website is intended to provide tax, legal, or investment advice and nothing in this website should be construed as a recommendation to buy, sell, or hold any investment or security or to engage in any investment strategy or transaction. You are solely responsible for determining whether any investment, investment strategy, security or related transaction is appropriate for you based on your personal investment objectives, financial circumstances and risk tolerance. You should consult your business advisor, attorney, or tax and accounting advisor regarding your specific business, legal or tax situation.

THERE CAN BE NO ASSURANCE THAT ROSE TECHNOLOGY INVESTMENT OBJECTIVES WILL BE ACHIEVED OR THE INVESTMENT STRATEGIES WILL BE SUCCESSFUL. PAST RESULTS ARE NOT NECESSARILY INDICATIVE OF FUTURE RESULTS. AN INVESTMENT IN A FUND MANAGED BY ROSE INVOLVES A HIGH DEGREE OF RISK, INCLUDING THE RISK THAT THE ENTIRE AMOUNT INVESTED IS LOST. INTERESTED PROSPECTS MUST REFER TO A FUND’S CONFIDENTIAL OFFERING MEMORANDUM FOR A DISCUSSION OF ‘CERTAIN RISK FACTORS’ AND OTHER IMPORTANT INFORMATION

He was not a lone wolf. He was an active member of PauseAI with six community roles. His Discord handle was "Butlerian Jihadist." His Instagram was a feed of doomer content: capability curves captioned "if we do nothing very soon we will die," Venn diagrams placing us at the intersection of The Matrix, Terminator, and Idiocracy. Four months before the attack, he recommended Yudkowsky and Soares' If Anyone Builds It, Everyone Dies to his followers.

His name is Daniel Moreno-Gama.

He had his own Substack. In January he published “AI Existential Risk,” estimating the probability of AI-caused extinction as “nearly certain.” He called the technology “an active threat against anyone who is using it and especially towards the people building it.” He concluded: “We must deal with the threat first and ask questions later.” He wrote a poem imagining the children of AI developers dying, asking their parents why they did nothing. “May Hell be kind to such a vile creature,” he wrote about the builders.

PauseAI has already deleted his messages from their Discord.

For an investing newsletter, I know this is not what most of you are here for. The goal here is to to explain where my worldview is coming from, so that the longer term calls start to make more sense. My ideas behind the “New New Deal” are intended as a direct response to where this is going.

All I am doing here is running their model forward, and connecting the dots.

Here is the framework. It has three moving parts.

Start with certainty. Yudkowsky’s position is that if anyone builds sufficiently intelligent AI, every human being on earth dies. Not probably. Not maybe. Everyone. Your children. His daughter Nina, whom he invokes by name. He published this in TIME. He wrote it in a book called If Anyone Builds It, Everyone Dies. He said we should airstrike data centers, and that the risk of nuclear exchange is preferable to a training run completing.

Purity spiral aka escalation. Within this community, members compete to demonstrate commitment by raising the stakes. P(doom) numbers climb from 50% to 90% to 99.99999%. The Center for AI Safety's national spokesperson said on camera that the correct response is to "walk to the labs across the country and burn them down." PauseAI activated something called a "Warning Shot Protocol" declaring an AI model "a weapon of mass destruction." One of PauseAI's leaders said an Anthropic researcher "deserves whatever is coming to her." When someone flagged this rhetoric in PauseAI's Discord, the mods deleted the post.

The day before the attack, Nate Soares, Yudkowsky's co-author on the very book the kid recommended, tweeted that Altman was "doing terrible stuff."

Then cheap talk gets tested. Game theorists have a term for this: cheap talk is costless signaling that eventually meets reality. When you make the stakes existential for the human race, you can justify any level of extremism if it lowers the hallowed p(doom). These aren't isolated incidents. They are a series of escalating and mutually reinforcing claims around an eschatological philosophy that, taken to its conclusion, would accept killing 99% of the world to save the last 1%.

It was only a matter of time before someone took the framework at face value. The kid read the book. He joined the community. He wrote his own manifesto. In a memoir for his community college English class, he described himself as a consequentialist: “I give very little credence to intentions if the results do not match.” He chose “Butlerian Jihadist” as his name. On December 3rd he wrote in PauseAI’s Discord: “We are close to midnight it’s time to actually act.”

Then he acted.

They gave him a trolley problem. One life versus all of humanity. The kid pulled the lever.

There is a final irony that deserves attention. If the doomers truly hold their stated beliefs at their stated confidence levels, they should be more honest about what those beliefs imply. A few weeks before the attack, a journalist asked Yudkowsky: if AI is so dangerous, why aren't you attacking data centers? His answer, relayed by Soares: "If you saw a headline saying I'd done that, would you say, 'wow, AI has been stopped, we're safe'? If not, you already know it wouldn't be effective."

Notice what that answer is not. It is not “because violence is wrong.” It is “because it wouldn’t work yet.” The restraint is strategic, not moral. And the community knows it. The dark undercurrent is an unspoken agreement: the kid’s greatest sin was bad timing.

This is what I mean by intelligence not equaling power, and it is the deepest flaw in the entire doomer worldview.

Yudkowsky’s framework rests on the conflation: a sufficiently intelligent AI will necessarily acquire the power to destroy humanity because intelligence converts automatically into capability. Most of his followers are not technical. They do not build AI systems or work on alignment engineering. They possess a particular kind of verbal intelligence that lets them construct elaborate arguments about risk, and they have convinced themselves this entitles them to a priestly authority over the technology. They can construct the argument. They cannot build the system.

This isn’t accidental. It’s baked into the foundational texts. Yudkowsky’s Harry Potter and the Methods of Rationality literally models a world where the person who reasons best deserves to override every institution around him. The Sequences build the liturgy: a small caste of correct thinkers, epistemically and morally superior, whose rationality entitles them to govern what the rest of humanity is allowed to build. It’s not a safety movement. It’s a priesthood with an origin story written in fanfiction.

Yudkowsky can distance himself from the kid with the Molotov. But he cannot distance himself from the syllogism. If the builders are going to kill everyone, stopping the builders is self-defense. That is the central claim, stated plainly. The only question was always when someone would take it at face value.

They should stop acting surprised when their own logic shows up at 3:45 AM with a bottle full of gasoline.

Disclaimers

I am not advocating for or against any position on AI safety. I am observing that a framework built on certainty of extinction produces predictable consequences. The suspect is innocent until proven guilty.

These views do not represent those of any investors, clients or affiliates of Rose.