№ 029WeeklyJun 7 20269 min read

AI CapexMacro

Capital Has Somewhere Else to Be

Bonds, IPOs, and AI are all sucking money out of Crypto

By Akram Ayyash

The Signal

Two markets, one risk environment, opposite outcomes. On the one hand you have the equity market: the S&P 500 closed above 7,600 for the first time on June 1, and the Nasdaq printed its own record. On the other, you have digital assets: Bitcoin spent the same week sliding under $63,000, its lowest since February and roughly 51% below its October high near $126,200. Ether hit a 14-month low. Roughly $3 billion in leveraged positions were liquidated over two days. Spot bitcoin ETFs have now seen outflows for 13 straight sessions.

But then on Friday, a strong jobs report put the first real dent in equities too, it was a bright red day, all assets sold off.

There are plenty of reasons crypto sold off last week. The main one: the marginal dollar finally has somewhere better to be, and crypto is what gets sold to free it up.

The Setup

Start with the thing crypto is losing the bid to.

US equities are at all-time highs and feel stretched at them. Into that, the market is preparing to absorb three of the largest IPOs in history, nearly at once: SpaceX, OpenAI, and Anthropic. SpaceX alone is targeting something close to a $2 trillion valuation, its S-1 already public and its roadshow underway. The three together carry targeted valuations near $3.6 trillion, with combined cash to be raised running north of $200 billion. Goldman puts total 2026 IPO proceeds around $160 billion, roughly four times last year.

And it is not only IPOs. The incumbents are tapping the same well. Alphabet raised equity this week, and Meta is weighing a raise of tens of billions to fund its own AI build-out. Add it together and the equity market is being asked to absorb new supply on a scale that competes with every existing risk position for the same pool of dollars.

That is a withdrawal. Every dollar that funds an OpenAI allocation comes out of some other risk position, and the position that gets sold first is the one the holder is least attached to. The question allocators are actually asking is whether a supply event this size marks the top of the cycle rather than feeds it. You do not need an answer to see the setup. You only need to know which assets are the funding source.

My read is that crypto has been the funding source all year.

Crypto Was Already the Weak Link

This was not a one-week event. Bitcoin has spent the year giving back ground while equities made new highs, and last week was the last chapter of that story, not the first. The tape underneath is worse than the headline. Ether traded to a 14-month low. Solana broke $70. XRP slid under $1.15. As of this week, the total crypto market capitalization sits near $2.2 trillion, a long way from where the cycle's optimism had it going.

The Saylor Sale

And then there is Michael Saylor's Strategy. His company last week sold 32 bitcoin. About $2.5 million worth. For a company that holds more than 843,000 BTC, that is a rounding error on a rounding error, and Saylor called it "inconsequential." Mechanically, he is right.

But the market did not like it. The man who built a company, and a following, on the promise never to sell, was selling. He was doing it to service the dividend on his preferred stock, an obligation of roughly $1.5 billion a year. The pledge met the coupon, and the coupon won.

Strategy's machine runs on a premium. For crypto-native readers, the relevant number is mNAV, the ratio of the company's market value to the bitcoin it holds. When that ratio sits well above 1, Strategy can issue stock above the value of its coins, buy more bitcoin with the proceeds, and leave existing holders with more BTC per share. Issue, buy, repeat. That premium has now collapsed from roughly 3.9x in late 2024 to about 1.22x as of early June. At parity, the engine stalls, and the dividend has to be funded by selling the asset instead of issuing against it.

Reflexivity runs in both directions, and the market is only pricing one of them. If Saylor sells a little more bitcoin and rebuilds the cash buffer, the preferred stock trades back toward par. A preferred back at par is a preferred he can issue against again, and issuance funds bitcoin buying again. The forced seller and the renewed buyer are the same person, on opposite sides of the cash balance. I think the sign on the whole thing depends on whether the buffer gets rebuilt. That is not a capitulation story or a forced liquidation story. It is a balance-sheet story.

The Cascade

When the most-watched holder in the asset class blinks, the leverage underneath everything gets tested. Forced selling fed forced selling, the standard mechanism: prices fall, positions cross margin thresholds, the system sells them, prices fall again.

The names that had stayed above the fray went last and hardest. Hyperliquid, NEAR, and Zcash had each outrun the market into early June, and each gave it back as the trader most associated with the trade rotated out of them. The "immune" cohort turned out to be one crowded book, not a structural exception.

Zcash is the cleanest illustration. ZEC had been one of the year's standout performers. Then a researcher disclosed a soundness bug in its Orchard shielded pool, a flaw that in principle could mint counterfeit ZEC without detection. Developers shipped an emergency fix within days. The precision matters: this was a disclosed vulnerability, not a confirmed exploit, and because of the privacy design there is no way to prove whether it was ever used. The price did not wait for the distinction. Most of the outperformance was gone in days.

So after a week like that, where does the money actually want to be?

Where the Money Wants to Be

It has options it did not have a year ago.

It has three IPOs of genuinely consequential companies. And it has the ten-year Treasury at roughly 4.5% as of Friday, a real, liquid, risk-free yield that competes directly with every speculative allocation in a book.

The rate backdrop is the quiet driver, and it has traveled a long way in a short time. Markets entered 2026 pricing roughly two rate cuts. After April CPI printed 3.8%, the hottest reading in nearly three years, that conversation did not just close. It inverted. And the print is substantially an oil story: the war that has effectively closed the Strait of Hormuz since late February drove Brent up more than 45% and produced what the IEA has called the largest disruption in the history of the oil market. Energy is doing the work, and the Fed cannot cut into it.

So the market repriced. By the December meeting, CME fed funds futures now price roughly a 70% chance the policy rate sits above today's level, with a single 25 basis point hike the most likely single outcome and a cut down near 1%, as of June 5. A month earlier, the same meeting priced a hold as the base case. The Fed sits on hold at 3.50% to 3.75% today and keeps flagging upside inflation risk, but the market has already moved to pricing the next move as up, not down.

When the risk-free rate might be heading up rather than down, cash and short duration stop being a parking lot and start being a position. Fixed income becomes its own liquidity sink, and it pulls hardest from the far end of the risk curve. Crypto sits at the very end of it.

The way I see it, this is the capital rotation happening now: AI for the upside, Treasuries for the floor, and crypto as the funding source for both.

The Other Side of the Rotation

There is a complication in the destination.

The AI complex absorbing this capital is spending at a rate that is starting to strain the balance sheets behind it. Hyperscaler capex is tracking toward roughly $700 billion in 2026, and the free cash flow that funds it is going the other way. Combined free cash flow at Amazon, Alphabet, Microsoft, and Meta is projected to fall toward around $4 billion in the third quarter, down from a post-pandemic quarterly average near $45 billion, with at least one of them on track to run free cash flow negative for the year. Evercore has called it a red-flag moment, the point at which the largest companies in the index spend enough to run cash-flow negative to feed the build-out.

The point for this note is not whether AI is a bubble. It is concentration. The money leaving crypto is rotating into the part of the market with the most operating leverage to a single capex thesis, at the same index weight that makes those names the market. If that thesis wobbles, the side of the rotation that looks safe has a tail of its own, and it is not a small one.

In short: the marginal dollar found better homes (record-high equities, three trillion-dollar IPOs, and a risk-free rate near 4.5%), and crypto was the funding source. Two triggers lit the fuse last week: a strong jobs print that takes Fed cuts off the table, and the biggest buyer of Bitcoin becoming a seller. The rotation can reverse as fast as it came, so watch how the summer unfolds around these storylines.

What I'm Watching

The IPOs at launch. SpaceX first, OpenAI and Anthropic into the fall. How they price, and how the broad equity market trades alongside them, is the cleanest read on whether this is a rotation inside a bull market or the top of one.

The Fed's read on the labor market. The next meeting on June 17 is expected to be a hold, but the market has moved to pricing hikes by year end. The question that matters is whether the committee takes the 172,000 headline at face value, which points that way, or starts looking for the cracks underneath it. That read, more than any single print, sets the path from here.

Oil and the Strait of Hormuz. The inflation forcing the Fed's hand is largely an energy story. Any further escalation that keeps Brent elevated keeps the hawkish path intact, and the pressure on risk assets with it. A genuine de-escalation is the cleanest path back to lower inflation, and the scenario the market seems least prepared for.

Michael Saylor's next move. Does he de-risk the company by raising cash, common or preferred, and rebuild the buffer? Or sit on his hands and remain a small, steady, forced seller to service the dividend? That single choice determines whether the reflexivity that started this turns back the other way.

Drop me a line:

What do you think? Do you like this? Do you not like this? I would love to hear your thoughts, so please reach me at akram@span.blog

Enjoyed this? Subscribe for the weekly deep dive, or drop me a line at akram@span.blog.