Why an AI Bubble Burst Could Fuel the Crypto Bull Market

Speaking on the Bankless podcast, Arthur Hayes gave a blunt diagnosis of the disappointing cycle: “AI took all the money.” His argument is that every marginal dollar of printed fiat that historically flowed into Bitcoin and Ethereum got redirected into AI infrastructure instead. Bitcoin’s run from $69K to $125K, explosive by any prior cycle’s standard, felt muted precisely because capital was being absorbed elsewhere even faster. As he put it, “the only thing that investors want to allocate to is AI, whether that’s on the debt or the equity side.”

The Opportunity-Cost Problem

Hayes is careful that this isn’t a knock on crypto’s fundamentals, it’s about relative returns. Investors chase the fastest horse, and right now AI is winning that race. He framed it starkly: “If Bitcoin went to $500,000, let’s call that a 10x return, SK Hynix went up 10x this year. You could find some second or third derivative component necessary for data center buildup, put your entire net worth into that, and be 20x in six months. Why would you buy Bitcoin?”

The logic is simple and, for crypto, uncomfortable. “It’s always an opportunity cost,” Hayes said. “Everyone knows that this unit of fiat will be worth less in the future. What can I buy now that’s going to outperform that? I want the fastest horse. AI is the fastest horse.” Until that changes, he argues, the marginal dollar keeps going to AI, not Bitcoin.

The Data Behind the Drain

Hayes and Saylor describe the “suction effect” anecdotally, but institutional data backs the mechanism up. The Bank for International Settlements, in a January 2026 bulletin, described AI investment as reshaping corporate balance sheets “on an unprecedented scale,” with firms increasingly moving beyond operating cash flows and turning to debt and private credit to fund the buildout. As the BIS put it, that shift creates a suction effect in the financial system: as investors and corporations prioritize multi-billion-dollar AI opportunities, marginal capital gets diverted away from other asset classes, altering the liquidity landscape for the broader market, the exact dynamic Hayes and Saylor describe hitting crypto.

J.P. Morgan Asset Management puts numbers on the scale. By its estimates, 2026 capital expenditure for just the five largest US hyperscalers has reached roughly $697 billion, up $173 billion since the start of the year. More telling is the ratio: AI capex for these firms has surged from about 33% of their collective operating cash flow in 2023 to an estimated 93% in 2026, meaning these companies are now plowing nearly all their available operational cash into the AI buildout. That concentration, J.P. Morgan notes, carries its own risk: as hyperscalers exhaust their cash to fund the buildout, the market increasingly demands evidence of real AI demand to justify the spending, setting up potential “catch-up or catch-down” volatility once it tests whether AI revenue can support these historic expenditure levels.

That last point is where the institutional data and the Hayes-Saylor thesis converge. The same concentration of capital draining other asset classes is also what makes the trade fragile: if the spending outruns the revenue, the correction Hayes describes becomes the catalyst, and the capital currently locked in AI starts looking for a new home.

READ MORE:

Dan Tapiero on Bitcoin’s Path to $1M and Market Cycles

Why a Burst Could Be the Catalyst

Here’s the inversion at the heart of the thesis. If AI is absorbing capital on the way up, an AI crash may release it on the way down, and Hayes’s broader argument is that the same fiat-debasement trade currently funding AI capex might rotate violently into hard assets and crypto the moment AI stops delivering. The setup, in his framing: AI absorbs everything on the way up, crypto absorbs everything on the way down. His ideal, if difficult, trade is to “ride the AI bubble into cash, let everything correlate to one on the downdraft, then buy the bottom.”

In that light, Hayes sees current prices as opportunities rather than warnings. “Ether is 30% below the 200-week moving average,” he noted. “They are strictly deals in this moment.” His read is that BTC and ETH are cheap relative to how much fiat has been printed globally, the missing ingredient is the catalyst to send that fiat back toward crypto.

Saylor: A Temporary “Suction Effect”

Michael Saylor, speaking in an appearance referenced on The Pomp Podcast, offers a similar diagnosis with a clearer timeline. He frames Bitcoin’s weakness not as a structural problem but as a temporary capital queue. AI deals are so large and attractive that Wall Street is actively marketing them, creating what he calls a suction effect pulling money from every other asset class, Bitcoin included.

His numbers put scale on it: “SpaceX, Anthropic, OpenAI, Google, Meta, we’re talking about $500 billion of capital they’re working to raise to power up their AI data centers. That’s creating a hot set of deals that Wall Street is marketing. Everybody wants to get into the deal.” The Bitcoin impact, he stresses, is mechanical rather than philosophical: only “1% or 2% of that capital is coming from Bitcoin.”

Why Saylor Sees It Reversing

Saylor’s case is that the rotation runs both ways. Hot money enters AI deals, flips them, then rotates back, and as lockups expire, AI winners tend to diversify, which historically has meant Bitcoin. “Once the deals have gone through, the early hedge funds and traders will flip it and rotate back the other way. The hot money will come back.” His estimate is a “12 to 24 week cycle,” with the AI “suction activity” easing by the end of 2026.

He adds an Austrian-economics twist: a falling Bitcoin price makes Bitcoin more attractive to buy, not less. “As the equity and the Bitcoin fall, there’s more reason to buy it.” In his telling, the same mechanism draining capital now is quietly building the case to re-enter.

The Honest Caveat

Both arguments are compelling, and they share a structure: crypto is cheap, the capital is elsewhere, and a rotation back is coming. But it’s worth being clear that this is a bullish thesis, not a certainty, and it hinges on a premise that may not hold. The entire case depends on AI either crashing (Hayes) or its capital naturally rotating out (Saylor). If AI keeps delivering returns and the capital stays put, the catalyst never arrives and crypto’s prices could keep reflecting that drain. Saylor’s 12-to-24-week timeline is an estimate, not a schedule, and Hayes himself admits the “ride it into cash and buy the bottom” trade is easy to describe and hard to execute.

What both are really describing is a conditional setup: crypto positioned as the destination for capital that is currently committed elsewhere. Whether that capital actually rotates, and when, is the open question. The thesis is coherent, but it’s a bet on a sequence of events, an AI peak, a reversal, a flight to hard assets, that has to play out in order. For now, it remains an argument about what could happen, not confirmation that it will.

The post Why an AI Bubble Burst Could Fuel the Crypto Bull Market appeared first on Coindoo.