Arthur Hayes is making a blunt liquidity argument for why Bitcoin has struggled to sustain rallies.
In his Benzinga essay, Hayes says the “AI bubble” absorbed capital that otherwise could have flowed into cryptocurrencies. His headline thesis is mechanical, not mystical. He points to a near match between new U.S. money supply growth and AI debt issuance.
The liquidity swap Hayes claims
Hayes frames the problem around U.S. M2. He says US M2 rose by about $1.5 trillion from November 2022 to “today.” Over the same window, he claims AI companies issued roughly $1.5 trillion in debt to fund data center construction.
The takeaway in Hayes’ telling is simple. If AI financing took in that new liquidity first, there was less remaining fuel for assets like Bitcoin. Hayes writes, “Bitcoin never had a chance.” He also argues the earlier rebound from the November 2022 Bitcoin low happened for a specific reason, not because of crypto-specific demand.
Why the timing matters
Hayes says the “only reason Bitcoin rallied strongly off the November 2022 low” was that the AI debt binge only truly kicked into gear later, starting in 2025.
only reason Bitcoin rallied strongly off the November 2022 low
That’s the crux of his counterfactual. Bitcoin did have room to move in the period before AI borrowing fully ramped. Once the AI debt cycle intensified, Hayes argues the marginal dollar went elsewhere.
He further claims Bitcoin peaked in October 2025, “precisely …” the period when the AI debt dynamics were already established. The piece, as provided in the source excerpt, cuts off before the exact rest of his timing explanation, so readers should treat that portion as Hayes’ claimed linkage rather than a fully documented sequence.
What this means for Bitcoin as an asset with risk
Hayes’ case is not a technical roadmap critique. It’s macro liquidity framing. That matters because liquidity narratives often explain why catalysts can fail even when crypto fundamentals look fine on paper.
If Hayes is right, Bitcoin’s next sustained move would depend less on protocol news and more on whether the AI-financing cycle slows or reverses. The risk for holders is that “bubble pop” is not a single event. Liquidity can tighten gradually, then re-open, and cross-asset flows can shift without a clean unwind.
Also, Hayes’ argument is only as strong as the implied accounting. His essay claims a near one-to-one mapping between M2 growth and AI debt issuance over a specific window. If that mapping misses other channels, like how that debt spending circulates through the broader economy, the conclusion weakens.
Still, the skepticism is the point. Hayes isn’t selling an improvement plan for Bitcoin. He’s arguing Bitcoin got outbid for new dollars.
The bottom of the thesis
Benzinga’s excerpt ends with Hayes’ strongest line. “Bitcoin never had a chance,” he wrote. His specific condition for a better outcome is that the AI bubble must pop.
That sets a clear, if uncomfortable, bar. Until liquidity stops being net-absorbed by AI financing, Bitcoin may remain stuck competing for scraps.