Hayes’ core claim: fiat creation didn’t reach BTC
Bitcoin has been stuck under $63,000, and BitMEX co-founder Arthur Hayes has a blunt explanation. In a recent blog post, Hayes says the AI boom absorbed a large share of newly created dollar liquidity. That, he argues, explains why bitcoin struggled to rally further even as money supply expanded more broadly.
Hayes revisits a long-held thesis: crypto markets are largely driven by fiat liquidity. But he now concedes he “may have overlooked an important factor” in that framework. The missing piece was not just how much money was created. It was where that liquidity actually went.
The “great AI bubble” drew the money first
Hayes ties the shift to the commercial launch of ChatGPT in November 2022, which he calls the start of the “great AI bubble.” In his telling, bitcoin should have performed much better across the same period because dollar creation rose. Instead, AI-linked investments pulled in more capital than crypto.
He points to relative performance. Hayes cited Nvidia’s roughly 11x increase compared to BTC’s 7x gain over a similar timeframe. He also says AI outperformance accelerated from late 2024 onward, while bitcoin later slid sharply from its peak.
The implication is simple. If incremental liquidity gets absorbed by a different trade, bitcoin can look “underfunded” even when headline money printing looks supportive.
Capital intensity matters more than liquidity headlines
Hayes says his earlier models focused on the headline amount of fiat creation and assumed the rest would eventually find its way into bitcoin. That assumption, he argues, failed because AI is extremely capital intensive.
According to Hayes, AI requires sustained investment in data centers, electricity generation, specialized chips, and the supporting infrastructure around them. He links the urgency to data center spending that started expanding in 2024 and accelerated in 2025.
Hayes then brings in debt issuance figures compiled from public disclosures. He says AI-related firms issued about $1.5 trillion in debt between November 2022 and the present. He estimates around $1.3 trillion of that was raised from 2025 onward as AI infrastructure spending surged.
He compares that to growth in the US M2 money supply over the same period, which he estimates also increased by about $1.5 trillion. From that comparison, Hayes concludes that AI effectively absorbed nearly all newly created dollar liquidity, writing “AI sucked up all created dollars.”
Where the argument lands for BTC holders
This isn’t a claim that bitcoin has no liquidity sensitivity. Hayes’ point is about timing and routing. If AI funding needs are large enough, they can soak up marginal capital before it reaches crypto.
That matters for how readers should interpret “why bitcoin isn’t following money supply higher.” Hayes is arguing the missing variable is not just macro liquidity. It is the competing use of that liquidity.
The article also flags skepticism from market observers about bitcoin’s near-term setup. The newsroom’s caution here is about what we can actually verify from the text. Hayes provides quantified estimates for AI debt and M2 growth, but his broader interpretation is still a model. The cycle commentary comes from a named analyst, but it is still scenario-based.
Cycle talk: more volatility is the expectation
In the same CryptoPotato piece, market analyst Doctor Profit says bitcoin has entered the fifth stage of a six-stage bear market cycle. He characterizes this phase as one with higher volatility and emotional stress.
Doctor Profit’s view, as reported, is that the recent pullback is not the final bottom. He flags a $40,000 to $48,000 range as the most likely area for BTC’s eventual cycle low. He also suggests that low could arrive between September and October 2026.
The key practical consequence for readers is that Doctor Profit’s framework expects turbulence after the current drop, not a smooth return.
Fact snapshot from the source
| Topic | Figure or claim (as stated) | Source in text |
|---|---|---|
| Bitcoin price context | BTC struggling below $63,000 | CryptoPotato |
| AI start point Hayes cites | Commercial launch of ChatGPT in Nov 2022 | CryptoPotato quoting Arthur Hayes |
| Nvidia vs BTC relative move | Nvidia ~11x, BTC ~7x | CryptoPotato quoting Arthur Hayes |
| AI debt issuance (total) | ~$1.5 trillion in debt from Nov 2022 to present | CryptoPotato quoting Arthur Hayes |
| AI debt issuance (since 2025) | ~ $1.3 trillion from 2025 onward | CryptoPotato quoting Arthur Hayes |
| US M2 growth (same period estimate) | ~$1.5 trillion increase | CryptoPotato quoting Arthur Hayes |
| AI liquidity conclusion (Hayes quote) | “AI sucked up all created dollars” | CryptoPotato quoting Arthur Hayes |
| Cycle-range estimate | $40,000 to $48,000 possible cycle low zone | CryptoPotato quoting Doctor Profit |
| Possible timing for cycle low | Between Sep and Oct 2026 (as stated) | CryptoPotato quoting Doctor Profit |