BitMEX co-founder Arthur Hayes has advised crypto investors to go long Bitcoin (BTC) and Ethereum (ETH) rather than attempt to short an artificial intelligence stock collapse, arguing that future monetary rescues will drive digital assets to record highs.
Trading the Rescue Instead of the AI Bubble
In an October 15, 2026 post on X, Hayes outlined his investment thesis, stating he expects the ongoing AI data center expansion to hit a wall around late 2027 or 2028. Drawing on historical technology cycles that experience overbuilding followed by contraction, he believes central banks and governments will ultimately respond with multi-trillion-dollar bailouts larger than those issued during the 2008 financial crisis.
Hayes argues that newly printed fiat currency will search for store-of-value assets, sending Bitcoin toward $1 million or higher. He cited the 2020 pandemic monetary expansion—during which the Federal Reserve doubled its balance sheet and Bitcoin rallied from under $5,000 to over $60,000—as precedent for this mechanism. However, market metrics reflect recent divergence: US money supply hit a record high in August, yet Bitcoin remains roughly 35% below its October 2025 peak, with both BTC and ETH underperforming AI equities throughout 2026.
Key Takeaways
- Arthur Hayes recommends buying BTC and ETH to capture liquidity from prospective government bailouts rather than shorting AI stocks.
- Hayes projects that the AI data center expansion will peak and stall around late 2027 or 2028.
- Market history demonstrates precedent from 2020, when balance sheet expansion propelled Bitcoin from under $5,000 to over $60,000.
- Fundstrat's Tom Lee presents an alternative outlook, comparing current AI growth to 1997 rather than 1999, with an estimated peak in 2033.
AI Debt Risks and Differing Market Perspectives
Concerns surrounding AI sector leverage spiked after CoreWeave shared a chart of falling share prices. CoreWeave stock dropped 15% just one day after CNBC host Jim Cramer identified it as his single trusted AI cloud pick. In a September newsletter, Hayes claimed major US insurance firms were already insolvent at market value due to exposure to AI debt, echoing broader warnings about how rising US borrowing costs impact leveraged balance sheets.
Not all analysts share Hayes' timeline for an imminent crash. Fundstrat co-founder Tom Lee has argued that the AI development cycle resembles 1997 rather than the late-stage bubble of 1999, projecting that any structural peak is unlikely until around 2033.
Why It Matters
Hayes' commentary shifts the focus for crypto market participants from timing an equity market top to positioning for structural central bank responses. If AI infrastructure investments prove over-leveraged, central banks may be forced to choose between systemic financial instability and renewed monetary easing. For digital asset markets, this highlights how long-term valuations remain deeply tied to macroeconomic liquidity injections rather than immediate sector correlation.
