TLDR
- Arthur Hayes argues that a slowdown in AI spending could eventually benefit Bitcoin if it leads to government or financial-system support.
- His thesis centers on the large amount of debt being used to finance AI data centers, chips and computing infrastructure.
- Hayes says weaker AI demand could pressure private credit and insurers exposed indirectly to AI-related debt.
- Regulators acknowledge risks around private credit, but current data does not show that the U.S. insurance industry is insolvent because of AI exposure.
- Bitcoin was trading near $85,700 early Tuesday after gaining more than 6% during the previous session.
Former BitMEX CEO Arthur Hayes says a downturn in artificial intelligence investment could eventually create conditions that benefit Bitcoin. His argument is that weaker demand for AI computing could stress heavily financed data centers and private-credit markets, eventually prompting measures that increase dollar liquidity.
Check out my new essay "Safety First".
Trump has a choice, print or print.
"Did you hear that? The AI bros suddenly developed a conscience and are worried about humanity’s survival in the face of their almost silicon-God’s ascendance. It’s been almost silicon-God for some time… pic.twitter.com/xMRydV3lFf
— Arthur Hayes (@CryptoHayes) September 22, 2026
Hayes presented the thesis in a September 22 essay titled Safety First. He questioned whether recent calls to slow frontier AI development could partly reflect weaker economics, although OpenAI and Anthropic have publicly cited safety and security concerns rather than falling customer demand.
AI Debt Is Central to Hayes’ Bitcoin Thesis
Hayes argues that lower demand for AI training and inference could undermine assumptions supporting data centers, semiconductor purchases and private credit. Apollo has separately estimated that around $5 trillion could be spent on AI infrastructure through 2030, while AI-related financing could support more than $2 trillion of additional investment-grade debt.
Those figures demonstrate the scale of financing flowing into AI, but they do not establish that an AI credit crisis will occur. Apollo has warned that weaker-than-expected cash-flow growth among major hyperscalers could result in wider credit spreads and lower capital spending.
Hayes sees insurers and private credit as another possible pressure point. His scenario assumes that losses or downgrades on AI-related debt could create problems for institutions holding those assets, particularly where leverage or affiliated reinsurance structures are involved.
The National Association of Insurance Commissioners does say private credit carries lower liquidity, less pricing transparency and less frequent valuations than public debt. However, the regulator says these issues require continued monitoring rather than indicating broad deterioration across insurers’ portfolios.
Hayes Expects Any Rescue to Increase Liquidity
Hayes outlines two possible responses if AI infrastructure economics deteriorate. One involves the U.S. government supporting demand for computing capacity, while another involves financial assistance if private-credit losses threaten insurers and policyholders.
Neither policy has been announced by U.S. authorities in response to an AI debt crisis. Hayes argues that either route could involve additional borrowing or liquidity, which he expects would support Bitcoin and other scarce assets.
Current Federal Reserve policy complicates that thesis in the near term. The Fed raised its benchmark rate by 25 basis points on September 16 to a range of 3.75% to 4%, saying inflation remained elevated.
AI spending also remains strong despite recent safety debates. Nvidia recently outlined financing initiatives involving major investment firms, while SoftBank has begun marketing more than $11 billion in bonds to help fund its OpenAI investment.
Hayes’ Bitcoin forecast therefore depends on a future chain of events rather than conditions already in place. Bitcoin was trading near $85,700 early Tuesday after climbing more than 6% during the previous session.
