Arthur Hayes: AI Debt Could Cause Bitcoin to Drop and Then Rebound

Could a debt crisis related to AI ultimately benefit Bitcoin? Arthur Hayes, co-founder of BitMEX, believes that the "Safety First" narrative adopted by several industry players also masks questions about the demand for their services. He outlines a possible chain of events: a slowdown in spending, deterioration of data center financing, public intervention, and then a return of liquidity to riskier assets. However, this interpretation remains a personal hypothesis, not an established forecast. But we are used to Mr. Hayes.

Key Points

• Arthur Hayes suspects AI companies are using safety to justify a potential slowdown in their investments.
• He mentions over $1 trillion in quality debt at risk, without providing detailed calculations.
• Oracle, Meta, and TeraWulf illustrate the increasing reliance on debt to finance data centers.
• According to him, a credit crisis would first lower Bitcoin before prompting monetary intervention that could support it.

Behind the "Safety First", Arthur Hayes Suspects an AI Slowdown

The concerns expressed about the safety of artificial intelligence are very real. Several industry leaders have recently called for a slowdown in certain research or for stricter controls. However, there is no evidence to suggest that these announcements primarily serve to conceal a drop in demand. This is Arthur Hayes's interpretation.

His reasoning is based on the gap between committed investments and generated revenues. Microsoft, Alphabet, Amazon, and Meta have allocated approximately $380 billion to their investment spending in 2025. A large portion of this was related to data centers and AI, but this amount does not fully correspond to spending directly linked to artificial intelligence.

He also cites a study from MIT's NANDA project, which found that about 95% of the generative AI pilots studied had not yet produced measurable effects on financial outcomes. This figure does not mean that 95% of projects have definitively failed: researchers mainly highlight issues of integration, organization, and adaptation to businesses.

Finally, skeptics are monitoring agreements made between capital providers and chip vendors. Nvidia has announced its intention to gradually invest up to $100 billion in OpenAI, depending on the deployment of new infrastructures using its own systems. This operation fuels criticism regarding the circular nature of certain financings, but does not constitute proof of manipulation of orders.

Arthur Hayes is known for publishing his economic analyses on Substack -- Source: X Account

Data Center Debt Concentrates Concerns

The need for financing indeed poses a more concrete risk. Morgan Stanley estimated as early as 2025 that the global development of AI infrastructure could leave an external financing deficit close to $1.5 trillion by 2028. Global debt issuance related to AI could approach $570 billion in the year 2026 alone.

Several operations illustrate this evolution. Oracle issued $18 billion in bonds in 2025. This issuance should not be confused with the approximately $18 billion in loans related to the Jupiter data center project, some of which are now trading at a discount.

Meta has created a joint venture with Blue Owl to finance the Hyperion campus in Louisiana. The debt is primarily in a non-consolidated structure, but Meta discloses the existence of this agreement and maintains economic commitments. Therefore, describing it as a completely hidden off-balance-sheet debt seems excessive.

TeraWulf, which originated from bitcoin mining and is now heavily involved in computing hosting, has also placed over $3 billion in secured debt to finance its infrastructure. However, the company still maintains mining activities: it is diversifying into AI rather than completely abandoning it.

Arthur Hayes places insurers and their reinsurance subsidiaries at the center of his scenario. He cites an estimate of $1.54 trillion in reinsurance among affiliated entities. However, this amount does not represent a direct exposure to AI or a potential loss: it measures a much broader set of risk transfers within the insurance sector.

Bitcoin: An Initial Decline Before Possible Return of Liquidity

In Mister Hayes' scenario, a deterioration of AI-related debt would first lead to a decline in the markets. Bitcoin would fall alongside tech stocks and other risky assets, while companies combining mining, data centers, and debt would face particular pressure.

An intervention from the Federal Reserve or the Treasury could then change the situation. Our expert today relies on March 2020, when bitcoin had sharply dropped before benefiting from the Fed's asset purchases, as well as on the banking crisis of March 2023. The BTFP (Bank Term Funding Program) launched after the collapse of Silicon Valley Bank was, however, a collateralized lending mechanism, not a traditional quantitative easing program.

The second part of the reasoning remains conditional. A crisis of AI-related debt does not guarantee either a public bailout or a sustainable rise in bitcoin. It would likely first cause a general reduction in risk, before the nature and extent of any potential U.S. response determine the reaction of cryptocurrencies. Arthur Hayes also has interests in projects combining AI and blockchain, an element to consider when interpreting his often very bullish views on Bitcoin.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.