According to BitMEX co-founder Arthur Hayes, the colossal, debt-fueled expansion of artificial intelligence infrastructure has the potential to conclude in a credit crisis on par with the 2008 financial collapse. He has predicted that the government liquidity measures that would follow such a crisis could ultimately propel Bitcoin (BTC) to a price of $1 million or even higher.

In a blog post published on Tuesday, Hayes contended that investors are incorrectly perceiving the massive spending on data centers and power infrastructure. Instead of seeing it as a high-growth technology investment, he argues it should be viewed as a form of leveraged real estate. Hayes anticipates that lenders will finance an excessive amount of construction, which will lead to a crisis when a slowdown in AI capital expenditure reveals the vulnerability of weaker borrowers.

This thesis links the trillion-dollar development of AI infrastructure to a potential new influx of liquidity into the cryptocurrency market. It is important to note, however, that the crisis Hayes envisions, along with the government bailout and subsequent Bitcoin rally, remains speculative.

Hayes characterized the AI boom as more of a “credit story like 2008 and not an earnings story like 2000.” In the near term, he projects that BTC could fluctuate between $60,000 and $70,000, with a potential dip to as low as $50,000. This period of consolidation would precede a recovery driven by the credit cycle and the resulting liquidity injection. Separately, Hayes forecasted that Ether (ETH) could reach $5,000 by the end of the year. He also stated that his fund, Maelstrom, is planning to establish a substantial position in the asset while simultaneously selling out-of-the-money ETH put options.

This latest outlook from Hayes expands upon his prior comments regarding the conflicting impacts of AI on crypto liquidity. He stated on May 13 that heightened competition in AI between the U.S. and China would stimulate bank lending and the creation of fiat currency, which would be beneficial for Bitcoin. Then, on June 4, Hayes revealed he had sold his HYPE and NEAR holdings, issuing a warning that major AI-related stock market listings could siphon capital away from crypto assets.

The sheer magnitude of the financial commitments supporting the AI boom is already becoming apparent. A Reuters report from Tuesday revealed that Microsoft, Meta, Oracle, Amazon, and Alphabet have collectively pledged approximately $1.09 trillion for leases that have not yet begun, with the majority designated for data centers.

These commitments are almost four times the roughly $285 billion in lease liabilities the companies have already recognized on their books. The Reuters report did note, however, that this $1.09 trillion figure cannot be treated as straightforward debt, as it comprises undiscounted payments that are scheduled to be spread out over a number of years.

Despite the massive overall figures, the financial burden is not uniform across the industry. A separate Reuters analysis found significant disparities. For instance, Oracle’s debt stood at about 4.3 times its earnings before interest, taxes, depreciation, and amortization (EBITDA). In contrast, the debt-to-EBITDA ratios for Alphabet, Amazon, Microsoft, and Meta were all below one.

Andrew Chang, an analyst at S&P Global, pointed to a specific vulnerability for Oracle. He highlighted that the company's long-term data-center leases, which have terms of 15 to 19 years, present a key risk. This is because Oracle's customer contracts are for much shorter periods, lasting no more than five years, creating a potential mismatch between its long-term liabilities and short-term revenue streams.