Hayes Warns AI Credit Bubble Could Drive Bitcoin Toward $1M

Arthur Hayes, the former co-founder of BitMEX, is warning that today’s surge in AI infrastructure spending could sow the seeds of a renewed credit crunch—one he believes may ultimately send Bitcoin to highly elevated levels.

In a Tuesday blog post, Hayes argued that the boom is being treated by investors as a high-growth technology earnings story, when he views it more like leveraged real estate. He expects lenders to fund aggressive data-center and power buildouts, only for a slowdown in AI-related capital expenditure to reveal weaker borrowers. From there, Hayes suggested, a government liquidity response could reintroduce significant risk assets into the broader market, with Bitcoin potentially rallying far beyond current ranges.

Key takeaways

• Hayes frames AI infrastructure expansion as a “credit story” rather than an “earnings story,” drawing a parallel to the 2008-style credit cycle.
• He expects banks to finance data-center construction and believes the exposure will become clearer when AI spending growth cools.
• Hayes said Bitcoin could churn in a range of $60,000 to $70,000, with downside risk to $50,000 before any credit-driven recovery.
• He forecast Ether could reach $5,000 by year-end and said his firm Maelstrom plans to accumulate while selling out-of-the-money ETH puts.
• Recent reporting highlights the scale of future AI data-center lease commitments, underscoring the leverage embedded in the buildout.

Hayes’ “AI is real estate” credit-cycle warning

Hayes’ latest argument centers on how the AI buildout is financed. In his view, spending on data centers and power infrastructure is not the same as investing in product-driven technology growth. Instead, he characterizes it as a leveraged commitment that resembles property finance—where cash flows depend on demand staying strong and credit remaining available.

That distinction matters because credit cycles can turn quickly when expectations are met too early or when capital expenditure slows. Hayes’ thesis is that lenders will continue extending funding while projects are still ramping, but problems may surface after AI capital expenditures weaken and borrowers face difficulty servicing obligations. In that scenario, he expects liquidity measures from policymakers to follow—potentially injecting fresh capital into financial markets.

From 2008 comparisons to Bitcoin’s speculative path

Hayes directly compared the dynamic to 2008, calling the AI boom a “credit story like 2008 and not an earnings story like 2000.” He stressed that the key driver for crypto, in his framing, would not be fundamental “earnings” growth from the AI sector itself, but rather the liquidity response that could follow a credit deterioration.

In the meantime, he outlined a near-term technical-style range for Bitcoin. Hayes said BTC could remain between $60,000 and $70,000, with potential downside to $50,000, before any recovery tied to the credit cycle and government liquidity response. He also floated the prospect that, if the cycle plays out as he expects, Bitcoin could eventually be driven to $1 million or higher.

It’s important to note that Hayes’ scenario is inherently speculative. The argument depends on a specific chain: overbuilding in AI infrastructure → weaker borrowers → a credit crisis → policy liquidity support → renewed inflows into risk assets like Bitcoin. While the general linkage between credit conditions and market liquidity is a recurring theme in macro finance, the timing and magnitude Hayes suggests remain uncertain.

What changes, and how Hayes positions within the market

Hayes’ outlook includes both a macro forecast and an options-and-positioning angle. He predicted Ether (ETH) would reach $5,000 by year-end and said Maelstrom intends to build a “significant position” while simultaneously selling out-of-the-money ETH put options. The structure signals a willingness to hold exposure while collecting premium that could cushion downside—though the payoff depends on where ETH trades relative to the strike prices and volatility conditions.

His thinking also builds on earlier public comments about how AI competition and capital allocation could affect crypto liquidity. On May 13, Hayes said US-China competition in AI would encourage bank lending and fiat creation—an environment he argued could benefit Bitcoin. On June 4, he previously said he sold HYPE and NEAR after warning that major AI-related listings could divert capital away from crypto.

Taken together, the throughline is that Hayes sees crypto’s near-to-medium term direction as sensitive to macro and liquidity flows, not just to crypto-native fundamentals. Where AI spending is framed as a credit lever, the opportunity for crypto comes from the knock-on effect: whether the broader system expands liquidity—or contracts it under stress.

Why leverage in AI infrastructure is getting attention

Hayes’ caution about financing risk comes as reporting has begun to quantify the scale of commitments behind the AI buildout. According to Reuters, Microsoft, Meta, Oracle, Amazon, and Alphabet have committed about $1.09 trillion to leases that have not yet commenced, largely for data centers. Reuters noted that this figure cannot be treated as a straightforward debt total because it reflects undiscounted payments spread across multiple years.

Still, Reuters highlighted that the commitments are nearly four times the roughly $285 billion in lease liabilities already recognized by the same companies. The gap matters because off-balance-sheet commitments can become a stress point if operating assumptions weaken, especially if the buildout timing and actual demand for capacity diverge.

Reuters also pointed to uneven strain across firms. A separate Reuters analysis cited that Oracle’s debt was about 4.3 times its earnings before interest, taxes, depreciation, and amortization, while Alphabet, Amazon, Microsoft, and Meta had ratios below one. Reuters further quoted S&P Global analyst Andrew Chang, who said Oracle’s data-center leases run for 15 to 19 years, while customer contracts last no more than five years—creating a mismatch that could increase risk if customers do not renew or expand on the expected schedule.

For crypto investors tracking Hayes’ thesis, the relevance is straightforward: if the AI infrastructure ramp becomes a drag on credit and financing markets, it could translate into broader liquidity constraints. Conversely, if policymakers respond aggressively to maintain stability, that same liquidity could later flow back into speculative assets—where Bitcoin has historically captured attention during risk-on phases.

Going forward, market participants will likely watch whether AI infrastructure spending and financing conditions begin to show signs of strain, and whether policy-makers move to support credit markets if they do; Hayes’ case hinges on that transition from construction optimism to a liquidity-driven response. Until there is clear evidence of a slowdown in capital expenditure or credit stress in the real economy, his BTC range and $1 million-plus scenario remain a high-volatility narrative rather than a confirmed forecast.