Arthur Hayes says AI spending mirrors 2008 credit bubble, not dot-com bust
Macro & Markets ·
The BitMEX co-founder argues a coming AI capital-spending slowdown could force government bailouts large enough to reignite a Bitcoin bull run, even as he flags near-term downside risk.
In an essay dated August 5, Arthur Hayes lays out a case that today's buildout of AI data centers and power infrastructure is being financed the way commercial real estate was before the 2008 crash rather than treated as a straightforward growth bet, according to cryptohayes.substack.com. Lenders, private credit funds and governments, he writes, are underwriting the expansion on the assumption that demand for AI compute will keep accelerating without limit, a premise he expects to unravel once capital expenditure growth slows in 2027 and then contracts.
Hayes frames this as fundamentally a credit story, distinct from the dot-com collapse, which he characterizes as an earnings-driven correction. Because AI infrastructure is viewed as strategically tied to national security, he expects governments to step in with bailouts and liquidity injections larger than those used after the 2008 financial crisis rather than allow major AI firms to fail, a scenario detailed in reporting from cryptopotato.com. That resulting wave of newly created money, in his view, would eventually flow into crypto markets and mark the start of a longer Bitcoin uptrend.
On near-term price action, Hayes says Bitcoin has not necessarily hit its cycle bottom, and could keep trading in a $60,000 to $70,000 band with a possible dip toward $50,000 while capital keeps flowing into AI projects before investors start questioning the returns. Separately, he points to Ethereum as a more immediate opportunity, suggesting growing institutional interest in tokenized real-world assets could lift ETH toward roughly $5,000 by the end of 2026 if that narrative takes hold.
The same argument — that AI capex resembles pre-2008 real-estate lending more than a dot-com-style equity bubble — has circulated across multiple outlets covering Hayes's essay, including analysis noted by leviathan.news and wublockchain.xyz, each emphasizing the expectation that central-bank-style easing would follow any AI-driven credit stress.
What remains unresolved is timing: Hayes ties the unwind to a 2027 slowdown in AI capital spending, but offers no fixed date for when a bailout response might materialize or how deep a Bitcoin pullback toward $50,000 could go before any recovery begins. Whether the tokenized-asset narrative gains enough institutional traction to push Ethereum toward $5,000 by 2026 is likewise an open question rather than a settled forecast.