Arthur Hayes: Bitcoin may fall to $50,000 before $1M

Arthur Hayes, co‑founder of BitMEX and chief investment officer at Maelstrom, wrote in an Aug. 4 essay that bitcoin could fall to roughly $50,000 before later reaching $1 million. He described the current artificial intelligence data‑center expansion as a leveraged credit bubble that could trigger large government support.

Hayes compared the AI infrastructure buildout to a real estate credit boom rather than a typical technology cycle. He said hyperscalers — the large cloud companies funding data centers to train and run AI models — are effectively financing real estate filled with hardware that loses value quickly as newer chips arrive.

Hayes argued lenders could end up holding facilities with increasingly obsolete equipment. He also noted market conditions that encourage continued lending: short‑term interest rates below nominal economic growth and higher longer‑term yields, which widen the spread between banks’ funding costs and their lending rates and can make data‑center loans more profitable.

The Federal Reserve Board’s May 2026 Financial Stability Report included market contacts who identified AI among key vulnerabilities, along with equity valuations, debt‑financed capital spending, labor‑market pressures and private credit. In interviews cited by the Fed staff, half of respondents listed AI as a potential shock, up from about 30% in a fall 2025 survey.

Hayes set out a timeline in which the peak pace of data‑center capital spending slows from mid‑2027 and becomes clearly visible in 2028. He likened the pattern of continued credit expansion through a slowdown to the 2006–2007 mortgage period, when lending kept growing after house‑price gains had slowed.

To describe how policymakers might respond, Hayes sketched a hypothetical support plan using emergency lending powers and the Exchange Stabilization Fund. He noted the ESF holds about $28 billion and applied a 10x leverage ratio used in past Fed‑backed special‑purpose vehicles to suggest Treasury authority could mobilize roughly $280 billion into struggling AI projects. He characterized such intervention as equity‑focused liquidity rather than standard quantitative easing.

Hayes linked those potential liquidity injections to a sharp rise in bitcoin prices. He wrote: “Once the authorities sufficiently panic because their AI‑created GDP growth is just another run‑of‑the‑mill property bubble, they will print money in sums greater than the 2008 GFC. This will ultimately drive bitcoin to one million and beyond.”

On bitcoin’s nearer‑term path, Hayes noted the cryptocurrency peaked in October 2025 and then lost about half its value as capital flowed into AI credit and equity. He warned the price could trade in a range before rising, writing that bitcoin could “chop between $60,000 to $70,000 for a while with a potential downside of $50,000.”

Hayes also referenced potential near‑term pressures, including possible sales by large corporate holders such as MicroStrategy. He described the scale of capital misallocation into AI infrastructure as already larger, relative to GDP, than the subprime episode and comparable to 19th‑century railroad investment.

The essay frames a multi‑year path in which credit expands through a slowdown, policymakers provide large liquidity support when losses mount, and bitcoin responds to that liquidity after an initial correction. Hayes has previously linked crypto price targets to shifts in credit and fiscal policy and reiterated a view that liquidity flows can shape bitcoin’s major moves.