Arthur Hayes: AI arms race fuels fiat credit surge, lifts Bitcoin

Arthur Hayes, former BitMEX CEO and Maelstrom CIO, wrote that the global race to develop advanced artificial intelligence has produced a historic expansion in fiat credit. He described central banks and commercial lenders as financing large AI capital expenditures and said that dynamic is increasing demand for Bitcoin.

Hayes outlined his view in a recent newsletter. He noted that much of the first wave of AI capital spending came from operating cash flow at the largest software firms but wrote that those cash reserves are nearing exhaustion. He added that the next phase of investment will require growth in the credit channel, meaning greater reliance on commercial bank lending alongside central bank support.

Hayes pointed to policy shifts in major economies. He noted that Chinese policy has redirected lender activity from the property sector toward technology, and that the Federal Reserve and the People’s Bank of China have eased financial conditions that can help finance AI build-out.

Researchers at Simple Mining described AI capex as a national-security concern and argued hyperscale cloud providers will eventually turn to commercial bank balance sheets once free cash flow is exhausted. They added that bank lending can have a multiplier roughly three times larger than central bank lending.

The national-security framing has been reinforced by recent U.S. government action. On May 1, the Department of Defense signed AI deployment agreements with eight contractors: Google, Microsoft, Amazon Web Services, Nvidia, OpenAI, Reflection AI, SpaceX and Oracle.

White House AI and crypto adviser David Sacks estimated AI capital spending could add about 2 percentage points to U.S. GDP this year and cited a note from Morgan Stanley that projects a contribution above 3 percent next year.

Hayes invoked Jevons Paradox to explain why computing demand may continue to climb even as models become more efficient. He wrote that improvements in efficiency reduce the cost per compute task, which can lead to more tasks and greater total demand for energy and specialized hardware.

On cryptocurrency markets, Hayes reported that Bitcoin bottomed near $60,000 earlier this year and projected a return to $126,000 is now highly likely. He predicted faster gains if Bitcoin clears $90,000, when forced covering by short sellers could amplify price moves. He said investors are watching AI infrastructure spending, central bank policy, electricity markets and upcoming technology IPOs for early signals of a cycle shift.

Hayes warned the expansion need not be permanent. He wrote that an oversized AI initial public offering or a major merger could halt the rally. He also flagged political and economic risks, including anti-AI rhetoric from potential 2028 candidates and rising electricity and commodity costs that could prompt public pushback ahead of the U.S. midterm elections.

Bitcoin advocate Simon Dixon criticized the framing that supports expanded monetary backing, calling it a “manufactured crisis” to justify emergency bond purchases and a narrative to allocate capital in an emergency.