AI Debt Could Keep Bitcoin Under Pressure After Fed Hikes

Minutes from the Federal Reserve’s Sept. 15-16 meeting identified private debt issuance for artificial intelligence infrastructure as one factor behind higher U.S. Treasury yields and term premiums. The financing demand could keep long-term borrowing costs elevated after the Fed ends its rate increases.

Treasury yields rose by about 35 basis points across maturities from two to 10 years between the September meeting and the previous policy meeting. The minutes listed stronger economic data, expectations for more rate increases, geopolitical developments and uncertainty about Treasury buybacks as additional factors.

The Fed raised its federal funds target by 25 basis points in September to a range of 3.75% to 4%. Most officials judged that another increase would probably be appropriate before the end of the year. A pause could reduce expectations for short-term borrowing costs without lowering longer-term yields if companies continue competing for financing.

The Bank for International Settlements estimates that the five largest technology companies will spend more than $1 trillion on AI-related capital expenditure in 2025 and 2026. Industry projections cited by the BIS place global AI investment at about $500 billion currently, with the figure potentially reaching $3 trillion to $4 trillion by 2030.

Technology companies initially financed much of the AI buildout with operating cash flow. That financing mix is changing as spending on data centers, chips and energy infrastructure exceeds earnings and free cash flow at some companies. Corporate bonds and private credit are becoming a larger part of the funding mix.

The Fed’s trading-desk manager noted that spreads on debt issued by major cloud providers remained wide because of the amount borrowed and the debt’s long maturities. Market participants also identified competition from AI-related private issuance as one factor lifting Treasury term premiums, which compensate investors for holding longer-term bonds.

The minutes did not quantify the portion of the rise in Treasury yields linked to AI financing. The Fed noted that changes in real interest rates accounted for most of the increase in longer-dated yields during the period. Real yields measure the return investors receive after inflation.

Treasury data placed the 10-year nominal yield at 5.28% on Oct. 7 and the inflation-adjusted 10-year yield at 2.92%. Treasury securities provide contractual income, while Bitcoin does not. Higher real yields can increase the return investors require from riskier assets.

AI-related equities have absorbed higher financing costs more easily so far. The Fed noted that companies benefiting directly from infrastructure spending outperformed the broader market, supported by stronger current and expected earnings even as valuation multiples declined.

The BIS has warned that the AI investment boom could create financial risks if companies build more capacity than future demand can support. Greater reliance on debt could increase the risk of financial stress and forced asset sales if revenue expectations weaken.

Arthur Hayes, co-founder of the defunct BitMEX exchange, expects the data-center race to produce excess computing capacity and a downturn. He has placed that risk around late 2027 or 2028, when new capacity could come online. Hayes expects policymakers to respond to financial stress with liquidity support, which he views as a potential benefit for Bitcoin and other crypto assets.

AI demand could instead grow fast enough to absorb new infrastructure, while higher productivity and profits could help companies service their debt. The Fed’s next meeting is scheduled for Oct. 27-28. Bitcoin investors will be watching 10-year yields and term premiums after the final rate increase.