Higher Treasury Yields Could Pressure U.S. Growth and AI Spending

Higher borrowing costs could shave 0.2 percentage point from U.S. growth in 2027 under a baseline outlook, with the drag exceeding 0.5 point if yields remain near current levels, leaving artificial intelligence (AI) investment exposed.

Bond yields have climbed to multidecade highs around the world, while U.S. economic growth and equities have so far remained firm. The baseline outlook assumes the Federal Reserve raises rates once more in December 2026, then cuts three times in the second half of 2027. The 10-year Treasury yield reaches 4.4% at the end of 2026 in that scenario, which puts the rate burden on 2027 growth at 0.2 percentage point; if yields stay near current levels, the hit rises above 0.5 point.

The same rate outlook carries a fiscal cost. U.S. debt is projected to reach 124% of gross domestic product in 2035 under the baseline and 132% if current rates persist, adding to pressure from large federal deficits and higher interest expense.

For households and businesses, the rate channel is already visible in mortgages: the 30-year mortgage rate has topped 7%, putting pressure on housing investment. A broader estimate for developed economies finds that each 100-basis-point rise in rates would reduce housing investment by about 6%, consumer spending by 0.2% and business capital spending by 1.8%. Together, those declines would trim growth across developed economies by about 0.6 percentage point.

AI spending has financial cushions. Hyperscalers account for more than two-thirds of U.S. AI capital spending, and AI-related issuers rated AAA to AA have interest coverage of 27.7 times and net leverage of 0.5 times. But investment links between AI firms, leverage added through exchange-traded funds and derivatives, and uncertainty over sustained revenue growth create exposure if borrowing costs stay high.

The U.S. economy remains vulnerable, and AI activity could be among the first areas to weaken if rates stay at current levels or higher for another one or two quarters. AI-linked activity and stock-market wealth effects are estimated to add 0.6 to 0.7 percentage point to growth of about 2%, leaving non-AI growth around 1% to 1.5%.

The S&P 500’s price-to-earnings ratio fell from 23 to 19 this year, while earnings grew more than 30% over the past 12 months. Still, a rise in the 10-year yield of more than 50 basis points in a month would match a historical stress pattern, and market breadth is at its lowest level since the dot-com era. Banks and credit markets remain broadly manageable, but if volatility returns to 2022–23 highs, it could undermine yield-dependent institutional investors’ confidence and materially pressure demand for investment-grade corporate debt.