By Sophie Lindqvist

U.S. net interest payments reached $1.05 trillion in fiscal 2026—$2.9 billion a day and an estimated 3.2% of GDP, breaking the 1991 record. Those costs underpin a growing market concern that fiscal risk, not AI borrowing alone, is driving long-term yields.

The evidence cited by Robin Brooks points to a rising fiscal risk premium—the additional return investors demand for holding Treasuries as debt and deficits mount. Brooks, a senior fellow at the Brookings Institution, places that view at the center of the latest contest over yields.

“The midterm elections and the AI sell-off that’s ensued have given us the perfect natural experiment to disentangle what’s going on with long-term yields. They’re up because of mounting fiscal stress, not because of AI,” he wrote.

The yield surge

Before easing in recent weeks, the 10-year Treasury yield rose more than 100 basis points (one-hundredth of a percentage point each), topped 5% and reached a 24-year high. Investors then sought shelter in Treasuries.

The Iran war began in late February, lifting oil prices and reviving expectations that the Federal Reserve would raise rates against resurgent inflation. AI infrastructure operators known as hyperscalers also flooded the bond market with debt to finance massive capital spending.

That spending has kept the U.S. economy hot and inflation sticky. The resulting corporate supply competed for investor dollars as the Treasury tapped the bond market to finance a $2 trillion budget deficit and pay interest on $40 trillion of federal debt.

Some Wall Street observers argue that AI-related debt crowded out U.S. bonds, forcing Treasury auctions to offer higher returns. Brooks disagrees, citing historically high public borrowing and corporate issuance that remains below early-2000s IT-bubble levels.

The sell-off test

In a Saturday Substack post, Brooks described the episode as a natural experiment separating fiscal and AI explanations. He put himself among those seeing a fiscal risk premium, against the view that hyperscaler issuance explains the rise without signaling trouble in fiscal policy.

Brooks used the latest monthly jobs report as a cross-check. Payrolls grew by more than double the consensus estimate, yet the two-year Treasury yield rose by more than hot economic growth alone could explain. The move weakens both the claim that U.S. growth depends on AI and the argument that AI borrowing is driving higher yields.

He also examined the 10y10y forward yield, the market-implied return on a 10-year Treasury held for another 10 years. It gauges the premium investors demand for holding Treasuries over a long horizon, including debt and deficit risk. If an AI bubble were pushing long-term yields higher, that measure should have fallen with the 10-year and two-year yields. It did not, which Brooks viewed as further evidence for a fiscal premium.

The conclusion remains contestable. Oil prices, expected Federal Reserve tightening and private AI investment can still lift yields; one market episode cannot isolate their respective contributions. Brooks’s test relies on comparative bond pricing rather than a controlled measurement of every investor’s motive.

Fiscal arithmetic

The Treasury Department’s final fiscal 2026 tally, released last month, showed:

• The federal budget deficit widened to $1.99 trillion, 12% worse than the $1.78 trillion gap in fiscal 2025.
• Revenue rose 3% to $5.4 trillion, while spending increased 6% to $7.4 trillion.
• Social Security and Medicare ranked as the largest line items. Interest ranked third, exceeding defense spending and Medicaid.
• Gross interest costs reached $1.36 trillion, with an average interest rate of 3.53%. That exceeded twice the 2021 level, just before the Federal Reserve began aggressive inflation-fighting tightening.
• Net interest payments increased 8% to $1.05 trillion, outpacing growth in total spending, revenue and gross domestic product. Their increase exceeded real GDP growth of about 2% and nominal GDP growth of about 6%.
• Fiscal 2026 net interest equaled $2.9 billion a day. The Peter G. Peterson Foundation estimated it at 3.2% of GDP, breaking the previous record set in 1991.

These figures explain why higher borrowing costs can compound the problem. A larger debt stock requires more interest; higher interest raises the cost of replacing maturing debt and can add to future borrowing needs. The feedback loop tightens as deficits and debt remain historically high.

What does this change for bond investors?

It shifts attention from corporate AI issuance alone to the premium attached to long-dated U.S. debt. Investors receive more yield, but the market signal says fiscal deterioration may be the more persistent force setting that required return.

Committee for a Responsible Federal Budget President Maya MacGuineas said in September: “Higher interest begets more debt, and greater debt begets higher rates and higher interest. Eventually, the whole thing could come crashing down.”

MacGuineas called for immediate fiscal intervention by the President, the Treasury Secretary and congressional leadership. She cautioned that there are no easy fixes and Washington cannot wish the problem away. Her warning offers no specific remedy, and the market signal does not dictate how large any fiscal adjustment must be. The practical issue extends beyond whether AI can finance ambitious capital spending; it is whether Washington can stabilize debt costs without adding pressure to an already hot economy.