Are US Treasuries the 'Ultimate Killer' of the AI Bubble? Institution Warns: 10-Year Yield Breaking 5% Could Be the Trigger

Ruchir Sharma, chairman of Rockefeller International, believes that this round of US debt excess stems from the government rather than corporations. High fiscal deficits are pushing the 10-year Treasury yield toward the 5% mark. Since the AI infrastructure's funding gap of hundreds of billions of dollars relies heavily on external financing, a break above 5% in Treasury yields would create a severe crowding-out effect, significantly raising the cost of debt and equity financing for AI companies, potentially bursting the AI bubble and hitting stock market valuations.

The main driver of this round of US debt excess is the government, not corporations. This structural difference renders the traditional "corporate deleveraging" narrative ineffective and places the 10-year Treasury yield at a critical juncture that will determine the fate of the AI boom.

Ruchir Sharma, chairman of Rockefeller International, wrote in the Financial Times that US federal debt has surpassed $40 trillion, and the 10-year Treasury yield has risen to about 4.8%; once it "decisively breaks above 5%"—the upper bound of the range since the dot-com bubble era—the AI bubble could burst.

The crowding-out effect from the government issuing debt at high interest rates is pushing up the bond and equity financing costs for AI companies. Annual revenue from AI applications is estimated at about $200 billion, just a fraction of the over $1 trillion spent on infrastructure such as data centers. The gap relies heavily on external financing, and a surge in long-term interest rates will directly impact this pace.

Current market pricing implies an assumption that "AI investment is not constrained by macro interest rates," but every major bubble over the past 300 years has ended with a significant rise in borrowing costs for core companies. This time, the source of risk has shifted from corporations to the government. Once the 5% threshold is triggered, it could lead to a narrative reversal.

Debt Excess Falls on the Government: The Bubble's Structure Has Changed

Looking back 300 years, every major bubble burst began with a significant rise in borrowing costs for core companies, from the 19th-century railroad bubble to all bubbles in the modern central bank era. Typically, during market frenzies, companies borrow heavily to bet on hot themes, inflation and interest rates rise accordingly, and eventually the bubble bursts. Only then does the government step in to take over debt and stimulate the economy.

The difference this time is that the government continues to stimulate even when the economy is doing well. Since the 2020s, the US fiscal deficit has consistently accounted for about 6% of GDP, more than double the average of previous decades. During the same period, households and corporations did not significantly increase leverage for a long time. It was only in the past year that tech giants, after depleting their cash surpluses, began borrowing heavily for massive AI infrastructure—but their leverage levels remain manageable for companies of such size.

Sharma argues that in this cycle, the substantial excess borrowing has so far accumulated on the government's balance sheet, and the problem starts precisely where the excess is deepest.

Crowding-Out Effect Emerges: 5% Becomes the Life-or-Death Line for Financing

The expansion of government debt is creating a crowding-out effect in the bond market. The outstanding balance of US public debt has risen to $40.05 trillion, breaking the $40 trillion mark for the first time. Interest expenses for this fiscal year reached $1.17 trillion. The 30-year Treasury yield climbed to 5.32%, the highest since 2007. Public debt interest payments have doubled over the past five years to over 3% of GDP, a record for the US and the fastest increase and highest level among major developed economies.

Fiscal concerns combined with rising energy prices are pushing up global bond yields. The elevated risk-free rate is in turn raising financing costs for AI companies—which account for the largest share of newly issued corporate bonds. Sharma identifies the 10-year Treasury yield as the key indicator to watch: once it "decisively breaks above 5%," it signals the start of a tighter monetary era, making financing for mega AI projects more difficult. When large tech companies must compete for funds with government bonds yielding over 5%, corresponding to inflation-adjusted returns above 2.5%, many companies will be crowded out of the bond market.

Expectation Gap and Valuation Threat: A Constraint the Market Has Not Priced In

Annual revenue from AI applications is currently estimated at about $200 billion, just a fraction of the over $1 trillion spent on infrastructure such as data centers. AI companies increasingly rely on new bond and equity issuance to fill the gap, and a 10-year yield breaking above 5% would simultaneously slow both channels—a level that would also exceed the earnings yield of US stocks, which has historically been a headwind for the stock market.

If the 10-year yield breaks above 5% before November, it would mean a rise of more than 75 basis points within six months. Historically, surges of this magnitude have ended bull markets. Some analysts argue that breaking 5% is merely a return to the 1990s—when the 10-year yield remained above 5% throughout and US stocks were strong. But back then, the US was far less dependent on debt: that decade ended with a fiscal surplus, whereas today public debt is approaching 100% of GDP, and debt servicing costs are much higher than before. Rising public borrowing costs will squeeze other borrowers more quickly and hit the bubble-inflated AI market harder.

Going forward, close attention must be paid to how quickly the 10-year Treasury yield breaks above 5% and the threat it poses to AI capital spending and US stock valuations.