What if the most dangerous leverage in markets is not in Silicon Valley’s balance sheets, but in Washington’s? That is the uncomfortable logic behind Ruchir Sharma’s warning: the debt binge that matters most is the one parked on the government ledger, because public borrowing can quietly rewrite the price of money for everyone else. Markets often hunt for fragility in the wrong place. They stare at corporate exuberance, then discover that the real weak link was the rate structure all along.
The point is not that debt is new. Debt is as old as power, and power has always preferred to spend today and explain tomorrow. The difference now is scale. Sharma, chairman of Rockefeller International, argues that the current excess sits mainly on the government balance sheet rather than on corporate or household books. That matters because once sovereign borrowing becomes too heavy, the state itself can start to crowd out the cheap capital that fuels speculative booms. In that sense, the government can become the hidden short seller of the market cycle.
The threshold he singles out is stark: a decisive breach of the 5% 10-year Treasury yield. Around the time of his warning, the 10-year was cited at roughly 4.8%, with the 30-year at 5.32%, the highest since 2007. Those are not just numbers on a screen. They are the market’s vote on whether inflation, borrowing, and growth can still be balanced without breaking the furniture. Once long rates move from pressure to punishment, every stretched assumption must reprice.
Sharma’s claim is not that 5% is a magical line drawn by the gods. It is that bubbles often die not from disbelief alone, but from the cost of money rising enough to expose the financing behind the story. He put it plainly: “If you end up getting the 10-year yield above 5%—if you look at every single bubble in history, it requires higher interest rates for those to burst”. That is an old lesson in a modern costume. Speculation can survive ridicule longer than it can survive a tighter discount rate.
He added another warning: “What we’ve seen just now is only a trailer… that could be the bill which signals that the market as a whole can’t finance this kind of borrowing”. The word trailer is apt. In markets, the opening scene is usually harmless. The danger arrives later, when the financing bill arrives and the audience realizes the movie was never really about the product. It was about the cost of carrying the dream.
The AI story is often sold as a clean contest between progress and pessimism, but Sharma’s framing is more austere. He notes that AI application revenue is estimated at about $200bn annually, while more than $1tn is being spent on data-center and infrastructure buildout. That leaves a financing gap that depends on debt and equity issuance. In other words, the machine that is supposed to deliver the future must first be funded by today’s capital markets. If rates climb, the whole structure becomes less like a rocket and more like a bridge under load.
This is where investor psychology becomes brittle. When people hear “AI,” they think of software-like margins and winner-take-all economics. But the infrastructure race is much closer to heavy industry than to pure code. Data centers consume capital, land, power, and financing. Those are not airy promises; they are concrete claims on cash flow. History is full of elegant narratives that collapsed because the bill for scale arrived before the profit engine did. Railroads, telecoms, and the dot-com era all learned that lesson in different accents.
The government’s role makes the current cycle stranger than the usual private-sector binge. US federal debt has crossed $40 trillion, and budget deficits have run about 6% of GDP this decade, more than double the prior multi-decade average. Public-debt interest payments have more than doubled in five years to above 3% of GDP, a US record. That means debt service is no longer a side note. It is becoming part of the economic weather. Once interest costs take on that kind of weight, they reduce room for policy, investment, and error.
This is the hidden asymmetry. A corporate bubble can burst and burn only its own capital structure. A sovereign borrowing surge can change the price of capital for the entire system. The state does not merely borrow; it sets the benchmark. If that benchmark rises persistently, risk assets must justify themselves against a higher hurdle. The market likes to believe it can float above gravity, but gravity is negotiated through yields.
Sharma’s other claim is that a 75-basis-point rise in the 10-year yield within six months has historically coincided with the end of bull markets. That is not a prophecy; it is a pattern. Markets love to confuse patterns with comfort, but pattern recognition is only useful if it makes you less sentimental. The relevant question is not whether this time is identical to prior cycles. It is whether the financing architecture today can tolerate the same kind of rate shock that ended earlier expansions.
Game theory helps here. Every participant in a boom prefers to believe someone else will blink first. Investors want growth to continue. Borrowers want cheap funding. The government wants the debt burden to remain manageable. But when yields rise, the equilibrium shifts. The first actor to retreat may suffer less than the last. That is why bubbles are social phenomena as much as financial ones. They depend on coordination, and they die when coordination fails.
The signal to watch is not any single day’s volatility. It is whether the 10-year Treasury yield decisively breaches 5%. That is the line Sharma identifies, and it is also the level that would test the market’s appetite for long-duration assets built on future earnings. If the yield keeps climbing, the discount rate eats into every optimistic valuation model. Promises far in the future become less valuable when the present demands a larger toll.
There is a deeper irony in the AI trade. The technology is often described as transformative and antifragile, yet the funding structure around it may be fragile in the old-fashioned sense. The more capital-intensive the race becomes, the more dependent it is on stable financing conditions. If public debt pushes rates higher, the market may discover that the most advanced sector in the economy is still governed by the oldest constraint in finance: the cost of capital.
The common mistake is to think fragility lives only in debt-laden companies. In reality, fragility often migrates upstream. When the sovereign borrows too much, it does not merely weaken itself; it changes the terrain under every leveraged strategy. Investors then act surprised when a macro issue appears inside a micro story. But markets are not separate compartments. They are interlocking beams. Bend one long enough, and the load shifts everywhere else.
That is why Sharma’s warning is worth taking seriously even for people who never trade Treasuries. A debt-heavy state can accidentally become the catalyst that ends a speculative era elsewhere. The market does not need a dramatic collapse in earnings to correct. Sometimes it only needs the price of money to reassert itself. Interest rates are the slow poison of overconfidence. They do not shout. They seep.
The useful habit in markets is not optimism or pessimism, but inversion. Ask not what could keep the AI story alive, but what would make it harder to finance. Ask not whether a boom is exciting, but whether its capital stack can survive a yield repricing. On Sharma’s reading, the answer becomes uncomfortable once the 10-year gets through 5%. That is the kind of level at which the bond market stops being a background actor and becomes the plot.
If the current debt binge has a real consequence, it may not show up first in government spreadsheets. It may show up in the cost of funding the next great technology wave. That is the paradox of public borrowing: the bill is issued in Washington, but the stress test lands in every corner of the market.
