This year’s surge in oil prices, Canadian tariffs, and pre-midterm political acrimony have not slowed the U.S. economy. Quite the opposite: It appears to have gained momentum in the second half of 2026.

That may be hard to see in the nation’s mediocre headline GDP numbers, but only because the 2.8% pace of growth in final demand—the total purchases of businesses, governments, and consumers—was offset by an inventory drawdown and an increase in imports. But demand leads supply, and the Atlanta Fed’s GDPNow estimate currently puts third-quarter growth at a blistering 5%.

That estimate is almost certainly too high; GDPNow tends to run hot when the underlying data are still thin. Even so, the U.S. economy is clearly showing strength. Politically, Republicans could see a welcome October surprise when the first estimate of third-quarter growth is released shortly before the midterms—especially if the recent improvement in job growth continues.

Alas, the foundations of this economic acceleration are rotten. Growth is being driven largely by an inflating AI technology bubble that is generating trillions of dollars in new “wealth,” even as household incomes are being subsidized to the tune of $6,000 per person by the yawning federal budget deficit. This is an overheating economy, not a decelerating one, as evidenced by falling savings rates and a widening trade deficit.

But the billion-dollar question is when will it end? The recent sea change in the bond market suggests that the reckoning may be approaching. The 10-year U.S. Treasury yield has risen 100 basis points since the start of the year, reaching its highest level in two decades. Oil-driven inflation risk explains only part of that increase, and the Fed’s recent move to cool inflation by raising short-term interest rates has not had the desired impact on the long end of the yield curve.

Over the last year, Beacon Economics has been correct in our assertion that neither tariffs nor the oil shock would derail the U.S. economy—but the bond market is a different issue altogether. What we’re seeing is the economic equivalent of a barometer dropping rapidly on a sunny, late-summer afternoon at the beach. The sky may still be clear, but it’s time to prepare for a sudden change in the weather.

An Overheating Economy

Despite weak consumer sentiment and high gasoline prices, U.S. consumer spending rebounded from a soft start to the year, contributing 2.1 percentage points to overall GDP growth in the second quarter. The problem is that households appear to be spending money they do not have, given that real disposable income has been essentially flat over the past year.

The personal saving rate has fallen from 5% in 2024 to 3%, close to its lowest sustained level since the overheated subprime run-up to the Great Recession—hardly a reassuring comparison. The trade deficit is also widening, another sign of an economy consuming too much. The immediate cause of the nation’s weak disposable income growth is the recent oil-driven spike in prices, which has offset a 4% year-over-year increase in nominal income. But that 4% income growth is itself weak, given that nominal GDP grew by well over 6.5% during the same period. Instead, national income growth has been dominated by corporate profits, up an astonishing 22%.

Aggregate worker compensation growth has slackened since the beginning of the Trump administration’s aggressive campaign against undocumented immigration. Those policies have contributed to stalled population growth and a 1% contraction in the U.S. labor force—something that has never happened in a full-employment economy.

What population taketh away, the stock market giveth back.

From a per capita spending perspective, wealth accumulation has more than made up for weak income growth. The Federal Reserve’s Financial Accounts show real average U.S. household net worth reaching $1.4 million in the second quarter, the highest level on record. That increase has been driven largely by an approximate $20 trillion gain in financial asset valuations, particularly, though not exclusively, corporate equities.

Is the market in a bubble? Despite the increase in corporate profits, valuations have grown even faster and Robert Shiller’s cyclically adjusted price earnings ratio now exceeds 42, a level surpassed only during the heady final days of the late-1990s “new economy” bubble. The only way to justify such a valuation would be if all this investment were starting to pay off in productivity growth. It hasn’t; worker productivity has actually slowed. 

There is an intense debate over whether AI will usher in a new dawn for humanity or wipe us out. On the more mundane question of its present economic impact, however, there is little evidence that AI is doing much beyond driving investment and inflating equity values.

And it is not just stock prices that are elevated relative to profits. Margin trading is also booming, with equity investors borrowing $1.4 trillion to leverage their investments. This is classic speculative behavior and increases both the likelihood of financial instability and the damage that could follow if the bubble deflates. That is hardly unusual in the history of transformative technologies, whether physical—the railroad, telephone, and internet—or financial, such as subprime lending and margin trading.

As in the 1990s, soaring equity prices are also fueling a surge in business investment in information technology. Investment in hardware and software has contributed more than one percentage point to GDP growth this year, driven by massive spending on AI and the data centers required to operate its systems. This is yet another source of economic overheating.

Whatever AI’s long-run implications, its clearest short-run effect is the creation of a financial bubble. Like previous financial bubbles, this one will eventually burst. Today’s most important question is what happens when it does.

A Collision of Structural and Cyclical Imbalances

While AI is overheating the economy in the short run, the consequences can be understood only in the context of the nation’s central structural problem: an unsustainable annual federal deficit of $2.2 trillion.

That deficit amounts to an annual subsidy of roughly $6,000 for every person in the United States, or about 9% of per capita disposable income. In inflation-adjusted terms, the deficit was larger during both the Great Recession and the pandemic, but the current deficit is unprecedented for an economy operating at or near full employment.

Today’s deficit is the product of a quarter-century of fiscal mismanagement. The nation now owes a record $40 trillion, an amount exceeding 120% of GDP—roughly the same ratio recorded in 1946, at the end of World War II. But the comparison ends there. In 1946, the U.S. economy could look forward to a peace dividend and a baby boom that would help it expand dramatically, bringing down the debt burden relative to GDP. Neither force is at work today.

This is why the current technology bubble is more dangerous than the bubble of the 1990s—not necessarily because it is larger, but because policymakers have far less fiscal capacity to absorb the revenue losses and increased public spending that would accompany a crash. The U.S. government responded to the 2001 downturn from a position of relative fiscal strength. It no longer has that luxury.

Of course, this analysis assumes that a bursting technology bubble would help trigger a fiscal crisis. The causality could run in the opposite direction: An emerging fiscal crisis could cause the stock market to collapse.

The link between the two is interest rates. The United States doubled its debt-to-GDP ratio after the turn of the century, yet federal interest payments remained remarkably stable at roughly 2% to 2.5% of GDP. Rising debt was offset by falling interest rates.

That tide has turned. Higher rates have caused federal interest payments to surge from roughly $500 billion to $1.25 trillion in just six years. And more increases are  coming. The blended interest rate on outstanding federal debt is approximately 3%, while the one-year Treasury bill yields 4.4% and the 10-year Treasury is near 5%.

If rates remain at these levels, federal interest payments could exceed $2 trillion within two or three years, pushing the annual deficit above $3 trillion. That would accelerate debt accumulation, force the government to borrow still more, and place additional upward pressure on interest rates. It’s a vicious cycle that ultimately risks a fiscal crisis.

The Federal Reserve’s recent decision to tighten policy appears to rest on the assumption that rates are rising primarily because the oil shock has temporarily increased inflation risk. Under that interpretation, higher short-term rates should eventually produce lower long-term rates.

But there is little evidence that the increase in yields is strictly related to inflation, as inflation expectations have barely moved despite the rise in oil prices. This suggests that markets understand that oil shocks typically have, at most, a transitory effect on inflation and economic growth.

The recent rise in interest rates likely reflects more fundamental forces: The technology bubble is increasing the demand for capital, while investors—particularly global investors, as suggested by the weakening U.S. dollar—may be gradually recognizing that the federal government’s fiscal trajectory is unsustainable. Bond prices are therefore behaving like a barometer. The further they fall, the more severe the approaching storm is likely to be.

The longer this problem is ignored, the harder the eventual return to earth will be. A crisis is not inevitable. Congress could confront the deficit, although that appears highly unlikely amid the political acrimony surrounding the midterms, the nation’s bitter partisanship, and, more broadly, the “miserabilism” that has overtaken Americans despite record incomes and wealth. The AI boom could eventually produce meaningful productivity growth, although that seems unlikely to happen fast enough to save the nation from its current predicament.

For now, the sun is still shining. The economy is expanding, consumers are spending, equity markets are soaring, and federal borrowing continues.

But the barometer is dropping.