Workers’ share of America’s income is at a record low before the AI boom even begins

But analysts are starting to wonder: richer for whom? Workers’ share of U.S. income has already fallen to its lowest level on record, while corporate profit margins keep breaking records quarter by quarter.

According to Gregory Daco, the EY-Parthenon chief economist, the productivity gains that explain that divergence largely predate the AI boom. “Productivity growth protects margins, not income,” Daco wrote in a note Thursday. 

Economic output grew 1.7% in the second quarter, just based on 0.3% more hours. Compensation rose 2.6%, which, set against a spring and summer of oil-driven inflation, comes out to “flat to slight contraction” in real terms, Daco told Fortune in an interview. 

Margins hit a record 14.9% of GDP, while the labor share fell to 52.8%, the lowest since the government started counting in 1947. Daco said that 50% isn’t a floor. “As long as you continue to see concentrated gains on the capital side, and within a certain number of firms,” labor’s share could keep plummeting, he said. 

The productivity behind those numbers, after all, is a decade of good old automation, some cost discipline (hiring pulling back after some post-pandemic bloat), and capital spending, as opposed to AI. All AI has delivered so far is further concentration. 

“You tend to have greater concentration and more of a winner-takes-all type of environment when you have these technological advances,” Daco said. In almost every technological revolution—the railroad boom of the late 19th century, or the 90s dot-com revolution—large, vertically integrated firms initially capture the gains, while smaller ones face “persistent cost pressures, persistent policy uncertainty, higher interest rates,” Daco noted. 

In the 90s, a handful of companies at the technological frontier front-loaded the capital investment—and thus reaped the capital gains—but the productivity growth through cheaper software spread quickly throughout the economy, and wage growth followed. But there is no guarantee that AI follows the same timetable. 

After all, the boom is uniquely, historically, capital intensive. Data center investment is expected to reach $31 trillion, nearly the size of our current GDP, by 2050, according to analyst firm PricewaterhouseCoopers LLP. While the engine sputters out in other sectors, construction and manufacturing is roaring now because of data centers; otherwise, the industry would be in recession, a Chicago manager said in the Federal Reserve’s Beige Book this week.

That sounds a lot like growth; companies spending hundreds of billions on equipment that should eventually allow the economy to produce much more with less. But much of that equipment isn’t made in America.

Imports of the large computers used in AI servers have exploded over the past year. Net imports of “large computers”—the Census category for GPU servers—hit a $450 billion annualized pace last month, an astonishing increase from roughly $50 billion a year through 2023, per Census data compiled by economist Joseph Politano. The GDP accounting works out such that an imported server adds to investment and subtracts as an import in the same amount. So the net contribution to GDP: zero.

That helps to explain an element of the AI economy so far. While capital spending is booming, productivity is improving, corporate margins are enormous, and conditions are “loose,” as Warsh points out, hiring is weak, housing is struggling under tight rates, and the share of income to workers keeps shrinking.

“While U.S. investment is booming, growth in gross domestic product has been modest,” wrote Jon Hilsenrath, the former Wall Street Journal Fed reporter who now advises hedge funds at Serpa Pinto Advisory.

That raises a tough question for Warsh and Bessent: let it rip, or do something about it?

Growth is not the same thing as broadly distributed income: if every dollar of output increasingly accrues to the owner of a data center, just passively collecting checks, or to a shareholder who owns the data center owner, then the fiscal math gets complicated. The economy may be getting richer while the tax base and political constituency that policymakers usually associate with a boom grow much more slowly. It’s not hard to imagine that fanning the flames of suspicion about the AI buildout and its benefits.

And the investment itself can and does carry costs and risk. Hundreds of billions of dollars of AI spending competes for capital in an economy where borrowing is getting more expensive. Higher long-term rates make mortgages expensive and suppress homebuilding, as the WSJ demonstrated in a startling chart this week.

That doesn’t mean that the AI productivity boom will fail; it just means that it’s not immediately obvious how it boosts labor’s share. “I don’t think there’s a floor,” Daco said.