U.S. Workers’ Record-Low Income Share Raises Stakes for the Coming AI Productivity Boom

U.S. workers are receiving their smallest share of national income on record, even as productivity and corporate profit margins remain strong. The imbalance has emerged before artificial intelligence has materially driven the economy’s recent productivity gains, raising questions about how the next technology wave will distribute wealth.

The latest pattern combines modest growth in labor compensation with faster gains in output and profits. In the second quarter of 2026, economic output increased 1.7% while hours worked rose just 0.3%, and compensation climbed 2.6%. That gap points to an economy producing more with limited growth in labor input, a development that benefits companies when additional output translates into higher margins.

The productivity improvement now shaping the debate largely began before the current AI investment boom, according to Gregory Daco, chief economist at EY-Parthenon. His assessment separates the economy’s existing efficiency gains from the larger productivity claims attached to generative AI and automated software.

Daco’s central warning is that productivity growth protects margins rather than automatically raising worker income. Treasury Secretary Scott Bessent and Federal Reserve Chairman Kevin Warsh have described an AI-driven productivity surge as a force that can make the economy richer and deflationary, but the labor-share data show that higher output alone doesn't determine who receives the proceeds.

Large, vertically integrated companies typically capture the early gains from major technological shifts, while smaller firms face higher costs, policy uncertainty and expensive financing. That pattern appeared during the railroad expansion of the late 19th century and the dot-com boom of the 1990s, and Daco expects similar concentration pressures from AI.

The result is a winner-takes-all structure in which firms with capital, data, distribution and technical infrastructure can scale productivity gains faster than competitors. Companies that adopt AI without comparable resources still face pressure to invest, cut costs or accept thinner margins. The effect on workers depends on whether firms pass efficiency gains into wages, expand employment or retain the gains as profits.

Artificial intelligence could deepen the divide between capital owners and workers if companies use it mainly to reduce labor requirements or strengthen pricing power. A productivity boom expands the economic pie, but the labor share determines how much of that expansion reaches households through pay.

That distribution question matters beyond individual occupations. If income flows increasingly toward profitable firms and asset owners, consumer demand can weaken even while output rises. Workers also face uneven exposure: highly specialized employees may gain from AI tools, while routine roles face tighter staffing and bargaining pressure. The technology’s economic value therefore depends on institutions, competition and wage-setting as much as on model performance.

Policy decisions will shape whether AI productivity produces broad wage growth or a larger concentration of income. Interest rates, competition enforcement, labor protections, tax rules and investment in worker training all affect how quickly firms can consolidate gains and how much negotiating power employees retain.

The issue also intersects with the federal government’s fiscal outlook. Bessent has argued that stronger productivity and the resulting economic growth can ease pressure from the national debt, while Warsh has emphasized AI’s possible disinflationary effect. Those benefits will be harder to translate into durable household prosperity if the labor share continues falling.

The immediate lesson is that productivity statistics won't settle the distribution debate. U.S. output has already grown faster than labor input, yet workers’ share of income has declined, showing that efficiency gains can arrive without a corresponding increase in labor’s claim on the economy.

The next test will come as AI moves from experimentation into routine business operations. Investors will track margins and output, workers will track pay and job security, and policymakers will track concentration and demand. The record-low labor share gives that transition a clear starting point: AI enters an economy where the gains from productivity are already flowing disproportionately toward companies.