U.S. Companies Lead AI Adoption but Struggle With Costs

American companies are racing ahead of global competitors in artificial intelligence adoption, but the financial reality of that leadership is proving more complex than anticipated. New data shows the gap between enthusiasm and economic return is widening as firms confront unexpectedly high costs.

Adoption outpaces the world

Approximately half of U.S. workers now use AI in their jobs at least several times annually, up from under 40% a year ago, according to recent Gallup polling. That puts American workers well ahead of their European counterparts, where only 32% report using AI on the job, research from the Brookings Institution found in March.

The disparity extends to the enterprise level. Seven percent of U.S. companies now deploy AI for producing goods and services, compared to just 4% of European firms. By May, more than half of American companies were paying for AI subscriptions, up from 46.8% at the start of the year, according to data from expense-management platform Ramp.

Why it matters

The productivity gains from AI remain modest—Brookings calculated time savings worth 2.3% of working hours in the U.S. versus 1.4% in Europe—but the costs are mounting rapidly. Companies that moved fast to adopt AI are now discovering that being first means being first to solve the profitability puzzle, with billions at stake and no clear playbook.

The budget reckoning arrives

What appeared to be low-cost experimentation has turned into a financial challenge. Uber reportedly exhausted its entire 2026 AI coding budget in just four months, driven largely by Claude Code usage. The company's chief operating officer, Andrew MacDonald, acknowledged last week that "the link is not there yet" between heavy AI spending and profitable consumer products.

Corporate demand for tokens—the units measuring AI usage—has exploded. Google CEO Sundar Pichai recently announced the company processes 3.2 quadrillion tokens monthly, a sevenfold increase from a year earlier. At tech firms where AI use is most prevalent, some employees now engage in "tokenmaxxing," competing to rack up the highest token counts.

The Wall Street Journal reported that Meta, Microsoft, and Salesforce are pushing employees to use AI more efficiently or limit usage altogether. Axios cited one unnamed company that spent half a billion dollars on AI after failing to implement usage caps.

Returns remain elusive

Despite the costs, return on investment remains unclear. An Accenture survey of U.K. firms published last month found 90% have yet to successfully integrate AI with their business operations. Productivity improvements have largely benefited individual workers rather than generating company-wide efficiency gains.

Some economists suggest AI could be creating up to $250 billion in economic activity that traditional metrics cannot capture, according to research from the Peterson Institute on International Economics. But those theoretical gains have yet to materialize in corporate financial reports.

American executives may be further along the adoption curve than their international peers, but that advantage comes with a distinct challenge: being first to determine whether the technology justifies its price tag. These details were first reported by Fortune.