US is leading the global artificial intelligence revolution, but slowing growth and persistent inflation are raising questions about the AI-led expansion
Our Bureau
New York, NY
The artificial intelligence boom has become one of the biggest sources of optimism for the US economy, driving enormous investment in semiconductors, data centers, cloud computing and digital infrastructure while pushing technology valuations and corporate profits higher. Yet the latest economic data suggest that the extraordinary AI investment cycle is not translating evenly across the American economy, raising the possibility that the country may be entering an increasingly unequal, two-speed recovery.
US real GDP growth slowed to an annualized 1.5 per cent in the second quarter of 2026, from 2.1 per cent in the first quarter, according to the Bureau of Economic Analysis. The BEA said the deceleration reflected “a downturn in government spending and decelerations in investment and exports.” At the same time, inflation pressures remained elevated, with the PCE price index rising 5.3 per cent and the core PCE index increasing 3.6 per cent. The combination is uncomfortable: growth is slowing, but inflation has not fallen sufficiently to give policymakers an easy path towards monetary easing.
There is, however, an important distinction within these numbers. Underlying private domestic demand remains relatively strong, with real final sales to private domestic purchasers increasing 4.2 per cent in the second quarter. Corporate profits also rose sharply, increasing by $400.9 billion during the quarter compared with just $74.4 billion in the first quarter. Much of the optimism surrounding the American economy therefore comes from the strength of corporate America and, increasingly, from technology-led investment.
The concern is whether that strength can spread to the rest of the economy.
Nuvama’s analysis is particularly relevant because it argues that “US real GDP growth excluding technology capex is particularly weak”. The implication is that artificial intelligence has become a powerful investment engine without yet producing comparable spillovers across traditional sectors. Retail activity in the US, Europe and China remains subdued, while real estate activity across major economies is still close to post-global financial crisis lows.
This creates a two-speed global economy in which chip-producing countries such as South Korea, Taiwan and China benefit disproportionately from the AI investment cycle, while traditional sectors remain much weaker. The United States is at the centre of this process because its technology companies dominate many of the most valuable parts of the AI ecosystem, but that leadership also creates a potential vulnerability if investment expectations begin to outrun economic returns.
Nuvama warns that a slowdown in global AI capital expenditure could trigger “a reversal of capital flows”, with Asian savings that have been recycled into US equities and other assets returning to the region. The current US current-account deficit and dollar strength have been supported partly by this recycling of Asian savings into American financial markets. If the AI boom weakens, Asian technology exports could also slow, reducing the flow of capital back into the United States.
That could create an unusual problem for Washington. A weaker AI cycle could result in a weaker dollar and continued pressure on US Treasury yields even as economic growth slows. Normally, slower growth would be expected to push interest rates and bond yields lower. But if foreign demand for US assets weakens at the same time, Treasury yields could remain elevated, complicating the Federal Reserve’s response to an economic downturn.
The potential gains are nevertheless enormous. McKinsey estimates that AI could unlock around $65 billion in annual recurring value in the global upstream oil and gas sector using technologies available today, with a “credible path to $230 billion” as AI matures and autonomous operations become more common. The report identifies production optimization, drilling, reservoir management and predictive maintenance among the areas with the greatest potential.
But McKinsey also cautions that AI is a “concentration play”, with the top 10 use cases accounting for nearly half of the identified value. The lesson for the wider economy is that technological adoption will not automatically benefit every company or sector equally.
America therefore faces a paradox. It has arguably never been better positioned to lead a technological revolution, yet the wider economy has not fully demonstrated that it can convert that technological leadership into broad-based growth.
The next stage of the AI story will consequently be more important than the first. Building data centers, buying chips and investing billions in AI can generate spectacular financial numbers, but the real measure of success will be whether the technology raises productivity across the economy, supports wages and consumption, creates new industries and strengthens America’s underlying growth potential.