‘Hard to recover the investment’: Gita Gopinath warns $1 trillion AI bet could correct, flags US recession risk

Gita Gopinath has warned that the AI investment boom could eventually face a market correction, as companies may struggle to generate enough profits to recover the $1 trillion being poured into the technology in 2026 alone.

AI has emerged as one of the biggest investment themes, with companies committing huge sums to infrastructure and development, while some are also cutting jobs to redirect resources towards AI.

Investors, meanwhile, have been chasing exposure to the theme, particularly in markets such as the United States, South Korea and Taiwan, which have benefited from the surge in the technology's demand.

"I personally think that there will be a correction," the Harvard professor and former chief economist of the International Monetary Fund (IMF) said while speaking on Raj Shamani's podcast on Saturday.

However, Gopinath did not specify when such a correction might happen. “It's impossible to tell when that correction will happen,” she added.

Why a correction may take place?

Gopinath argues that the two opposing forces are keeping global growth broadly stable: the AI boom is driving investment and growth, while geopolitical tensions, tariffs and the US-Iran conflict in West Asia are weighing on economies. As a result, global growth growth projections for 2025 and 2026 remain close to pre-shock estimates, masking major changes beneath the surface.

"We're in a world where everything has changed and yet somehow nothing has," she said. "If you look at just the overall numbers, you really will feel like, wow, much ado about nothing. But if you look underneath that, under the surface, that's when you really see the changes playing out."

The divide is particularly visible among economies linked to the AI supply chain. Taiwan and South Korea are growing at roughly twice their usual pace. To illustrate the scale, Gopinath said it would be like India is suddenly growing at 5% instead of its usual 7.5%.

Taiwan's statistics agency has raised its 2026 growth forecast to 11.05%, the fastest annual growth since 1987, citing AI demand. Meanwhile, West Asia economies, such as Saudi Arabia and the UAE face growth downgrades due to energy and US-Iran conflict-related pressures.

The United States is still growing at or slightly above expectations, with AI investment offsetting some of the drag from tariffs.

She described the arrangement as a gamble because the world is more interconnected than ever, with trade accounting for its highest-ever share of global GDP, even as geopolitical power becomes more dispersed than it was during the Cold War. The US, China and India now shape the global landscape, but there is little coordination among them on AI safety.

"We're completely in this gamble," she said. "We're rolling the dice. The world is rolling the dice right now."

Why a correction in AI may mean a US recession

Gopinath's primary concern is that the returns may not justify the scale of investment. With intense competition across the industry, companies and AI labs expected to spend nearly $1 trillion globally in 2026 will need to generate sufficient revenue and profits to recover that investment.

"It's hard to explain how each of these labs and companies who are investing close to a trillion dollars now globally in AI just in 2026 will be able to make the kind of profits that are needed to recover that investment," she said.

This circular flow of money is what could allow a correction in one part of the AI ecosystem to spread across the wider market. NVIDIA, a listed company, puts money into AI labs, which use it to buy NVIDIA chips.

"If NVIDIA were to collapse or any one of the AI labs were to collapse, that would have implications for NVIDIA," she said. “Those kind of linkages we don't see.”

The macro consequence is direct. Gopinath warned that a sharp stock-market correction could push US growth to zero or slightly begative, triggering a recession, although one that would likely be short-lived if driven purely by market losses. The irony is that the AI boom is currently supporting the US economy through wealth gains and increased spending.