Stiglitz warns AI bubble could burst, urges India to protect its services sector

Speaking at the Hindustan Times Leadership Summit 2026, the Nobel laureate says AI’s ability to replace coders threatens India’s export miracle, even as tech firms struggle to deliver promised returns.

New Delhi: The artificial intelligence (AI) boom resembles a bubble, and its continued rise could spell trouble for India’s services sector, American economist and Nobel laureate Joseph Stiglitz warned on Saturday.

"I hate saying this, but I think you (India) should be worried. If you look at where AI has affected the labour force in the US, what is it good at? It is good at text. It is good at code. AI is really good at replacing coders, programmers. And those are the kind of jobs that it replaces," Stiglitz said at the Hindustan Times Leadership Summit 2026 in the national capital.

"And so, the services export miracle in India has been really based on those particular kinds of jobs. So the question is, how do you pivot? Where do you go?"

Missing economic model

Stiglitz believes AI lacks a viable macroeconomic model. While the technology itself could succeed, the companies that have poured massive investments into AI will not be able to generate the returns they expect, he said.

“And actually, it is an open question, not only in India, but in the US, in all countries. The difference is, in the United States, in principle, we could tax the winners of the AI revolution — the AI companies — and use some of that to support what I would call a progressive AI agenda, where you make investments in people and in society, and make it sustainable,” he said.

Across borders, however, that kind of redistribution will not work, he said.

The Nobel laureate was blunt about the numbers: between 30% and 32% of recent US growth has come from data centres alone—a pace he called unsustainable.

“I think the AI bubble is a bubble,” Stiglitz said. “You can’t tell a bubble until it breaks. But you can be fairly sure.”

Companies have promised returns they cannot deliver and are banking on monopoly power they will never achieve, he said. Competition is already eroding those profits, with some open-source models collapsing costs and US frontier firms themselves training on cheap Chinese models, he noted. The technology will survive, he insisted, even if the companies do not.