Nobel Economist Stiglitz Warns: AI Bubble Set to Burst, Prepare

Nobel laureate economist Joseph Stiglitz has just issued a warning that the current wave of AI investment is a bubble about to burst. He suggests that approximately one-third of last year's U.S. economic growth was fueled by AI investment — and when this bubble pops, the repercussions will be severe. The core problem: the market is betting that AI will be both technologically successful and face minimal competition — an assumption that is hard to uphold. Stiglitz also cautioned that the U.S. entirely lacks the infrastructure for worker retraining to address a potential wave of widespread layoffs. Nevertheless, he believes that in the long term, AI will support humans in working more effectively rather than replacing them.

Stiglitz is not one to cry wolf. He has witnessed enough economic cycles to distinguish between short-term shocks and long-term structural shifts. And this time, he states plainly: we are facing both simultaneously.

His bubble argument is actually quite simple. The enormous profits investors expect from AI depend on two concurrent conditions: the technology must be successful, and the market must have little competition. The problem is that these two conditions contradict each other. If AI is truly powerful and effective, the whole world will pour into it — from Google, Microsoft to Alibaba, DeepSeek. In that scenario, profit margins will be squeezed tightly. Wall Street is pricing in a technology monopoly scenario that is unlikely to materialize in reality.

The more worrying part is about labor. Stiglitz reiterates the lesson of the Great Depression: a surge in agricultural productivity displaced millions of farmers, but society had no mechanism to absorb them. It wasn't until World War II created massive demand for jobs that the problem was solved — through extremely strong government intervention. Today, no world war is waiting to rescue the labor market.

What Stiglitz points out, which few are willing to admit: the most vulnerable group this time is not factory workers but university-educated office workers — those who still consider themselves safe. Jobs involving analysis, drafting, and record processing — precisely what AI does best.

However, Stiglitz is not simply a pessimist. He points out that sectors representing the largest share of the economy — education, healthcare — cannot be replaced by AI for structural, not technological, reasons. Teachers require human-to-human interaction. The U.S. healthcare system is inefficient due to politics and monopolies, not a lack of data. AI cannot solve political problems.

The image of a plumber using AI to diagnose a problem is the best metaphor from the interview. The future is not about robots replacing humans. The future is about humans using AI to work smarter — if we survive the transitional period he calls "reallocation."

The real question is not whether AI is powerful. The question is: who will bear the costs of that transition?

Vietnamese-American engineers and programmers in Silicon Valley find themselves at the nexus of both opportunity and risk. Many are working directly in the AI industry or at companies heavily investing in AI. If the bubble bursts as Stiglitz describes, the next wave of layoffs in Big Tech could be even more severe than the 2022-2023 round — and this time targeting both AI engineers and positions in analysis and operations that are being automated.

Second-generation Vietnamese — many of whom are pursuing degrees in economics, law, finance, and marketing — particularly need to heed this warning. Stiglitz explicitly states that university-educated office jobs are precisely AI's primary target. The complacency of "I have a degree, so I'm safe" could prove to be a generational mistake.

H-1B visa holders awaiting green cards are the most vulnerable group. In an already tightened tech labor market, should the next wave of layoffs occur, they will have little time to secure new employment before losing their legal residency status. This presents a very concrete, not theoretical, risk.