The AI Bubble May Shrink, Not Burst

This summer has been turbulent for artificial intelligence stocks, as China’s advances in the field and other factors have driven down U.S. equities, particularly Nvidia and Micron. Pessimists quickly declared that the AI bubble was on the verge of bursting.

Their concerns are not entirely unfounded. Roughly 45% of the market capitalization of the S&P 500 index is tied to AI-related stocks, while the seven largest U.S. technology companies account for more than a third of the index’s value. Valuations have reached levels reminiscent of the dot-com bubble burst in 2000.

However, I do not believe history will necessarily repeat itself. The problem may not be a bubble bursting, but rather that AI is advancing at a slower pace than early expectations suggested. What we are witnessing may be a recalibration of market expectations, not a collapse of the technology.

For years, technology leaders have made exaggerated predictions, such as eliminating half of all office jobs or creating a permanent class of unemployed people. But the limits of AI have become clearer. Some companies that laid off employees in hopes of replacing them with AI have begun rehiring them, and a study by the Massachusetts Institute of Technology found that 95% of experimental generative AI projects fail to deliver measurable returns on investment.

Public skepticism has also grown. Data centers have sparked local opposition, while fears of job losses and poor-quality AI-generated content have fueled public discontent. Expectations have outpaced reality.

Yet this does not mean the technology is a failure. In areas such as data extraction, research, business automation, and customer service, AI is already making a tangible difference. The U.S. government uses it to detect fraud and expedite permit approvals, the Pentagon has adopted it, and Ukraine has demonstrated its utility in war.

Moreover, competition with China is accelerating innovation, while technology companies are working to address issues such as the water consumption of data centers.

This is, in my view, the future of AI: gradual adoption that resolves its challenges over time and integrates into society, without triggering an economic catastrophe. Markets are not collapsing; they are simply recalibrating their expectations to align with this reality.