MIT report dropped AI companies stock prices

MIT has published a report that has seriously cooled the hype around generative AI on Wall Street. The document, titled The GenAI Divide: State of AI in Business 2025, showed that the promised “gold rush” has not yet brought tangible returns to business.

The study is based on 150 interviews with top managers, a survey of 350 employees and an analysis of 300 AI implementations. The results were disappointing: less than 10% of pilot projects yielded a noticeable increase in revenue, and only 5% brought serious profits. The other 95% of initiatives had no impact on the company’s performance.

This caused an immediate reaction on the stock exchange. Nvidia shares sank by 3.5%, Arm Holdings – by 3.8%, Palantir lost almost 9%. Meta shares also came under pressure after news of the development of an AI division.

The author of the report, Adit Chalapalli, notes that the problem is not in the models themselves, but in their application. According to him, the greatest effect is demonstrated by small AI startups that work on specific tasks together with clients: their revenue in some cases has grown from zero to $20 million per year.

MIT also points to the inefficient distribution of investments: more than half of the budgets go to marketing and sales, while the real profit is provided by the “boring” automation of the back office and the rejection of outsourcing. At the same time, the use of external specialized solutions turned out to be successful in 67% of cases, while in-house developments yielded results in only a third of projects.

The words of OpenAI head Sam Altman were also heard in the background, who warned about the formation of an “AI bubble.” In his assessment, “some investors will lose a lot of money, but in the long run, society will benefit.”

Taken together, the MIT report and Altman’s statements have reinforced analysts’ doubts: the hype around generative AI increasingly resembles overheating, which could cost the market dearly.