Steve Eisman: The AI investment race could end in disaster like in 1999

Investor Steve Eisman, famous for predicting and profiting from the 2008 mortgage crisis, now raises his voice about the risks of artificial intelligence. After years of silence regarding the tech sector, Eisman returns with a strong forecast: the AI market could be heading toward a collapse similar to the dot-com bubble. His analysis, shared on his YouTube channel, is based on a concerning historical parallel that the entire industry should seriously consider.

Over $300 billion in CapEx spending without proven results

The concern expressed by Steve Eisman starts with an alarming number: major tech companies—Meta, Google, Amazon, and other giants—are collectively investing over $300 billion in capital expenditures (CapEx) related to AI development. This figure represents an unprecedented investment in a technology whose economic viability remains uncertain.

The paradox of this situation is that, while spending is growing exponentially, fundamental questions remain unanswered. What will be the actual return on this massive expenditure? Will these projects generate the expected value, or are we witnessing a repeat of historical patterns of uncontrolled speculation?

Forgotten lessons of 1999: How speculation devours profitability

Steve Eisman draws a historical parallel that is uncomfortable for the current tech sector. In the late 1990s, during the peak of the dot-com bubble, financial analysts confidently proclaimed that the internet would conquer the world. Their prediction was correct—internet indeed revolutionized society—but the timing was completely off.

The speculative gold rush led to monumental overinvestment between 1997 and 1999. Companies invested “too much, too fast,” in Eisman’s words. When the bubble burst in 2001, the result was a deep recession that not only destroyed companies but also kept the tech market stagnant for several years afterward, even after the economy began to recover.

Could the same happen with AI investment? Eisman suggests it could, as the analogy is evident, though he emphasizes the inherent uncertainty of any prediction. Speculative cycles tend to repeat because market participants often forget past lessons.

ChatGPT 5.0 doesn’t surprise: Early signs of innovation slowdown

A key element in Eisman’s analysis is the apparent slowdown in the pace of innovation. Although he admits that AI development is not his specialty, he cites serious industry critics who argue that the current development model—mainly based on scaling large language models—is reaching its limits.

Evidence of this is tangible: the recent release of ChatGPT 5.0 has not represented a significant leap over its predecessor, ChatGPT 4.0. Improvements are incremental, not revolutionary. This pattern suggests that the frontier of innovation through pure scalability may be hitting a ceiling—a problem developers have yet to convincingly solve.

If this trend is confirmed, the justification for maintaining the dizzying spending would weaken considerably. Investors would start questioning whether the invested money is generating the expected returns.

The risk of a “painful digestion period”

Eisman’s final projection is clear: if investment returns disappoint in the coming years, AI spending will experience a sharp slowdown from its current pace. Companies will halt ambitious projects, cut budgets, and seek profitability rather than speculative growth.

What would follow is what Eisman calls a “painful digestion period,” similar to what the tech industry experienced after 2001. During those years, even surviving companies were forced to demonstrate real profits and sustainable economic viability, not just future promises.

For investors and sector employees, this would translate into significant volatility, company consolidations, and a radical reevaluation of valuations. AI startups attracting unlimited investment today could disappear in contraction cycles. With his track record of accurate crisis predictions, Eisman suggests that preparing for this scenario is not paranoia but financial prudence.