Steve Eisman warns about overinvestment in artificial intelligence by big tech companies

Steve Eisman, the legendary investor who predicted the 2008 financial crisis and made significant profits from the collapse of the subprime mortgage market, has once again raised concerns in the markets. This time, his worry centers on the current massive spending on artificial intelligence by global tech corporations. The analyst argues that the scale of current investments bears alarming similarities to past speculative bubbles, particularly the tech bubble of 2000 that ended in a recession.

The 1999 historical lesson: a pattern that could repeat

To understand Steve Eisman’s perspective, it’s necessary to go back to the late 1990s. During that period, Wall Street analysts promoted the internet as the revolutionary technology that would conquer the world. Over time, that prediction proved correct. However, the speculative euphoria led to a frantic investment race: too much capital, invested too quickly, without immediate returns to justify the level of spending. The inevitable consequence was painful.

The excessive allocation of resources was largely the main trigger for the 2001 recession and the prolonged weakness of the tech stock market in subsequent years. During those years, tech sector stocks remained stagnant, unable to generate profits. Steve Eisman draws a potential analogy between that cycle of overinvestment and the current state of the artificial intelligence sector.

Massive CapEx spending: a recent unprecedented phenomenon

Major tech companies—Meta, Google, Amazon, among others—are collectively channeling over $300 billion in capital expenditures (CapEx) related to AI development. All of them aim to position themselves at the forefront of this technological transformation, creating a competitive investment dynamic with no apparent end. The magnitude of these resources is undoubtedly extraordinary, but the fundamental question remains: do these outlays justify the expected returns?

Early cracks in innovation: a concerning slowdown

Steve Eisman points out that there are already early indicators suggesting that the pace of innovation in AI may be losing momentum. Although he admits this is not his area of expertise, he cites observations from specialized critics warning about the limits of the current AI development model. Specifically, the prevailing strategy of continuously scaling large language models appears to be reaching diminishing returns.

A concrete indicator of this possible slowdown is the performance of ChatGPT 5.0, recently released, which according to multiple evaluations has not shown a substantial improvement over its predecessor, ChatGPT 4.0. This relative stagnation contrasts with the optimism that characterizes investments in the sector.

The return question: heading toward an inevitable correction

The fundamental uncertainty posed by Steve Eisman is as simple as it is decisive: no one can predict with certainty what the return on investment will be for this massive deployment of resources in AI. If the yields turn out to be disappointing in the early stages, the direct consequence will be a slowdown in the current rapid investment pace. This scenario would lead to a period of market contraction and adjustment—similar to what the world experienced in 2001—characterized by painful corrections and reevaluation of expectations.

Steve Eisman’s analysis, based on documented historical lessons, emphasizes that tech speculation cycles follow recognizable patterns. The warning is not that artificial intelligence will fail, but that the unsustainable pace of investment could generate significant disillusionment if returns do not materialize as expected.