Steve Eisman says the debate over whether artificial intelligence is a speculative bubble misses what he views as the most important fact: demand is already here.
In a new installment of the Steve Eisman podcast, the “Big Short” investor says he has spent significant time thinking through the bubble question and concluded that today’s AI buildout is fundamentally different from past market manias.
“I’ve given this bubble debate a great deal of thought, and here is what I think are the major issues. As for whether it’s a bubble, there is no question. Hyperscalers are buying chips like crazy from Nvidia, AMD, etc. because the hyperscalers have actual demand from customers. AI data centers are being planned and constructed because there is real demand.”
Eisman contrasts the current cycle with the dot-com era, noting that today’s spending is coming from established global companies responding to confirmed orders, not speculative hopes.
“None of this is imaginary. This is not like the dot-com bubble, where hundreds and hundreds of companies went public with little more than a business plan.”
He points to the scale and credibility of the buyers involved as a key distinction.
“The largest companies in the world, Amazon, Google, Meta, Oracle, are spending billions of dollars in CapEx because they have actual orders.”
Eisman says the momentum in AI spending is being reinforced by the technical trajectory of large language models (LLMs), which continue to expand in size and capability.
“Right now, there is a virtuous cycle because the LLMs that create AI continue to scale ever larger. And as they scale, they have an insatiable demand and need for more and more chips.”
He adds that the financial results flowing from that demand reinforce the case that the boom is grounded in reality.
“So the demand is real, and the money is being spent by large global companies. The fact that Nvidia’s third quarter 2025 revenue grew in excess of an amazing 60% is really all you need to know. The bulls argue that it is still very early in the AI story. For now, we should definitely expect the AI growth story to continue well into 2026.”
As to what could derail the AI trade, Eisman recently pointed to an argument on scaling laws that predicts LLMs lose their efficacy as updated versions are released. The investor noted that the improvement in ChatGPT-5 is “a hell of a lot less” over 4.0 than 4.0 over 3.0.
Disclaimer: Opinions expressed at CapitalAI Daily are not investment advice. Investors should do their own due diligence before making any decisions involving securities, cryptocurrencies, or digital assets. Your transfers and trades are at your own risk, and any losses you may incur are your responsibility. CapitalAI Daily does not recommend the buying or selling of any assets, nor is CapitalAI Daily an investment advisor. See our Editorial Standards and Terms of Use.