Marketdash
Back to Stock Market News

Steve Eisman Sees Two Major Threats to AI's Momentum in 2026

MarketDash Editorial Team
5 hours ago
The investor who famously predicted the 2008 financial crisis thinks the AI boom faces serious headwinds next year. Power shortages and questions about whether bigger language models actually work better could slow the whole party down.

Steve Eisman, the investor who made his name betting against subprime mortgages before the 2008 crash and got immortalized in "The Big Short," has some thoughts about where artificial intelligence is headed. And they're not exactly optimistic.

The Power Problem and the Scaling Question

Speaking on the Real Eisman Playbook podcast over the weekend, Eisman laid out two major risks that could derail AI's momentum in 2026. The first is pretty straightforward: we're running out of electricity.

"The current binding constraint on AI growth is power," he explained. "You can build an AI data center, but if you can't plug it in, what good is it?" He pointed to two recent examples in Santa Clara, California, where data centers sit without electricity due to massive shortfalls.

U.S. power demand is growing at roughly 3% annually now, but here's the problem: you can't just snap your fingers and create new energy infrastructure. "From drawing up the plans to completion, it's at least a two to three-year process," Eisman noted.

The second risk is more existential and potentially more damaging. Eisman thinks the industry's obsession with building ever-larger language models might be "a dead end."

He brought up OpenAI's release of ChatGPT 5.0, which critics noted wasn't really better than ChatGPT 4.0. "Since then, there has been a growing but still minority opinion that believes that continuing to scale LLMs is becoming a dead end," Eisman said.

If that view gains traction, the implications cascade quickly. "The need for more chips will diminish and tech hardware growth will slow across the board and the virtuous cycle will reverse," he warned.

According to Eisman, "the potential realization that the ever larger LLMs won't get us to the finish line"—meaning AGI, or artificial general intelligence—"is the fundamental risk to the entire AI story."

Reality Check Time

The market seems to be catching on. In JPMorgan Chase & Co. (JPM)'s 2026 outlook, a survey of CEOs shows executives are now demanding clear returns on investment rather than indulging in blind experimentation with AI.

But not everyone's hitting the brakes. Gene Munster of Deepwater Asset Management argues the AI trade is "still early," calling 2026 year three of a "5-year bull market." He expects the Nasdaq to end 2026 up by 10% or more, driven by AI-related gains.

Munster even suggested that "current bubble talk is actually constructive because it keeps expectations in check and lowers the risk of an actual bubble forming." Which is one way to look at it—or maybe it's just what people say when they're trying to convince themselves the party isn't over yet.

Back to Stock Market News

Steve Eisman Sees Two Major Threats to AI's Momentum in 2026

MarketDash Editorial Team
5 hours ago
The investor who famously predicted the 2008 financial crisis thinks the AI boom faces serious headwinds next year. Power shortages and questions about whether bigger language models actually work better could slow the whole party down.

Steve Eisman, the investor who made his name betting against subprime mortgages before the 2008 crash and got immortalized in "The Big Short," has some thoughts about where artificial intelligence is headed. And they're not exactly optimistic.

The Power Problem and the Scaling Question

Speaking on the Real Eisman Playbook podcast over the weekend, Eisman laid out two major risks that could derail AI's momentum in 2026. The first is pretty straightforward: we're running out of electricity.

"The current binding constraint on AI growth is power," he explained. "You can build an AI data center, but if you can't plug it in, what good is it?" He pointed to two recent examples in Santa Clara, California, where data centers sit without electricity due to massive shortfalls.

U.S. power demand is growing at roughly 3% annually now, but here's the problem: you can't just snap your fingers and create new energy infrastructure. "From drawing up the plans to completion, it's at least a two to three-year process," Eisman noted.

The second risk is more existential and potentially more damaging. Eisman thinks the industry's obsession with building ever-larger language models might be "a dead end."

He brought up OpenAI's release of ChatGPT 5.0, which critics noted wasn't really better than ChatGPT 4.0. "Since then, there has been a growing but still minority opinion that believes that continuing to scale LLMs is becoming a dead end," Eisman said.

If that view gains traction, the implications cascade quickly. "The need for more chips will diminish and tech hardware growth will slow across the board and the virtuous cycle will reverse," he warned.

According to Eisman, "the potential realization that the ever larger LLMs won't get us to the finish line"—meaning AGI, or artificial general intelligence—"is the fundamental risk to the entire AI story."

Reality Check Time

The market seems to be catching on. In JPMorgan Chase & Co. (JPM)'s 2026 outlook, a survey of CEOs shows executives are now demanding clear returns on investment rather than indulging in blind experimentation with AI.

But not everyone's hitting the brakes. Gene Munster of Deepwater Asset Management argues the AI trade is "still early," calling 2026 year three of a "5-year bull market." He expects the Nasdaq to end 2026 up by 10% or more, driven by AI-related gains.

Munster even suggested that "current bubble talk is actually constructive because it keeps expectations in check and lowers the risk of an actual bubble forming." Which is one way to look at it—or maybe it's just what people say when they're trying to convince themselves the party isn't over yet.