Jensen Huang Says AI Needs 1,000 Times More Power. Here's Why Oklo and NuScale Power May Still Suffer.

Despite a recent pullback, artificial intelligence (AI) stocks remain some of the market's best performers in recent years. And projections look very positive for additional AI spending and continued rapid adoption. Goldman Sachs now forecasts AI spending to surpass $1 trillion in 2026. Meanwhile, more than one-quarter of all U.S. businesses are now regularly using artificial intelligence products.

The recent pullback in AI stocks wasn't driven by uncertainty about demand. Instead, there are rising concerns regarding the pace of the global data center build-out. To scale adoption, more data centers will need to be built. Doing so, however, generally requires support from the communities these facilities will be located in -- support that is increasingly difficult to find. Plus, these facilities require substantial capital to build. The complete availability of that capital is currently in question.

"AI demand remains strong," affirms recent research from Barclays. "What the market is questioning is funding, as hyperscalers keep lifting capex while free cash flow evaporates."

Concerns about capital availability and public support for additional data centers are relatively new but intensifying. What isn't new is the market's concern over how AI companies will power their data centers.

"Capital is pouring into data center development, but there are real constraints on growth," McKinsey & Co. warns. "Incumbents can't meet demand for power."

Nvidia CEO Jensen Huang agrees. "The amount of energy that we need for [AI] computing is likely probably 1,000 times more than we currently have," he said earlier this year.

It is this need for more energy generation capacity that has driven enthusiasm for nuclear energy stocks, including Oklo (NYSE: OKLO) and NuScale Power (NYSE: SMR). Both of these companies are developing small modular reactors, or SMRs. These systems can, at least on paper, be built faster with lower initial capital requirements than conventional nuclear power plants. That makes them an ideal fit for the AI industry, which needs to scale as much new power generation as quickly as possible.

This investment thesis -- that SMR stocks will benefit from the AI industry's desperate need for more energy -- has driven much of the bullish sentiment around Oklo and NuScale Power. Recent developments, however, have called the strength of that thesis into question.

Here's how AI scaling concerns could hurt Oklo and NuScale Power

Despite industry hype, Bank of America's estimates don't predict real-world SMR adoption will scale meaningfully until the 2030s or even the 2040s. That's a problem for NuScale Power and Oklo, given that both companies remain unprofitable and the full commercialization of their SMR designs may still be years away.

Image source: Getty Images.

To sustain themselves financially, both Oklo and NuScale Power have sold large blocks of shares to raise cash, diluting shareholders. With capital and social support concerns on the rise, SMR adoption by AI and data center companies could be delayed, especially since lower-cost, lower-risk alternatives such as solar, natural gas, and wind remain accessible.

This year, shares of Oklo and NuScale are both down more than 60%. If the SMR build-out is stretched out, more stock issuances will be on the way, this time at rock-bottom prices. While I still believe in the SMR story over the long term from a business perspective, the investment case for SMR stocks is currently facing mounting uncertainty.

Bank of America is an advertising partner of Motley Fool Money. Ryan Vanzo has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.