Jensen Huang Says AI Needs "1,000 Times More Power Than We Currently Have." These 3 Industrial Stocks Will Deliver It. - The Motley Fool

What the report says

The Motley Fool reported that Nvidia CEO Jensen Huang told a Stanford University computer science class in May that AI computing could require vastly more energy than is currently available, while acknowledging the estimate could be off by “a couple orders of magnitude.” The article framed electricity supply, rather than memory chips, as a major constraint for future AI expansion.

The report cited outside forecasts to show the scale of demand. Goldman Sachs was said to expect U.S. data center power needs to rise from 31 gigawatts in 2025 to 66 gigawatts in 2027, which would lift their share of national power use from just over 4% to more than 8% at peak usage. Bloom Energy’s forecast, as cited by Motley Fool, put AI’s potential electricity use at up to 12% of U.S. production by 2030.

Motley Fool identified three companies it says are positioned to benefit from this demand: GE Vernova, Vistra and Caterpillar. GE Vernova, spun out of General Electric, was highlighted for natural gas turbines and grid-related equipment, with the article citing a $176 billion companywide backlog and growth in gas power equipment orders. Vistra was described as using long-term power purchase agreements, including deals with Amazon and Meta, to support data center electricity needs. Caterpillar was included because it makes large diesel and gas generators, battery storage and related systems that can serve as interim or off-grid power sources.

The significance is that AI infrastructure growth is increasingly tied to energy availability, permitting, grid connections and generation capacity. Motley Fool’s article presents this primarily as an investment thesis, not as a policy announcement or company guidance from Nvidia.