‘An industrial transformation’: Nvidia CEO Jensen Huang says AI will need '1,000 times more' energy than we have now

“When you see AI as essential infrastructure, the implications become clear,” he noted.

“AI starts with a transformer LLM. But it’s much more. It is an industrial transformation that reshapes how energy is produced and consumed, how factories are built, how work is organized and how economies grow.”

In that emerging reality, Huang said that “energy becomes central because it sets the ceiling on how much intelligence can be produced at all.”

Two months later, Huang doubled down on the tech sector’s need for more AI data center development power. The amount of energy that we need for [AI] computing is likely probably 1,000 times more than we currently have,” Huang told a Stanford University computing class.

He’s not alone. U.S. Secretary of Energy Chris Wright, in a June 9 speech at the Global Energy Forum, said the Department of Energy’s focus “is how do we help the energy system catch up, the electricity system, so we don’t slow the progress of AI?”

That ‘catch-up’ climb could be a steep one. New data from Bank of America estimates the U.S. will require more than 230 gigawatts (GW) of fresh generating capacity through 2031, but domestic regulated utilities are only likely to add 93 GW of power in that time frame, leaving a power gap of about 100 GW.

With demand for AI data center power rising in 2026, the U.S. energy sector could be pressed severely, as the BofA data shows. Here’s how that stress is impacting America’s power grid, for the short- and the long-term.

A much-needed U.S. power grid restoration could be looming

With AI developers demanding more power for their data centers, the biggest U.S. grid bottleneck could be electricity generation and transmission, and the supply of semiconductors. “Right now, those two are the biggest roadblocks to an otherwise unstoppable AI investment boom,” Pablo Gomez, director at FTI Consulting, an AI consulting and strategy firm, told Moneywise.

On the electricity side, even though the U.S. is the largest energy producer, that doesn’t translate to more domestic electricity, as the U.S. exports much of its energy. “Even if exports of energy decline in favor of more domestic usage, the power grids and transmission need to be rebuilt, or at the very least modified for more capacity,” Gomez said.

AI is increasingly ‘an infrastructure issue’

AI experts point to an incoming large shadow cast over the U.S. utilities system, as AI data center power demand grows.

“The most important implication of Jensen Huang’s comments is that AI is increasingly becoming an infrastructure issue, not simply a software issue,” Bhargavi Vepuri, director of technology lead at Prudential Financial, told Moneywise. “As AI adoption expands, organizations and policymakers will have to think simultaneously about compute capacity, energy availability, governance, workforce impact, security and who ultimately bears the cost of scaling these systems.”

Huang may be playing a power game of his own with Congress

Huang’s comment that AI will require 1,000% more energy capacity may be designed to shock public policy decision makers into action on power grid expansion.

“Numbers like that scare the hell out of elected officials, whether it’s on Capitol Hill or at City Hall in Plano, Texas, Christopher Lee, partner and senior political strategist with Foresight Strategic Advisors, told Moneywise. “Like your parents warning you about smoking, Huang’s overreach is deliberately fear-inducing.”

A better plan does exist

If Huang and the AI data center energy cohort could have run a NASA-type ‘Space Race’ type of campaign, the path to more data power may have been more expansive. “Here, you tell communities what it takes, how the industry is investing in resource alternatives and why it’s necessary to compete with China,” Lee said.