When people talk about the AI race, they usually talk about chips. Nvidia (NVDA)'s latest GPUs, the billions flowing into data centers, the battle for computing supremacy. But Hyliion Holdings Corp. (HYLN) CEO Thomas Healy thinks investors should be looking at something far less glamorous: power plants.

And the latest numbers, he argues, should be a wake-up call.

"Last year, China added approximately 500 gigawatts of electricity capacity to their grid," Healy told MarketDash. "In comparison, the U.S. only added approximately 50 gigawatts."

That's a tenfold difference. And Healy believes it could determine who wins the global AI race.

The AI Race Needs More Than Chips

The rapid adoption of artificial intelligence is pushing power demand to unprecedented levels. New AI data centers consume vastly more electricity than traditional cloud infrastructure, creating a growing challenge for utilities and grid operators. While technology companies continue racing to build larger AI clusters, Healy believes power availability is emerging as the industry's most important bottleneck.

"We believe the winner in the AI race will ultimately be who has access to power in order to operate these large data centers," he said in an email.

That view reflects a growing concern across the AI ecosystem. Hyperscalers are spending billions on infrastructure, but new generation capacity and transmission networks often take years to build. The result is a widening gap between AI demand and available electricity.

A Market That's Starting To Pay Attention

Healy believes investors are beginning to recognize the scale of the challenge.

"I think there are a few things happening at once that the market is starting to realize," he said.

Among them: AI demand is proving more durable than many expected, grid expansion projects are taking longer than anticipated, and electricity constraints are extending well beyond data centers into industries ranging from defense to critical infrastructure.

The implications could be significant for investors looking beyond the obvious AI winners. While semiconductors remain essential, the infrastructure required to power AI may become just as important over the coming decade.

The Next Big AI Investment Theme?

For years, discussions about AI leadership centered on computing power. Healy argues the conversation is shifting.

"The addressable market for high-quality distributed power generation is enormous," he said.

His thesis is that companies capable of delivering reliable electricity where and when it is needed could become increasingly valuable as AI workloads scale. Whether Wall Street fully appreciates that opportunity remains an open question.

But if AI's future depends on electricity as much as it depends on chips, the gap between China's and America's power buildout may be more than an energy statistic. It may be an early indicator of who is best positioned for the next phase of the AI race.