• Alger Investments' Brad Neuman recently told BI that the firm continues to bet on AI-power stocks.

• Data centers are expected to use 25% of US electricity by 2030, up from 10% today, Alger projects.

• He highlighted three stocks to play the trade: GE Vernova, FTAI Aviation, and Bloom Energy.

A few months back, semiconductor stocks were hardly the only ones to take it on the chin when investors got spooked about a potential pullback in hyperscaler spending.

Stocks across the AI value chain were hit with force, including energy infrastructure firms. From June 22 to July 29, the Defiance AI and Power Infrastructure ETF (AIPO) fell 24%. If investors were going to start pressuring AI's big spenders to tighten their purse strings, then these power companies wouldn't be so in demand.

Since the end of July, a couple of catalysts have driven the stocks higher: hyperscalers like Microsoft and Amazon have shown they can monetize their investments, and AI infrastructure stocks have shown in earnings that there's no slowdown in spending happening.

Brad Neuman, the director of market strategy at Alger Investments, which manages $42 billion in assets, says to stick with the trade as it rebounds.

He said if there's any risk to the trade, it's not a pullback in AI spending. It's instead the prospect of energy firms being unable to fulfill demand.

"We're more worried about supply than demand. In other words, the ability to bring on data centers, build them, connect them to the grid, etc., rather than demand, which is I think growing very fast for end user tokens," he told Business Insider in an August 12 interview.

He added: "I think that's growing over 10x year on year, and it's hard to meet that demand."

Alger is estimating that data centers will use 25% of total US electricity by 2030, up from roughly 10% today. To meet that demand, energy generation will have to increase accordingly, he said. The firm is projecting that terawatt-hours generated in the US will double between now and 2040.

Neuman said the firm will continue to invest in the AI power bottleneck, betting on companies positioned to help build the infrastructure to meet the heightened demand levels. He shared three stocks to ride the theme.

3 stocks to play the bottleneck

First, Neuman highlighted GE Vernova (GEV).

The company produces natural gas turbines, which power data centers, Neuman said.

GE Vernova has a five-year backlog in orders for their turbines, he said, which will boost its earnings growth into the 2030s. However, investors still aren't pricing in that growth, he said.