By Ben Levisohn

Investors looking for the next hot AI stock could do worse than Chevron and Exxon Mobil.

On the surface, that seems like a wacky notion, akin to suggesting that William T. Vollmann should write young adult novels. Chevron and Exxon are dinosaurs, after all, and not the kind you find in children's books. They're exploring for last century's energy, rather than preparing for the next.

But as can be seen by the sudden sexiness of drab utilities like Vistra and Constellation Energy, and industrials that service power plants like GE Vernova and MasTec, AI needs power -- and lots of it. And what are Exxon and Chevron if not companies designed to help supply power? For investors willing to look past their preconceptions, the two oil majors might just be the best AI bargains going.

It's tempting, of course, to forget branching out beyond Nvidia, Meta Platforms, and other Big Tech stocks that look to have the AI edge. Yet figuring out who is ultimately the winner or loser among them may be harder than it looks, according to Gurvir Grewal, a global research analyst at William Blair. AI may allow companies to cut costs and innovate their products, but that won't necessarily translate into higher profit margins. For companies without strong moats, AI tools may just be the cost of doing business, and they will have to cut prices to keep their customers.

"AI will likely be a strategic investment rather than an economic investment for many businesses as they invest in AI not for immediate revenue or cost savings but simply to remain competitive," Grewal writes.

It's one reason so many AI plays simply fizzle out. Super Micro Computer, which rallied 318% from the end of 2023 through its March 2024 high after becoming an AI darling -- before dropping 85% through its November 2024 low -- comes to mind. And it's also why it's difficult to know whether a stock like Duolingo, which has rallied on AI hopes, can sustain its gains.

One of the best places to be might be power. AI is a power play, quite literally. The massive data centers driving the artificial-intelligence revolution require massive amounts of electricity to run, and power production is finally growing after years of stagnation. That has helped lift everything from the utilities providing the power to the companies helping to build and maintain them, including Amphenol and Vertiv Holdings.

The risk, of course, is that too much power is being built out, perhaps more than will ever be needed. Overbuilding was a problem for independent power producers in the past, with Calpine filing for bankruptcy in 2005 and Energy Future Holdings following suit in 2014. Such overbuilding is unlikely to occur now, according to Raymond James analyst J.R. Weston. For one, there's very little building going on. Instead, the utilities have been focusing on upgrading and restoring mothballed power generators to handle the new power needs. It's also unlikely that utilities will be able to overbuild due to regulations and other factors.

On the flip side, investors may be underestimating just how much power is needed. Data centers are just the beginning, according to 22V Research's Jordi Visser, who sees it being used in robots, self-driving cars, smart appliances, and other products that will require even more power. "Unlike traditional data centers that are concentrated and centrally planned, embodied AI will generate a vast, decentralized layer of 24/7 compute demand embedded in warehouses, hospitals, homes, vehicles, and factories," he writes. "These edge applications won't tolerate latency or downtime, making local, dispatchable power essential."

The change has been significant enough to cause a repricing of independent power producers across the utility sector. Vistra now fetches 25.1 times 12 month forward earnings, up from 15.5 times one year ago, while NRG Energy trades for 18.6 times, up from 11.3 times. And it's a big reason that Raymond James initiated the two stocks with Strong Buy ratings and Constellation and Talen Energy at Outperform. "We continue to like the setup for most names on the broader power theme," Weston writes.

Exxon and Chevron, however, are still viewed as stodgy oil companies, not AI beneficiaries. That could change, 22V's Visser claims. He argues that both companies are primed to benefit from demand for natural-gas-fired power, especially if there is likely to be an energy shortfall. Chevron, for instance, has partnered with GE Vernova and Engine No. 1 to expand gas power for data centers, Visser says, while Exxon has touted its ability to supply power to Big Tech firms long before nuclear power can become a realistic option.

"Despite these developments, the market continues to value Chevron and Exxon Mobil primarily as cyclical oil companies," Visser writes. "It has not yet priced in the possibility that these firms will become direct electricity suppliers to hyperscalers, a role that could warrant higher valuation multiples."

Even following their recent surge thanks to rising oil prices, Chevron and Exxon trade at 17 and 16 times 12-month forward earnings, respectively. If an AI premium gets included in the stock, they could be worth a heck of a lot more.

Write to Ben Levisohn at ben.levisohn@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

June 20, 2025 01:45 ET (05:45 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.