Every investor knows the AI story by now. Chips. Servers. Data centers. Software. All of it has been picked apart for months.
But there’s one piece of the story that still doesn’t get enough respect: electricity.
AI data centers use enormous amounts of power. They don’t just need electricity during business hours. They need it all day, every day, with very little room for failure. That changes the investment math. The winners are not limited to chipmakers. The companies that generate power, sell power, and build the equipment that keeps the grid running are becoming just as important.
I’m not saying every utility is suddenly an AI stock. That would be lazy. But a handful of companies sit right in the middle of the power problem. Here are three I’d keep on my watchlist.
Vistra (VST)
Vistra owns a mix of natural gas, nuclear, solar, battery storage, and retail power operations. That combination matters because AI data centers need reliable power first. Clean power is nice. Cheap power is nice. But if the lights go out, nothing else matters.
The company reported first-quarter revenue of $5.64 billion, ahead of the $5.24 billion analysts expected. Ongoing operations adjusted EBITDA came in at $1.49 billion, a record first quarter. Management also reaffirmed 2026 adjusted EBITDA guidance of $3.93 billion to $4.73 billion.
The stock has been trading in the mid-$190s recently after a huge run. That means you don’t buy it blindly. But Vistra has something the market wants: existing power assets in a country that suddenly needs much more electricity. Building new plants takes years. Existing capacity has value right now.
Constellation Energy (CEG)
Constellation is the largest nuclear power operator in America. That makes it one of the cleanest ways to invest in the AI power shortage.
Nuclear plants are expensive to build and hard to replace. But once they’re running, they provide steady power around the clock. That is exactly what data centers need. Wind and solar have their place, but AI workloads don’t pause when the weather changes.
Constellation reported adjusted earnings of $2.74 per share in Q1, ahead of the $2.53 analysts expected. Revenue came in around $11.1 billion. The company also reaffirmed full-year EPS guidance of $11 to $12.
The stock recently traded around $289, down roughly 18% year to date and about 30% below its 52-week high. That pullback is what makes it interesting. The long-term nuclear thesis hasn’t gone away. If anything, AI demand has made it stronger.
GE Vernova (GEV)
GE Vernova is different from the first two. It doesn’t just sell electricity. It builds the equipment that helps produce and move it.
Gas turbines. Grid equipment. Electrification systems. Nuclear services. These are the unglamorous pieces of the energy system, and right now they’re in high demand. Utilities, data center developers, and industrial customers all need more capacity. GE Vernova sits right in that supply chain.
The company’s first-quarter results beat expectations, and management raised full-year revenue and free cash flow guidance. Its HA gas turbine fleet recently passed 4 million commercial operating hours worldwide, which is a useful reminder that this isn’t a concept stock. The equipment is already running.
The caution is valuation. Shares recently traded around $965. The stock is up sharply over the past year, even after pulling back from its 52-week high above $1,180. GE Vernova may be one of the best-positioned industrial names in the AI power cycle, but price still matters.
The AI boom will not be powered by speeches, headlines, or investor enthusiasm. It will be powered by electricity.
That’s why I keep coming back to this group. Chips may get the attention, but the grid gets the bill.
