The artificial intelligence boom is hitting a wall, and it's not about chips or computing power,it's about electricity. Data centers need staggering amounts of power to run AI systems, but the electric grid can't keep up. Three companies are now positioned to solve this crisis, and their financial results show just how urgent the demand has become.

What's Creating the Electricity Shortage for AI Data Centers?

Building a new data center typically requires a connection to the local power grid, but utilities are overwhelmed. Securing a grid connection for a large data center campus can take years, and the equipment needed to build new power plants is in short supply. This mismatch between demand and available infrastructure has created a genuine bottleneck that's slowing down the entire AI infrastructure build-out.

The scale of the problem is staggering. On the same day that chip designer Broadcom announced a financing platform designed to enable more than 20 gigawatts of AI compute capacity through 2028, with an initial $35 billion tranche, even automakers started getting involved. General Motors announced it is developing a sodium-ion battery aimed at energy storage for data centers and the grid, signaling how widespread the electricity crisis has become.

Which Companies Are Winning the Race to Power AI?

Three publicly traded companies are emerging as the primary beneficiaries of this electricity shortage. Each takes a different approach to solving the problem, and their recent financial results reveal explosive growth driven by AI infrastructure demand.

• Bloom Energy (NYSE: BE): Makes solid oxide fuel cells that generate electricity on-site from natural gas without combustion, allowing data centers to bypass the grid entirely. The company's first-quarter revenue soared 130 percent year over year to $751.1 million, with product revenue jumping 208 percent. Management raised its full-year 2026 revenue outlook to $3.4 billion to $3.8 billion, representing about 80 percent growth at the midpoint.
• GE Vernova (NYSE: GEV): Supplies power equipment directly to utilities and manufactures the gas turbines that utilities are waiting in line to order. The company's first-quarter orders surged 71 percent on an organic basis to $18.3 billion, with its gas turbine backlog and slot reservation agreements growing from 83 gigawatts to 100 gigawatts in a single quarter. Management now expects to reach at least 110 gigawatts by the end of 2026.
• Vistra (NYSE: VST): One of the largest competitive power producers in the U.S., with a generation fleet spanning natural gas and nuclear. The company has signed long-term power purchase agreements with Amazon for up to 1,200 megawatts of nuclear power and with Meta Platforms for 2,609 megawatts of nuclear energy and capacity. Vistra reported first-quarter adjusted earnings before interest, taxes, depreciation, and amortization of about $1.5 billion.

The scale of these backlogs is extraordinary. GE Vernova's electrification segment, which makes grid equipment like transformers and switchgear, booked $2.4 billion of equipment orders to support data centers during the first quarter alone, more than in all of 2025.

How Are These Companies Solving the Power Problem?

Each company is addressing the electricity bottleneck through distinct strategies that reflect different aspects of the power infrastructure challenge. Understanding these approaches reveals why the demand is so robust and why these companies are seeing such explosive growth.

• On-Site Generation Strategy: Bloom Energy's fuel cells allow data centers to generate their own power without waiting for grid connections. In April, Oracle announced that Bloom fuel cells will fully power Project Jupiter, its AI data center campus in New Mexico, with up to 2.45 gigawatts of capacity, replacing the gas turbines and diesel generators originally planned. Notably, management said more than half of Bloom's data center backlog comes from customers other than Oracle, indicating broad adoption across the industry.
• Grid Infrastructure Expansion: GE Vernova is expanding the grid itself by manufacturing the equipment utilities need to distribute power more efficiently. The company's free cash flow more than quadrupled year over year to $4.8 billion, and management raised its 2026 guidance, reflecting confidence in sustained demand.
• Long-Term Power Contracts: Vistra is locking in AI companies with multi-decade power purchase agreements. Last year, the company signed a 20-year power purchase agreement with Amazon's cloud unit and 20-year agreements to supply Meta Platforms with nuclear energy. The company is also adding natural gas capacity through its pending acquisition of about 5,500 megawatts of generation from Cogentrix, targeted to close in the second half of 2026.

"Bloom is rapidly becoming the standard and go-to choice for on-site power," said KR Sridhar, founder and CEO of Bloom Energy.

KR Sridhar, Founder and CEO at Bloom Energy

What Risks Could Derail This Growth?

Despite the explosive demand signals, these companies face real risks that could slow their momentum. Project timing is one concern. Bloom Energy shares fell this week after a partner reportedly paused work on a data center site in Wyoming, demonstrating how sensitive these stocks are to construction delays.

Additionally, all three companies operate in highly regulated markets where permitting, environmental reviews, and grid connection approvals can take years. The bigger risk may be the demand side itself. All three stocks have rallied on the assumption that AI capital expenditures keep climbing, and AI infrastructure stocks have pulled back recently on concerns about the pace of that spending. If the build-out decelerates meaningfully, backlogs could stop growing and these valuations could compress quickly.

However, the demand signals keep arriving week after week, and from new directions. Financiers are committing capital one day, automakers are entering the market the next, and major cloud providers are signing multi-decade power agreements. For investors who believe the electricity bottleneck is real and durable, these three companies arguably offer a more grounded way to invest in the AI boom than chasing the chipmakers themselves.