5 power players winning the $700 billion AI energy race

The AI industry’s insatiable demand for computation has created a new chokepoint, shifting the primary bottleneck from graphics processing units to the electrical grid itself. While Nvidia captured the first wave of investment and memory makers like Micron followed, a new class of infrastructure and utility companies is emerging to win Big Tech’s projected $700 billion spend on energy, threatening to reshape the AI supply chain.

“The increasing adoption of artificial intelligence applications is driving massive power consumption at hyperscale data centers, creating opportunities for alternative and reliable energy providers,” Zacks Investment Research noted in a May 8 report.

This power demand is already showing up in utility earnings and the stock prices of specialized energy firms. WEC Energy Group, Inc. (NYSE:WEC) saw electricity consumption by its large industrial customers surge 2.7 percent in the first quarter of 2026, helping drive revenues up 9 percent year-over-year to $3.43 billion. Meanwhile, shares of FuelCell Energy (NASDAQ:FCEL) soared 11.6 percent in a single session on May 8 after it launched a new platform designed for large-scale data center operations.

For investors, this signals a critical new front in the AI trade. The focus is expanding from the chip designers to the companies that provide the picks-and-shovels of electricity generation and grid infrastructure, such as GE Vernova and Bloom Energy. This shift could lead to a significant re-rating of valuations for these previously staid industrial and utility stocks.

From Silicon to Grid: The New AI Bottleneck

The AI buildout has unfolded in stages, each defined by a critical bottleneck. Initially, the constraint was the GPU, with Nvidia Corp. (NASDAQ:NVDA) becoming the undisputed winner. As GPU supply caught up, the bottleneck shifted to high-bandwidth memory needed to feed the processors, causing shares of Micron Technology (NASDAQ:MU) and Sandisk (NASDAQ:SNDK) to surge in early 2026, with Sandisk gaining 40 percent in just five trading sessions.

Now, the industry is facing a more fundamental limit: energy. The sheer electricity demand from clusters of AI accelerators is straining local and regional power grids. Global electricity demand from data centers could double by 2030, according to some estimates. This has turned the business of power generation, transmission, and management into the next critical path for AI expansion, creating a massive opening for companies that can deliver reliable, large-scale power.

Utilities and Fuel Cells: 2 Power Plays

The investment implications are already visible in both traditional utilities and emerging energy technology firms. WEC Energy Group, a provider of regulated natural gas and electricity, delivered a profit of $2.45 per share in its first quarter, up from $2.27 a year earlier and beating estimates by $0.15. The company reaffirmed its long-term earnings per share growth target of 7 to 8 percent annually, citing data center demand as a key driver.

In a more focused play, FuelCell Energy is targeting the data center market directly with on-site power generation solutions. The company's shares jumped over 11 percent after announcing a new 12.5-megawatt platform aimed at hyperscale clients. This approach offers a way for data centers to bypass constraints on the public grid by creating their own reliable power source, a model that is gaining traction as grid connection queues lengthen, with waits of up to seven years in some regions.

This article is for informational purposes only and does not constitute investment advice.