The artificial intelligence (AI) boom has stocks soaring to new highs and electricity demand surging to record levels.
It has become pretty clear these last couple of years that this AI frenzy is fast becoming one of the biggest infrastructure and energy stories in America. And as hyperscale data centres multiply across the nation, utilities are facing an unprecedented challenge: ensuring enough generation capacity, grid infrastructure, and capital to meet the ballooning electricity demand.
“Data centers need electricity 24/7, and that makes contracted power assets look more like digital infrastructure than traditional energy assets. Storage additions are also rising rapidly as grid reliability becomes more important,” says Baron Lamarré, petroleum economist and former senior oil trader at Petronas.
This situation is resulting in a growing belief across the industry that scale is key and acquisitions and partnerships could become defining features of the next phase of the power sector.
We saw this with NextEra Energy’s massive $67 billion Dominion takeover deal, the largest energy acquisition of this century, to build the world’s largest utility to dominate the AI data center expansion.
This transaction made it clear that controlling the grid is no longer just an energy play but a technology play, a logic now reshaping the capital-intensive and heavily regulated power industry. The driving force behind it is the unprecedented demand for power.
After being flat for two decades, America’s electricity consumption is projected to rise to 4,283 billion kWh in 2026, up from a record 4,097 billion kWh in 2024.
That is primarily because of data centres, which could consume up to 17% of total U.S. electricity generation by 2030, up from about 4% today. Notably, these estimates are about 60% higher than its previous 2024 forecasts, reflecting the extraordinary pace of AI-related development.
Meanwhile, S&P Global forecasts the grid power demand of U.S. data centres to almost triple to 134.4 GW by 2030. These numbers show the level of investment needed across generation, transmission and storage infrastructure.
For utilities, this is the greatest opportunity but also the biggest financing challenge they have ever faced.
In response, the power and utilities sector has been making deals that totaled $142 billion across 157 transactions in 2025, up from $28 bln in 2024.
This includes Constellation Energy’s $29.4 billion acquisition of Calpine and NRG Energy’s $12 billion purchase of LS Power assets. The capstone of this cycle, so far, is the NextEra-Dominion combination, where the latter holds nearly 51 GW of contracted data-center capacity and counts tech giants like Amazon, Microsoft, Meta, Equinix, CoreWeave, and Alphabet among its customers.
But the more important question now is whether this deal creates a blueprint for the rest of the industry.
“Not all utilities have the balance sheet strength or regulatory positioning to pursue mega-mergers,” noted Lamarré, who’s also the co-founder of the International Digital Exchange (INDEX).
“Smaller utilities will increasingly look for mergers, asset swaps, joint ventures, or private-capital partnerships because the investment needed for data centers, grid upgrades, gas generation, nuclear life extensions, renewables and storage is simply too large for many balance sheets,” he added.
But without the scale of NextEra, the path to adequate capitalization is simply too narrow. According to Lamarré, smaller utilities “are now facing a brutal reality: the AI power boom requires capital at a scale many of them cannot finance alone. The sector is moving from regional utility management to industrial-scale infrastructure consolidation.”
This shift is already attracting attention from infrastructure investors. Private equity firms, pension funds, and sovereign wealth funds now increasingly see power generation assets as strategic infrastructure linked directly to the growth of AI.
As per Lamarré, the most attractive acquisition targets will be those “with exposure to high-load-growth regions, especially PJM, ERCOT, the Southeast and data-center corridors.”
On the buyer side, he points to strategics like Southern, Duke, AEP, Exelon, Constellation, Vistra and NRG and expects BlackRock/GIP, EQT, Blackstone, Brookfield and pension funds to be “highly active” too. “The AES, TXNM and Calpine transactions already show that private capital sees power assets as core AI infrastructure, not just utility exposure,” added Lamarré.
All these deals, however, are just a beginning, per Lamarré’s, as “the U.S. power sector is being repriced around scarcity: scarcity of grid connections, dispatchable power, transmission capacity, and reliable baseload generation. When scarcity appears, assets with real electrons become strategic.”
Instead of buying utilities just for dividend yield, he noted that the market is “buying power access, grid position and the right to serve the AI economy.”
This means “an acceleration in dealmaking,” with regulatory scrutiny being “the gating factor.”
But will this consolidation and the promise of scale-driven efficiency materialize in lower bills? That’s not entirely impossible, “as scale could improve efficiency in capital deployment and grid optimisation,” said Lamarré, only to add that it “depends heavily on regulatory frameworks passing savings through to consumers.”
“The bigger driver of bills will remain capex for grid upgrades and generation build-out—not consolidation per se. I reckon that these mergers can help solve the supply problem, but they do not make the investment cost disappear. Consumers may get short-term bill credits, but the real test is whether scale produces cheaper electricity over ten years, not cheaper headlines on day one,” Lamarré said.
