AI is no longer just reshaping software. It is now pushing its way into the power business, and NextEra's deal for Dominion is the clearest sign yet.
NextEra Energy has agreed to buy Dominion Energy in a transaction valued at about $66.8 billion, according to Reuters, creating what would be the largest regulated electric utility in the world and the biggest U.S. utility by market value. The deal, announced on Monday, lands at a moment when electricity demand from AI data centers is forcing investors to treat the grid as a core part of the AI stack rather than a slow-moving backdrop.
That shift matters far beyond Wall Street. For years, utilities were viewed as regulated, low-growth businesses that attracted income investors and little else. Now they are becoming strategic infrastructure assets, because every extra model training run, inference cluster, and hyperscale campus needs reliable power, and the companies that can supply it are gaining leverage. Reuters reported in February that U.S. utilities were already lifting capital spending plans to keep up with data center demand, with several operators describing the surge as a generational load-growth event.
NextEra is especially well placed to benefit from that rerating. The company already sits at the center of the clean power market, and its scale in renewables gives it a direct route into the energy needs of AI operators that want long-term contracts and a lower-carbon supply mix. Dominion, meanwhile, brings a large regulated footprint and a utility system tied to fast-growing demand regions on the East Coast, which is exactly the kind of geography where data center expansion has become most aggressive.
The timing is not accidental. Reuters reported in April that U.S. electricity use was set to hit record highs in 2026 and 2027 as AI demand surged, with data centers a major driver of that growth. In the same period, the market has watched power companies reprice themselves around one question: who can build generation, transmission, and interconnection capacity fast enough to serve the next wave of AI infrastructure?
That is why the NextEra-Dominion transaction is more than a utility merger. It is a signal that the companies closest to electrons are becoming as important to AI investors as the chipmakers and cloud platforms that sit higher up the stack. If the grid is constrained, compute deployment slows. If power supply is locked up through long-term contracts or utility partnerships, the operators that secure it first can move faster than competitors who are still fighting for capacity.
The deal also underscores how much bargaining power is shifting toward energy owners. Reuters has reported that some utilities are facing rising pressure from customers and regulators over affordability, even as they pour billions into new infrastructure to meet data center demand. That makes the balance of power more complicated, because utilities want to capture growth, but they also have to defend the political case for those investments.
What founders should watch
For AI infrastructure founders, the immediate lesson is straightforward: access to power is becoming a competitive moat. Data center operators that can secure long-term electricity contracts, on-site generation, or utility partnerships will be better positioned than peers that depend on congested regional grids. That is especially true for startups building around high-density compute, cooling, grid software, or energy optimization, where the product may be technically elegant but still useless without dependable megawatts.
It also changes procurement dynamics. A utility that is consolidating scale can offer more than baseline service. It can become a partner in siting, transmission planning, renewables procurement, and long-horizon capacity planning. For startups, that means the customer relationship may increasingly run through the utility or a utility-backed ecosystem, not just through the cloud buyer or data center landlord.
Investors should read the same signal. AI infrastructure is moving from a purely software-led story into one defined by physical constraints, and physical constraints tend to reward incumbents with scale, balance-sheet strength, and regulatory reach. NextEra's move shows that the market is beginning to price those advantages more aggressively. It also suggests that energy infrastructure startups will have a clearer path to strategic relevance, because the buyers of their products are under more pressure to secure every part of the stack.
What looks like a giant utility merger is really a report on the next phase of AI buildout. The software layer may still get the headlines, but the power layer is where the bottlenecks are forming, and the companies that control it are becoming harder to ignore.
Also read: Standard Chartered brings Zodia Custody onto its balance sheet, signalling banks will own crypto services • US and Philippines fast-track 4,000-acre AI hub as compute politics shift • Baidu's AI agent push is turning into real revenue
