NextEra Energy (NYSE:NEE) is reportedly in advanced talks to acquire Dominion Energy (NYSE:D), and yes, on paper, that sounds like a utility merger. A large one, too. NextEra is worth nearly $200 billion, while Dominion carries a market value of roughly $50 billion. But the more interesting story is not just consolidation. It is power.
For the past two years, investors have treated Nvidia (NASDAQ:NVDA) as the face of artificial intelligence. That made sense. Chips were scarce. GPUs were gold. Data centers became the new factories.
But factories need electricity.
That is where this reported deal gets interesting. Dominion serves Virginia, including Northern Virginia’s famous data-center corridor. NextEra already sits at the center of American power generation, renewables, nuclear, storage, transmission, and regulated utility growth. Put those together, and this starts to look less like a utility deal and more like an AI power grab.
The AI Bottleneck Is Moving From Chips To Power
The first phase of the AI trade was easy to understand. More AI meant more chips. More chips meant more Nvidia. More Nvidia meant more data centers. Wall Street loves a clean story, and for a while, this one was almost too clean. But AI has a habit of creating second-order problems. The next one is simple: where does all the electricity come from?
NextEra’s latest earnings call made that point very clearly. Management said electricity demand is accelerating, not slowing. It also said customers need power now, and “speed to power” has become essential. That phrase matters. AI companies do not just need cheap power someday. They need huge amounts of reliable power soon. That changes the market value of utilities, power developers, grid operators, and transmission owners.
This is why the reported Dominion talks are not just about adding customers or rate base. Dominion has exposure to one of the most power-hungry regions in America. Northern Virginia is packed with data centers. These facilities do not sip electricity. They gulp it. Dominion has said peak demand on its system could double by the end of the 2030s. That is not a sleepy utility story. That is an infrastructure bottleneck hiding in plain sight.
Investors have spent years debating chip supply. That debate is not going away. But a server rack without electricity is just expensive furniture. The next constraint may be less glamorous than a GPU. It may be a transformer, a substation, a gas plant, a battery project, or a transmission line. That is not as flashy. It is also much harder to fake.
Dominion Gives NextEra A Front Row Seat In Data Center Alley
Dominion’s Virginia footprint is the part of this story that deserves the biggest spotlight. Northern Virginia is often called “Data Center Alley” for a reason. It is one of the most important digital infrastructure hubs in the world. Cloud companies, enterprise customers, and AI workloads all depend on that region. That makes Dominion more than a regulated utility. It makes it a gatekeeper to one of AI’s most important power markets.
This is where NextEra’s reported interest starts to make strategic sense. NextEra already owns Florida Power & Light, the largest electric utility in the U.S. It also has a huge renewables and storage business, nuclear assets, transmission capabilities, and large-scale development experience. Dominion brings geography. NextEra brings scale. In an AI power cycle, that combination is not boring. It is valuable.
There is also a timing angle here. Utilities are being asked to serve demand growth that the system was not built to absorb this quickly. AI data centers need firm, round-the-clock power. Manufacturing, electric vehicles, and population growth are adding to the same pressure. That means utilities with the right territories may become strategic assets. Dominion’s territory sits near the heart of the AI infrastructure buildout.
Of course, this does not mean a deal would be simple. Utility mergers face regulators, politicians, customers, and consumer advocates. Ratepayer protection will matter. So will balance sheet discipline. But if NextEra wants to become the national power platform for hyperscalers, Dominion’s Virginia exposure is hard to ignore. In this market, location is not just about real estate. It is about access to megawatts.
Utilities Are Becoming AI Infrastructure Platforms
The phrase “utility stock” still sounds sleepy. It brings to mind dividends, rate cases, and people arguing about allowed returns on equity. Not exactly cocktail-party material. But the AI boom is changing that. Utilities are no longer just defensive yield plays. The right ones are becoming infrastructure platforms for the digital economy.
NextEra’s own numbers show why. Florida Power & Light has about 21 gigawatts of large-load interest, with roughly 12 gigawatts in advanced discussions. Management said some of that demand could begin being served as soon as 2028. It also said each gigawatt of large load under its tariff could represent around $2 billion of capital investment. That is the bridge between AI hype and utility earnings. More data centers can mean more regulated capex, more rate base, and more long-term infrastructure demand.
NextEra is also building a national data-center hub strategy. It has more than 30 hubs and aims to secure roughly 40. Its base case is 15 gigawatts of new generation for large-load customers by 2035. The upside case is 30 gigawatts or more. That is not a side project. That is a roadmap.
The key phrase is “Bring Your Own Generation,” or BYOG. The model is simple. Hyperscalers need power. NextEra builds the infrastructure. Hyperscalers pay their fair share. Everyday customers do not get stuck with the full bill. That matters politically. It also matters commercially. AI power demand can be attractive, but only if it does not trigger a ratepayer backlash. Utilities now have to sell both electrons and fairness.
The Grid Needs Gas, Solar, Storage, Nuclear & Transmission
The AI power story is not a one-fuel story. That is important. It would be neat if one technology solved everything. It would also be fiction. AI data centers need speed, scale, reliability, and affordability. No single resource checks every box all the time. That is why NextEra keeps talking about building all forms of energy infrastructure.
Renewables and storage are fast to deploy. NextEra added 4 gigawatts of long-term contracted renewables and storage projects to its backlog in the quarter. Its total backlog reached about 33 gigawatts. Battery storage is especially important because it helps manage timing. Power demand does not politely arrive only when the sun is shining. Storage helps bridge that gap.
Gas still matters too. NextEra was selected to build 9.5 gigawatts of new gas-fired generation tied to large-load demand in Texas and Pennsylvania. Gas plants can provide firm capacity, but they take time. Turbines, labor, permitting, pipelines, and interconnection all matter. This is where the physical world pushes back against the spreadsheet.
Nuclear is another piece. NextEra and Alphabet (NASDAQ:GOOGL) have been working on restarting the Duane Arnold nuclear plant in Iowa, with service targeted no later than the first quarter of 2029. NextEra is also evaluating small modular reactor opportunities at existing nuclear sites. Nuclear is not quick. It is not cheap. But for AI customers that want clean, constant power, it remains part of the conversation.
Then comes transmission. This may be the least glamorous part of the story. It may also be the most important. Power generation is useless if it cannot reach the load. NextEra has secured more than $5 billion in new transmission projects since 2023. That is the unsexy backbone of the AI economy.
Final Thoughts
NextEra’s reported talks with Dominion show how quickly the AI story is expanding. The market started with chips. Then it moved to cloud capex and data centers. Now it is moving into power plants, grid connections, transmission corridors, nuclear restarts, gas supply, and utility regulation. That is a very different kind of AI trade. It is slower, heavier, and more political. It is also harder to replicate.
The bull case is clear enough. NextEra has scale, development experience, a large balance sheet, and exposure to many parts of the power value chain. Dominion would add a strategic Virginia footprint at a time when data-center demand is reshaping electricity markets. The caution is valuation. As of May 15, 2026, NextEra traded at about 11.07x LTM revenue, 21.79x LTM EBITDA, 38.46x LTM EBIT, and 23.71x LTM P/E. Those are not distressed-utility multiples. They already reflect a serious growth premium. For the stock to work from here, the AI power thesis has to translate into durable earnings growth, disciplined capital spending, and acceptable regulatory outcomes. That is possible, but not automatic.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
