As he promised to do in his State of the Union address last week, President Trump on Wednesday secured commitments from leading technology companies to pay for their own power generation.
Under the so-called “ratepayer protection pledge,” the companies say they will either build or buy generation to power new data centers; pay for any new grid infrastructure or upgrades required as a result of the new load, such as transmission; negotiate separate rate structures with utilities and state government; hire from the local community; and, where possible, make backup resources available when it’s needed. The group had representatives from Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and Elon Musk’s xAI.
President Trump has framed the deal as a way to encourage the infrastructure needed for artificial intelligence to be built in the U.S., without raising rates for everyday consumers. In fact, he said (without evidence) that rates could drop “very substantially” as a result.
However, the industry has largely greeted the pledge with a shrug: a decent set of principles, but unlikely to substantially change the trajectory of what big tech companies were already doing. And Trump himself acknowledged that this deal is largely about public perception of data centers in light of affordability concerns. “They need some PR help because people think that if a data center goes in there, electricity prices are going to go up,” he said at the signing.
Crucially, the pledge has no binding mechanism, and Trump did not mention how or if he plans to hold hyperscalers accountable. And even if it were enforceable, it’s unlikely that it would be effective operating at the federal level alone, given that electricity is regulated largely by state and by region.
As Nick Zenkin, analyst for Latitude Intelligence, explained, FERC doesn’t have the authority to enforce the commitments, and while state commissions could theoretically incorporate the pledge terms into tariff conditions or interconnection agreements, “nothing in the pledge requires them to.” To make this enforceable, he added, would require concerted, coordinated efforts by FERC, RTOs, and state commissions.
“Without that translated into binding contracts or tariff language, it’s basically a big press event,” Zenkin said. “It just tells hyperscalers to sort out their own power problem however they want, and calls it ratepayer protection.”
Furthermore, there’s a question of which data centers are actually subject to the terms of the pledge. Is it only new data centers, or those that already exist or are under construction?
As Abe Silverman, a research scholar at Johns Hopkins, put it on LinkedIn, “one interpretation of the Ratepayer Pledge is that it is intended to put ratepayers back in the position they would have been but-for the data center revolution. Another interpretation is that existing data centers should be exempt from the Pledge, with plenty of options in between!”
The origin of the pledge is the increase in electricity prices in the U.S. Utilities have requested a combined $31 billion in rate increases in 2025, and as a result ire has turned to data centers — and Trump has zeroed in as well.
In the lead-up to the ratepayer protection pledge, Microsoft, OpenAI, and Anthropic all vowed to pay their own way. But after President Trump framed these actions as a major win for ratepayers in his SOTU address last week, Microsoft’s former head of energy Brian Janous expressed major skepticism. “They are meaningless because data centers have been paying their own way from day one,” he wrote last week, arguing that this is simply “ratemaking 101.”
The dramatic expansion of data centers, of course, has put pressure on these existing processes. And many states have already developed separate tariffs for data centers and other large loads, as the ratepayer protection pledge encourages.
“Many utilities already have large-load tariffs with exactly these features — separate rate schedules, minimum billing provisions, and demand charges — that apply regardless of consumption,” Zenkin said. “The pledge describes what a good, large load tariff design already looks like in markets that got ahead of it.”
Tech companies are making noises in support of the pledge. However, there is also the sense from some that this doesn’t represent a significant change in how they are already doing business. Amanda Peterson Corio, Google’s global head of data center energy, described the company’s signature as affirming its “long-held commitment to protect ratepayers, create jobs and keep the grid reliable as our business grows,” and citing the company’s pre-existing work on advanced conductors with CTC Global as well as the company’s December acquisition of Intersect Power’s development arm.
Certain stakeholders have reacted with more alarm than ambivalence, in part because the pledge doesn’t specify what kind of generation data centers should build. Of course, it would have been highly unrealistic to expect that the Trump administration would encourage clean firm solutions (much less renewables with storage) over fossil gas, but green groups are expressing concern that the pledge could encourage an even faster build-out of fossil fuels to power AI.
The Sierra Club, for instance, cited its own gas tracker which finds a nearly 50% increase in planned gas plants over what is currently online, and argued that clean energy would be the lowest cost and lowest risk option for meeting data center demand.
“This pledge is like asking the fox to guard the hen house,” said Sierra Club principal advisor Jeremy Fisher in a statement. “This is a pinky promise, nothing more. Big Tech companies must now take real actions and show up in state regulatory venues and hold data centers accountable for every penny of their energy infrastructure costs, making sure data centers are building lower cost, lower risk clean energy, and minimizing impacts to local residents.”
Transition-AI 2026, Latitude Media’s flagship event at the center of the AI-energy infrastructure buildout, will convene developers, utilities, regulators, hyperscalers, and capital providers to move towards coordinated, sustainable solutions for AI-driven demand. Register now to secure the lowest rates and join us in San Francisco April 13-14.
