What Ratepayer Protection Can and Cannot Fix

Oregon is the proof that Washington's ratepayer protection framework can work. Under the state's POWER Act, data center electricity rates rose 29 percent while residential rates fell 1.3 percent, a shift projected to save non-data-center customers $900 million over thirty years. ^[1] The mechanism is straightforward: make the company building the server farm pay for its own grid interconnection rather than spreading the cost across every household on the system. Oregon did it, and the numbers are real. The framework now extends well beyond one state. In June, the Federal Energy Regulatory Commission issued a show-cause order to six grid operators requiring them to prevent cost-shifting to ratepayers and mandating tariff reform proposals within sixty days. ^[2] A wave of state-level mandatory tariffs has passed in Oklahoma, Wisconsin, Pennsylvania, Louisiana, and Delaware, each requiring data centers to fund their own grid upgrades. ^[3] The Trump administration is expanding its voluntary Ratepayer Protection Pledge, originally signed by Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI, to now include electric utilities, state governors, and third-party data center operators. ^[4] Taken together, this is a real and binding cost-allocation architecture. It is not a sham. And yet household electricity bills are rising anyway. Ohio residential bills are projected to average $800 this summer, a 17 percent increase. ^[5] National electricity prices have climbed 33 percent over five years, adding roughly $420 to annual household bills. ^[6] Maryland passed its own tariff law designed to save ratepayers $150 annually, and its prices are still going up because the regional grid operator cannot keep pace with data center demand. ^[7] Cost allocation and market-level scarcity are two different problems. Making a data center pay for its own substation solves the first. It does nothing about the second. PJM, the grid operator serving 65 million people across thirteen states, projects a potential 60-gigawatt shortfall over the next decade while its current plan secures only 15 gigawatts. ^[8] When electricity supply cannot meet demand, the price of power rises for everyone on the system, regardless of who paid for the last transformer. No tariff can fix a shortage. The capital spending required to close that gap will flow through utility rate cases to all ratepayers. Fifty-one investor-owned utilities plan $1.4 trillion in capital expenditures through 2030 to modernize grids for AI demand, and as the energy research firm PowerLines notes, that spending is "a leading indicator of future rate increase requests." ^[9] AEP, one of the country's largest utilities, has announced a $78 billion capital plan to nearly double its generation capacity to 63 gigawatts. Ninety percent of that expansion is driven by data centers. Its Ohio electric revenue already rose 22 percent year over year. ^[10] The CEO pledged affordability is "top of mind." The spending that makes bills rise is the same spending the pledge cannot touch. There is a historical irony here. From 2015 to 2024, data centers actually lowered average U.S. retail electricity rates, because their large, steady demand spread fixed grid costs across more kilowatt-hours. ^[11] That era is over. The same researchers note that future rate stability "faces significant headwinds" as supply constraints bite. The mechanism has reversed: demand once lowered prices by spreading costs; now it raises them by outstripping supply. If the framework cannot stop bills from rising, what is it for? The answer is in the politics. When Trump unveiled the original pledge in April, he said data centers "need some PR help" and framed the commitment around the public fear that a new data center would raise their electricity bills. ^[12] The fear is well-founded. Seventy-one percent of Americans now oppose construction of AI data centers in their local areas, a higher level of opposition than to nuclear power plants. ^[13] Moratoriums and bans have spread to Ohio, Massachusetts, Missouri, and Maine. Massachusetts State Senator John Velis described what he called a horrifying pattern: data center operators make representations about their costs, but the costs far exceed those initial representations, and in some instances the residents of the respective communities end up on the hook financially. ^[14] Community organizer Morgan Harper captured the broader sentiment, describing a public that feels it has lost control over its own life while others profit on its back. ^[10] The pledge offers an answer to that backlash. Tech companies will "pay their own way," as Trump put it. ^[9] It gives governors who sign it a "protect ratepayers while building" frame rather than a "ban to protect ratepayers" frame. Maine Governor Janet Mills, a Democrat, vetoed the first proposed statewide moratorium on data center construction, illustrating exactly the political lane the pledge opens. ^[13] The framework is not a solution to rising bills. It is a permission structure for building anyway. You can make the tech company pay for its own transformer and the household bill still hits $800, because the shortage sets the price of power for everyone. The pledge, by design, addresses the first sentence and not the second.