Microsoft, one of seven large technology companies that agreed to cover any additional grid infrastructure costs stemming from their data centers under Trump’s original Ratepayer Protection Pledge, supports the bill, describing it as an “important step to help ensure American families are protected from rising electricity costs.” Google, another signatory, generally backs the idea of specialized large load tariffs that allocate network costs back to the hyperscalers.
But … why? After all, these companies are voluntarily putting themselves on the hook for what could be billions of dollars in costs that would typically be socialized to all the customers on the grid.
The Data Center Coalition, a trade group including several hyperscalers, has been more circumspect about the bill. Cy McNeill, the group’s senior director of federal affairs, told me in a statement that the group “is reviewing the details of the Ratepayer Protection Act with our members and looks forward to engaging with policymakers on this important topic.”
Evans, Castor, Guthrie, and and the rest appear to be acting not out of hostility towards the AI industry, but rather from a desire to protect it from public backlash fed by rising electricity prices. Earlier this month, Guthrie co-signed a letter to FBI Director Kash Patel, among others, raising concerns that China had “engaged in a coordinated effort to slow U.S. growth in AI development and the building of infrastructure supporting AI data centers” by fomenting domestic opposition — hardly the interpretation of someone working against the industry.
The explanation, perhaps, lies in the answers to two big questions about the Ratepayer Protection Act:
1. Are data centers responsible for higher electricity prices now, or will they be in the future?
2. And would the approach taken in the law actually work to protect ratepayers?
As to the first question, analysts have come up with a nuanced answer. The electricity cost increases we’ve seen in the last five or so years have been largely driven by expenses associated with the distribution grid, including the poles and wires themselves. In some states, like California, the costs come back to wildfires; in others, like Maine, to storm remediation. Looking backwards to 2019, researchers have not been able to find a regular relationship between load growth and price hikes.
In fact, several states “absorbed large industrial and data center load additions while reducing inflation-adjusted retail prices,” according to researchers at Columbia University’s Center on Global Energy Policy. By contrast, some states with little load growth from industry or data centers, such as Maine or California, have seen prices rise substantially.
Many analysts expect electricity prices to continue rising nationally, and data centers could be a driver going forward as demand hits a grid whose capacity to generate and transmit electricity is increasingly strained. This is likely already happening in the country’s largest electricity market, PJM Interconnection, where the system’s independent market monitor has claimed that current and forecasted data center demand has cost customers over $23 billion from recent capacity auctions.
To get prices to actually fall — or at least grow more slowly —it would require that “low-cost supply is available, existing infrastructure is more fully utilized, and cost allocation ensures that new demand contributes to system efficiency,” the Columbia researchers write. Under business as usual however, prices will likely continue to rise.
On the second question, there is much more cynicism.
Critics of the original Ratepayer Protection Pledge, including Harvard Law School’s Ari Peskoe, pointed out that the actual parties to ratemaking — utilities and state regulators — were not involved in the pledge at all. Already, there are accusations that projects developed by pledge signatories could lead to higher prices. Meta's sprawling planned data center project in Louisiana is responsible for the utility’s plans to buy a Texas natural gas-fired power plant, according to documents filed by regulators reviewed by the Times-Picayune. The $1.8 billion deal could lead to $8 a month in additional costs for typical Louisiana ratepayers.
The Ratepayer Protection Act would go a bit further than the pledge, amending the Public Utility Regulatory Policies Act to “establish a Federal standard relating to the recovery of the full, incremental costs of upgrades that serve large-load customers.” Peskoe, however, described this to me in an email as “largely symbolic” and noted that “Congress may not force state regulators to do anything” under current Supreme Court jurisprudence. “This section of PURPA is basically Congress asking state regulators to please take a look at the ratemaking standard.”
That being said, Peskoe noted that “many states and non-regulated utilities do tend to consider PURPA ratemaking standards,” but that there’s “no enforcement mechanism,” depriving the law of any teeth. “States can reject the ratemaking standards or adopt them in a way that deviates from what Congress may have intended.”
Still, it is likely in the political interest of state regulators to come up with something on large load tariffs, the Cato Institute’s Travis Fisher told me. He recommended that the National Association of Regulatory Utility Commissioners “spearhead an initiative to get every state regulator to sign a ratepayer protection pledge,” if only to insulate themselves from political backlash and maintain their power over retail ratemaking.
But even if states do adopt the cost allocation principle, determining exactly which infrastructure is being installed due to a data center and what serves all users can be tricky.
“Any real-world example of this is going to be quite complicated, and the devil’s always in the details,” Ben Schifman, a senior technology fellow at the Institute for Progress and a former attorney at the Department of the Interior and the Department of Justice, told me. While it might be possible to conclude that “a given substation is simply only needed for that data center,” he said, “as soon as you start zooming out into the larger, big-ticket investments, it’s quite complicated to attribute the cost to one user or one group of users.”
In summary, the Ratepayer Protection Act will ask state regulators to consider an approach to data center cost allocation that may not capture all of their costs and will likely do little to arrest the fundamental drivers of higher electricity costs. Viewed through this lens, the logic of the coalition supporting both the original Ratepayer Protection Pledge and the beefed-up Ratepayer Protection Act comes into focus.
Electricity prices are likely to continue to rise, and data center construction has powerful interests behind it. The public’s attitude towards data centers is rapidly souring, and no matter how many nuanced PDFs are published on the topic, people continue to blame data centers for higher electricity costs.
And if prices continue to rise, the big data center developers may be able to point to the Ratepayer Protection Act and say “well, it wasn’t me.”
