FERC Must Rethink Data Center Pricing Policies

The bill is coming due — and right now, the wrong people are paying it.

As hyperscale data centers plug into the grid at an unprecedented scale, a quiet but consequential debate is unfolding at the Federal Energy Regulatory Commission. Harvard electricity law expert Ari Peskoe has laid out the argument plainly: FERC's current transmission pricing policy is broken, and it's letting data centers off the hook for costs they're directly creating. The question isn't whether this gets fixed; it's whether regulators move fast enough before ratepayers and infrastructure developers absorb consequences they never signed up for.

Understanding FERC's Current Pricing Policy

FERC governs the rates, terms, and conditions of wholesale electricity transmission across interstate lines. Its pricing framework — built over decades of incremental rulemaking — was never designed with 100-megawatt data center campuses in mind. The existing model generally socializes transmission costs across broad customer classes, meaning everyone on the grid shares the expense of keeping power flowing.

That made reasonable sense when load growth was gradual and geographically distributed. It doesn't make sense when a single technology sector is responsible for driving the majority of new capacity demands in specific regional pockets.

The fundamental problem is one of cost causation: the entity creating the need for new infrastructure should bear the cost of that infrastructure. Right now, that principle is being honored more in theory than in practice. Utilities are investing in transmission upgrades to serve large industrial loads — data centers chief among them — while those costs get rolled into rate bases that every ratepayer shares.

Peskoe's argument, directed squarely at FERC, is that the commission needs to revisit this framework and require utilities to assign the full costs of service to power-hungry data centers. That's not a radical idea; it's how cost-of-service regulation is supposed to work.

The Rising Costs of Data Centers

The scale of data center growth over the past three years has genuinely caught grid planners off guard. Northern Virginia — the single densest data center market on earth — has seen demand growth that's straining both transmission infrastructure and interconnection queues. PJM, the grid operator serving the Mid-Atlantic and parts of the Midwest, reported interconnection requests totaling over 250 gigawatts in its queue as of recent counts. A significant portion of that backlog is tied to large commercial and industrial loads, with data centers representing an outsized share.

Nationally, data centers currently consume an estimated 2-3% of U.S. electricity. Multiple forecasts, including those from utilities and grid operators themselves, project that figure could double by the end of the decade. That's not incremental load growth — that's a structural shift in how the grid gets used.

When a single customer class drives transmission upgrades worth hundreds of millions of dollars, the math on "shared cost socialization" stops working. Georgia Power, for instance, has flagged the need for significant capital investment to serve data center load in its service territory. Those investments don't benefit residential customers in Savannah or small manufacturers in Macon — they exist to serve the specific connectivity and reliability needs of hyperscale facilities.

The energy intensity is part of what makes this unusual. Unlike a manufacturing plant that might run at variable loads, a hyperscale data center typically operates at high utilization around the clock. That constant, heavy draw creates specific transmission requirements — dedicated capacity, redundant pathways, tight voltage regulation — that standard grid infrastructure wasn't built to deliver without significant upgrades.

Why Utilities Should Bear Full Costs

To be precise about Peskoe's argument: it's not that utilities should absorb these costs. It's that utilities should *assign* them — to the data centers causing them. The distinction matters enormously.

Under proper cost-of-service principles, a customer whose load requires dedicated infrastructure upgrades should pay for those upgrades through their rates. This is how large industrial customers have historically been treated in many jurisdictions. A steel mill that requires a new substation doesn't expect residential ratepayers to split the tab. Applying that same logic to data centers isn't punitive — it's just accurate accounting.

The counterargument from the data center industry is predictably economic: full cost assignment raises operating expenses, potentially chilling investment and pushing development toward jurisdictions with more favorable rate structures. That's a real concern, but it's also an argument that could be made against almost any form of cost accountability. The risk of chilling investment has to be weighed against the certainty of inequitable cost shifting that's already happening.

There's also a less obvious argument for full cost assignment that often gets missed: it creates price signals that drive better infrastructure planning. When data centers bear the true cost of their grid impact, operators have a financial incentive to locate facilities where grid capacity already exists, invest in on-site generation or storage, and think more carefully about how they connect to the grid. Right now, those incentives are muted. Data centers can locate in capacity-constrained areas and let everyone else fund the upgrade.

Potential Consequences of Inaction

The near-term consequence of leaving FERC's pricing policy unchanged is straightforward: ratepayer cross-subsidization at scale. Residential customers and small businesses in data center-heavy markets will see rate pressure from transmission investments they didn't drive and don't benefit from. That's already beginning to show up in utility rate cases.

The longer-term risk is more structural. Grid operators and utilities are already struggling to site and permit new transmission at the pace load growth demands. If the cost signals are wrong — if data centers face no financial incentive to locate, size, or operate in grid-friendly ways — planning becomes harder and capital gets misallocated. Transmission projects that should be built to serve broad regional needs get crowded out by data center-specific upgrades that are socialized rather than targeted.

Infrastructure strain isn't a hypothetical future problem. PJM has already delayed interconnection decisions for thousands of projects because the queue management system couldn't handle the volume. When data centers jump to the front of that line without bearing proportionate costs, renewable energy projects, industrial manufacturers, and others pay the price in delays and uncertainty.

There's also a reliability dimension. The concentrated geographic clustering of data centers — Northern Virginia, Phoenix, Silicon Valley, Chicago — creates localized demand spikes that stress regional transmission systems in ways diffuse load growth doesn't. Pricing policy that ignores geographic concentration overlooks where the actual stress is occurring.

The Path Forward for Stakeholders

FERC has the authority to act here, and the political pressure to do so is building from multiple directions. State utility commissions are watching their rate cases get complicated by data center load. Environmental groups are flagging the renewable energy implications of unchecked demand growth. Some utilities are already pushing for tariff structures that assign larger cost shares to high-load customers.

What Peskoe's analysis does is provide the regulatory and legal framework for FERC to act with confidence. The commission doesn't need to invent new doctrine — it needs to apply existing cost-causation principles consistently to a customer class that has, until recently, largely escaped that scrutiny.

For developers, utilities, and infrastructure investors, the practical takeaway is this: transmission pricing reform is coming, and the direction of travel is toward greater cost accountability for large loads. Data center developers who get ahead of that shift — by co-locating with existing generation, investing in behind-the-meter storage, or selecting sites with genuine grid headroom — will be better positioned than those who assume the current pricing environment is permanent.

The grid doesn't care about market valuations or AI investment cycles. It runs on physics and capital, and right now, neither is being properly priced. FERC has the mandate to fix that. The commission's willingness to move — or not — will shape utility economics, infrastructure investment, and electricity rates for the next decade.

[INTERNAL LINK: FERC policies]

[INTERNAL LINK: data center growth]

[INTERNAL LINK: transmission pricing reform]