Blog Post 12.22.25
Alert
Alert
01.27.26
The rapid growth of data centers is exposing the limits of the U.S. electricity regulatory framework that was built for incremental, predictable load growth. That framework is now under scrutiny in proceedings before the Federal Energy Regulatory Commission (FERC) that could materially reshape how large electricity consumers like data centers obtain power and connect to the grid.
At its core, these FERC proceedings raise deceptively simple but consequential questions: Who gets to decide the rules for connecting large loads like data centers to the transmission system—the federal government or the states? Similarly, should large loads be able to access power generating stations directly? The answer to these questions will be shaped by the competing priorities of key stakeholders, including data center developers and owners, hyperscalers, grid operators and ratepayers, in addition to the federal and state governments themselves.
A Regulatory Framework Under Strain
Electricity regulation in the United States rests on a system of “cooperative federalism” embedded in the Federal Power Act (FPA). FERC regulates transmission and wholesale sales of electricity in interstate commerce while states retain authority over retail sales to end-use customers, local distribution, and utility service obligations. States also retain jurisdiction under the FPA to decide the location and type of power generators allowed to be built within their boundaries. For decades, this balance of power between FERC and the states has meant that even very large customers like steel mills, refineries, and, more recently, data centers are treated as retail load subject to state oversight because such large customers are end-use consumers of power.
That framework reflected long-standing assumptions: load growth would be relatively predictable, utility planning would be centralized, and large customers would have limited direct interaction with generation resources. Those assumptions are increasingly misaligned with today’s reality in which large corporate offtakers often contract directly with power generating stations, whether via a physical or virtual power purchase agreement. This trend of direct corporate procurement started over a decade ago, driven in part by the growth of renewable resources and corporate procurement goals designed to advance sustainability initiatives.
However, the explosive growth of data centers has exposed and intensified the strain. In many regions, traditional utility interconnection processes now involve multi-year delays, uncertain upgrade requirements, and service terms such as interruptibility that are incompatible with modern data center operations that require firm, 24-7 baseload electric service. What once functioned for incremental load growth is now a bottleneck for large-scale data center development. As a result, many data center developers and operators are exploring behind-the-meter solutions, though those options present their own set of unique challenges. Meaningful regulatory reforms are therefore needed to reduce interconnection delays and enable the build out and hardening of the grid to meet rapidly growing data center demand.
Forecasts for Data Center Load Growth
As of November 2025, the United States had approximately 5,427 data centers—nearly half of all data centers worldwide. Power demand from these data centers more than doubled between 2017 and 2023, driven largely by the rapid expansion of artificial intelligence (AI) services. That growth is expected to continue. According to a 2025 Berkeley National laboratory report, domestic energy usage from data centers could double or triple by 2028. In major data center hubs like Northern Virginia, these facilities already consume roughly 26% of the state’s total electricity, more than all the households in the region combined. For comparison, a small city typically requires enough power to serve 80,000 to 100,000 homes, while a medium-sized city serves 400,000 to 500,000 homes. A single data center, by contrast, can demand as much electricity as a small city and, in some cases, up to twice that of a medium-sized city. Nationally, data center power demand is projected to increase from 30 gigawatts today to 134 gigawatts by 2030—17% of the total U.S. peak electricity demand.
Competing Visions for the Future
As the federal government’s energy regulators consider the massive need to power data centers, regulatory tension came to a head in October 2025 when Secretary of Energy Chris Wright challenged the historical framework for powering and interconnecting data centers to the grid. Through an Advanced Notice of Proposed Rulemaking (ANOPR), the Department of Energy directed FERC to explore whether large loads greater than 20 megawatts (MW) that interconnect directly to the transmission system should be treated as transmission-level assets subject to federal jurisdiction, effectively bypassing local distribution interconnection entirely. In short, Secretary Wright argued that large data centers should be treated as the functional equivalent of wholesale power generators—exempt from state jurisdiction and eligible for expedited connection or direct powering from adjacent power plants (a practice called co-location).
FERC’s request for comment drew hundreds of responses from states, utilities, grid operators, consumer advocates, data center developers, and technology companies. While commenters broadly agreed on the problem—existing interconnection processes are too slow, inconsistent, and uncertain—they sharply disagreed on the solution to powering large data center load.
States and consumer advocates argued that the proposal represents an unprecedented expansion of federal authority into the exclusive province of the states over retail regulation. They warned that federalizing large load interconnection could undermine state-approved tariffs, interfere with retail rate design, and shift the costs of new transmission upgrades onto existing customers. This is possible, they say, because FERC-regulated transmission providers can pass new costs onto the states directly without state energy and public utility commission approvals. Several state commissions emphasized that they already have bespoke large load frameworks in place and cautioned against replacing them with a one-size-fits-all federal regime, particularly on an accelerated timeline. Both states and consumer advocates also clarified that the issue of powering and interconnecting large loads is being addressed at the state level through existing integrated resource planning and transmission planning processes. Some advocated for greater regional coordination among states and utilities to address cross-border transmission impacts while preserving the retail rate authority of the states.
Utilities and grid operators struck a more nuanced tone. Many support greater standardization and clarity from FERC but caution that creating parallel federal interconnection queues alongside existing state interconnection queues could conflict with regional transmission planning processes. Cost allocation emerged as a central issue: utilities emphasized that if large loads are permitted to interconnect at the transmission level, the governing framework must ensure that those large load customers bear the full incremental costs of network upgrades driven by their interconnection. In particular, the process should avoid shifting upgrade costs to existing customers or undermining long-standing transmission ratemaking principles where customers that cause costs—or benefit from investments—are responsible for paying for them.
Data center developers and technology companies, by contrast, largely support federal action. They point to a fragmented patchwork of state rules, multi-year delays in interconnection, and uncertainty around electric service terms as barriers to investment. From their perspective, predictable access to transmission-level interconnection is essential to maintaining United States leadership in artificial intelligence and advanced computing.
Why the Stakes Are So High
The outcome of the FERC expedited rulemaking on the subject, which may be decided by FERC by April 2026, will have significant implications for both federal-state regulatory authority and the practical aspects of powering data center development, including:
Federal Policy Signals
Until recently, it was unclear how FERC might reconcile the competing visions for a regulatory framework on large load interconnection. That uncertainty narrowed significantly in December 2025, when FERC issued an order requiring PJM Interconnection—the Mid-Atlantic region power grid operator overseeing wholesale transmission and generation for 67 million customers in all or parts of 13 states and Washington, D.C.—to accommodate a new category of large load interconnections associated with the co-location of generating facilities with data centers and other large loads. FERC ordered PJM to implement three new transmission services that will help facilitate large load and co-locating power generators. In a concurring opinion, Commissioner David Rosner concisely captured the broader theme: the grid has changed, load has changed, and tariff structures must evolve accordingly. Although the order formally applies only to PJM, it reflects an emerging principle that large loads can have transmission-level impacts warranting federal oversight. The order described ways that large load data centers can reduce or eliminate significant interconnection delays and transmission upgrade costs, including “bring your own generation” strategies that allow large load to connect power behind the meter, thereby reducing or eliminating the need to connect massive new load into the wholesale power grid.
That direction was reinforced again in January 2026, when PJM became the focus of a Statement of Principles jointly endorsed by the White House’s National Energy Dominance Council and governors from across the Mid-Atlantic and Midwest. The Statement urged PJM to conduct an emergency reliability auction in which data centers would bid for long-term capacity contracts designed to provide revenue certainty for new generation resources. Costs would be allocated to large loads that have not self-procured new capacity or agreed to curtailment, signaling a strong federal interest in tying large load growth to reliability and grid investment. Overall, the Statement seeks to quickly incentivize new power generation and is an effort by the White House and the governors involved to reduce potential costs to consumers of the massive growth in power needed to energize data centers.
Most recently, FERC approved Southwest Power Pool’s High Impact Large Load (HILL) framework—the most concrete regulatory development to date. The defining feature of the HILL framework is its integrated treatment of load and generation. Rather than forcing customers to navigate separate transmission service, load interconnection, and generator interconnection processes, the framework combines them into a single expedited study and approval process. In return for the benefits associated with registering as a HILL, a customer must comply with enhanced operational requirements, including detailed load forecasting and real-time telemetry, to support system reliability. The framework also establishes distinct tracks for large loads paired with existing or planned generation and for loads that bring new supporting generation as part of the interconnection request. In both cases, network upgrades are directly assigned to the relevant customer, reinforcing traditional cost-causation principles and reducing the risk that upgrade costs are shifted to existing ratepayers. FERC emphasized that this integrated approach preserves reliability while accelerating interconnection timelines. The effort provides greater certainty for data center developers and offers a potential template for addressing large load growth in other regions.
Taken together, these developments reinforce that data center powering and interconnection have become a central federal energy policy issue, not merely a state-level matter. These developments also offer hyperscalers and developers important strategies to consider when selecting data center sites and planning how to supply power to data centers.
Conclusion
The convergence of federal policy initiatives and recent FERC actions reflects a clear shift in federal regulatory posture. Data centers are no longer viewed as incremental load that can be absorbed through legacy state level interconnection processes. Instead, they are increasingly recognized as interstate infrastructure with system-level implications for transmission planning, reliability, and wholesale market design. How far that shift extends—and how it is reconciled with state authority—will define the next phase of large load regulation and determine the circumstances in which large loads can connect directly to power generation.