FERC puts a June clock on data-center power interconnection reform

FERC has put a date on the data-center power fight

FERC's April 16, 2026 order turned a long-running data-center power debate into a near-term regulatory clock. The Federal Energy Regulatory Commission said it intends to act by the end of June 2026 in its large-load interconnection docket, a proceeding focused on how very large electricity users connect to the interstate transmission system. The target matters because data centers are no longer a distant planning category for utilities. They are one of the clearest sources of new load growth in several U.S. regions, and their power requests are now large enough to affect transmission planning, cost allocation, reliability studies, siting decisions, and state-federal jurisdiction fights.

The FERC docket is formally broader than data centers. It concerns significant electrical loads, including industrial and commercial customers that may want direct transmission-level service. But data centers sit at the center of the current debate because artificial-intelligence demand has changed the scale and urgency of power procurement. A hyperscale campus can require hundreds of megawatts, and a cluster of campuses can reshape a utility's load forecast.

FERC's June action window does not mean every large-load problem will be solved by summer. It means the federal regulator is preparing to answer whether and how standardized interconnection reforms should apply to large loads seeking access to interstate transmission. The decision could affect who studies these projects, how quickly requests move, which costs large customers pay, how utilities protect existing customers, and whether flexible or self-supplied data-center designs receive different treatment.

For developers, the issue is time to power. For utilities, it is reliability and cost. For states, it is local economic development, land use, generation planning, and retail-rate protection. For FERC, the challenge is whether the interstate transmission system needs a clearer rulebook before the AI buildout pushes ad hoc negotiations past their limit.

The docket traces back to DOE's request

The current FERC schedule follows an October 23, 2025 directive from the U.S. Secretary of Energy under section 403 of the Department of Energy Organization Act. The Department of Energy asked FERC to consider an advance notice of proposed rulemaking on the timely and orderly interconnection of large loads to the interstate transmission system. DOE's request reflected the view that large-load interconnection, especially for data centers and advanced manufacturing, had become a national infrastructure issue rather than a set of isolated utility service requests.

FERC opened Docket No. RM26-4-000 for the proceeding. Its public docket page describes the focus as large loads seeking transmission-level interconnection and lists communities affected by data-center or transmission development, utilities, transmission providers, large-load customers, state regulators, consumer advocates, and grid operators as relevant stakeholders. That stakeholder list shows why the docket is difficult: nearly every decision changes someone else's risk.

DOE initially pushed for speed. Energy Secretary Chris Wright asked FERC to complete a final rule by April 30, 2026, according to contemporaneous coverage. FERC did not adopt that exact date. Instead, its April 16 order said the Commission intends to act by the end of June 2026. That gives the agency more time, but not much. For transmission regulation, two months is still a fast decision window.

The difference between April 30 and June is not only scheduling. It suggests FERC wants enough time to weigh jurisdictional boundaries, reliability concerns, and comments from utilities, data-center developers, state officials, and market participants before choosing the next step.

Why data centers are different from ordinary load

Utilities have always connected large customers. Factories, refineries, mines, ports, rail systems, and industrial campuses all require serious planning. Data centers are different in three ways.

First, their growth has accelerated quickly. AI training, AI inference, cloud migration, and digital-service demand have pushed data-center power needs higher and faster than many local load forecasts expected. The issue is not only one large campus. It is many campuses requesting service in the same constrained regions.

Second, data centers can be more flexible than traditional industrial loads in some circumstances, but not automatically. Certain compute workloads can shift across time or location, especially if they are not tied to real-time user demand. Other workloads need high uptime, low latency, and constant power. Regulators and utilities need to know which kind of load is actually being proposed.

Third, data centers may pair with generation, storage, or behind-the-meter resources in ways that blur old categories. Some developers want co-located generation. Some may use batteries, gas generation, renewables, demand response, or microgrid structures. Some may seek transmission service while claiming they can reduce grid impact through flexibility. Those claims need standard study methods and enforceable operating rules.

FERC's large-load docket is therefore not simply about making queues faster. It is about deciding how the grid should evaluate customers whose size, speed, flexibility, and power-supply strategies differ from conventional retail load growth.

Time to power is now a business constraint

For data-center developers, time to power has become as important as land, fiber, tax treatment, water access, and customer demand. A company can buy land, secure cloud customers, and design a campus, but the project does not become useful until it has power. If interconnection studies, network upgrades, transformer procurement, generation additions, or transmission constraints take years, the business case changes.

That is why the FERC docket has market consequences. A clearer process could help viable projects move faster, especially if they are willing to pay for upgrades, accept curtailment, bring flexible load, or pair with new generation and storage. But faster does not automatically mean cheaper or fairer. Existing customers may worry that data centers will push grid costs into general rates. Utilities may worry that speculative requests will clog planning studies. States may worry that federal rules will interfere with local utility regulation.

The phrase time to power can hide those conflicts. Developers use it to describe delayed business opportunity. Utilities use it to describe real infrastructure constraints. Regulators use it to ask whether the rules are allocating scarce transmission capacity fairly.

FERC's June action will likely be watched for whether it creates a fast lane, a clearer study path, a cost-allocation framework, or only another stage of rulemaking.

Cost allocation is the hard question

The central policy fight is who pays. If a data center requires transmission upgrades, substation work, new reliability equipment, or changes to regional planning, those costs must be allocated somewhere. The developer may pay directly, the utility may recover costs through rates, a regional market may socialize some upgrades, or a negotiated structure may split responsibility.

FERC's large-load order said the Commission would consider reforms designed to ensure timely, orderly, and equitable integration of significant electrical loads. The word equitable carries weight. Existing residential, commercial, and industrial customers do not want to subsidize speculative or private data-center expansion. Data-center developers do not want open-ended upgrade costs that make projects impossible to finance. Utilities do not want to be left carrying stranded costs if a large-load request disappears after studies begin.

This is why queue discipline matters. Large-load projects can be speculative. Developers may file requests in multiple places while deciding where to build. If each request triggers expensive studies or capacity reservations, the process can distort planning. On the other hand, if rules are too strict, serious projects may be delayed even when they are willing to fund upgrades.

A workable system needs milestone requirements, financial security, transparent study assumptions, and rules for what happens when a project changes size, location, timing, or operating profile. Without those controls, faster interconnection can become faster confusion.

Co-location adds another layer

FERC has also been dealing with co-located load, where large customers such as data centers sit near or behind generation resources. Co-location can look attractive because it may reduce the need for immediate grid imports or allow a data center to secure power from a nearby generator. But it raises difficult questions about transmission-system use, reliability, backup service, wholesale-market participation, and whether existing grid customers subsidize standby benefits.

FERC's earlier technical conference on large loads co-located at generating facilities listed many of these issues directly. The agenda asked how co-located loads receive wholesale-market services or benefits from the transmission system, how those benefits should be measured for cost allocation, what reliability and resource-adequacy effects may arise, and how studies should account for speculative requests and load flexibility.

In April 2026, FERC also acted on PJM-related co-located load issues, and law-firm analyses noted that data-center developers, generators, and transmission providers should watch how clarified PJM procedures affect siting and interconnection strategies. The PJM fight is not identical to the national large-load docket, but it shows the same policy tension: large customers want speed and optionality; grid operators need reliability and cost clarity.

Co-location will likely remain a major part of the June conversation because many AI-era data-center proposals are trying to solve power constraints by bringing generation closer to load.

Flexibility could change the economics

One reason data centers have attracted federal attention is the possibility that some loads can be flexible. A factory may not be able to shift its power demand easily without stopping production. A data center, depending on workload, architecture, and customer commitments, may be able to move some compute tasks, delay non-urgent processing, use on-site backup resources, discharge batteries, or reduce demand during grid stress.

Researchers are already studying this possibility. A 2026 paper on AI data-center flexibility modeled how flexible AI load might help reduce grid investment and operating costs in some scenarios. Other work on risk-aware transmission capacity allocation has focused on how rapid AI-driven load growth creates interconnection challenges. These studies are not regulatory rules, but they show the direction of the debate: if data centers can provide dependable flexibility, their interconnection treatment may differ from an inflexible 24/7 load.

The word dependable is doing the work. Utilities cannot plan around vague promises that a data center might reduce demand when asked. They need enforceable operating parameters, telemetry, notice periods, penalties, and clarity about which workloads can move. A data center serving latency-sensitive customer workloads is not the same as a batch AI training cluster that can pause or shift.

FERC's June action could therefore encourage clearer definitions of flexible large load. It could also leave much of that detail to regional transmission organizations, utilities, and state regulators. Either way, flexibility is likely to become part of data-center interconnection negotiations.

States are watching the jurisdiction line

Large-load interconnection sits near a sensitive jurisdictional boundary. FERC regulates interstate transmission and wholesale markets. States regulate retail electric service, generation siting in many contexts, local land use, and retail rates. Data centers touch both sides. They are local economic-development projects and large retail customers, but their power demand can affect interstate transmission planning and wholesale-market outcomes.

That is why state officials and members of Congress are watching the docket closely. A letter from Senator Jon Ossoff to FERC in April 2026 asked about strategies if load growth exceeds modeled forecasts and raised questions about DOE's proposed reforms for large loads like data centers. State and local governments are also dealing with community concerns over land, water, noise, tax incentives, and the visible buildout of transmission infrastructure.

FERC will need to act without appearing to override every local concern. A federal rule that speeds large-load interconnection may help national infrastructure goals, but it can create political backlash if communities believe they are absorbing costs or impacts without control.

The likely path is a framework that standardizes transmission-facing process while leaving some retail and siting decisions to states. The exact balance is what makes the June decision important.

Utilities need better demand evidence

Utilities cannot build around rumors. Many power companies have reported large data-center request pipelines, but not every request becomes a live facility. Some developers shop multiple regions. Some projects depend on customer contracts that may not materialize. Some requests change size as AI infrastructure strategies evolve.

That uncertainty complicates transmission planning. If utilities underbuild, real projects wait years and economic development shifts elsewhere. If they overbuild around speculative requests, existing customers may pay for assets that are not fully used. Large-load interconnection reform has to separate serious projects from placeholders.

A stronger process could require deposits, site control, commercial-readiness milestones, signed power agreements, evidence of flexibility, or other commitments before a project receives priority treatment. It could also require better data sharing between developers, utilities, regional grid operators, and regulators.

This is another reason June matters. A faster interconnection process without stronger demand evidence may simply accelerate the wrong projects. A slower process without reform may block real infrastructure. The policy goal is a queue that is both faster and more truthful.

What developers should prepare now

Data-center developers should not wait for June to begin preparing. The projects that move fastest under any new rule will likely be those with clear power strategies, credible milestones, and defensible cost-allocation proposals.

A developer should be able to explain the requested load size, phased ramp schedule, expected load factor, flexibility options, backup configuration, on-site generation or storage plans, grid-service capabilities, and willingness to fund network upgrades. It should also be able to show whether the project is speculative or backed by real customer demand.

Transmission providers and utilities will want clarity on operating behavior. Can the facility curtail? How much? For how long? Under whose instruction? What telemetry will be available? Will the data center draw backup service from the grid when on-site generation fails? Does the project need firm transmission service at all hours, or can some portion be interruptible?

Developers that answer those questions early may have an advantage if FERC's June action rewards project readiness. Developers that rely on generic claims about AI demand and economic benefits may face more scrutiny.

What consumers should watch

For ordinary consumers, the issue is whether data-center growth raises bills, strains reliability, or delays other grid investments. Data centers can bring jobs, tax revenue, and digital infrastructure. They can also require expensive grid upgrades and large amounts of generation capacity. The public-policy question is not whether data centers are good or bad. It is whether the costs and benefits are allocated clearly.

Consumer advocates will likely watch whether large-load customers pay for upgrades that primarily benefit them, whether utility planning assumptions are transparent, and whether existing customers are protected from stranded costs. They will also watch whether reliability risks are addressed before projects connect.

If FERC creates clearer large-load procedures, consumers may benefit from more disciplined planning. If the rules are too developer-friendly, consumer advocates may argue that the process shifts risk to ratepayers. If the rules are too slow or uncertain, data-center developers may turn to more behind-the-meter generation, which can create its own environmental and reliability questions.

The June action is therefore not only a technology story. It is a ratepayer and infrastructure story.

What June can and cannot solve

FERC can clarify transmission interconnection rules, study requirements, cost-allocation principles, and process expectations. It can push transmission providers and regional grid operators toward more consistent treatment of large loads. It can also signal how it views co-location, flexibility, and interstate transmission impacts.

FERC cannot instantly build transformers, transmission lines, substations, gas turbines, batteries, or renewable projects. It cannot eliminate local siting disputes. It cannot make speculative demand certain. It cannot guarantee that every data center gets power quickly without affecting anyone else.

The June decision should therefore be judged as a governance step, not a magic fix. If it creates a clearer rulebook, utilities and developers can make better decisions. If it leaves key questions unanswered, the market will keep solving the problem through one-off negotiations, local moratoriums, behind-the-meter deals, and regional disputes.

The stakes are high because AI infrastructure is moving faster than the grid normally plans. FERC's challenge is to make the interconnection process fast enough for real demand, strict enough to protect reliability, and fair enough that existing customers are not asked to quietly finance someone else's power race.