A Regulator Decides the Queue Is the Problem
The single biggest constraint on the AI data center buildout is no longer chips or capital. It is power, and more precisely the multi-year wait to connect large new loads to the electric grid. On June 18, 2026, the Federal Energy Regulatory Commission moved to attack that bottleneck directly. In a unanimous vote, the agency's commissioners issued tailored show-cause orders to six regional grid operators, demanding they justify or overhaul the rules that govern how data centers and other large energy users interconnect to the transmission system. It is one of the most consequential regulatory interventions in the data center era so far.
The orders target PJM Interconnection, the Midcontinent Independent System Operator, Southwest Power Pool, the California Independent System Operator, ISO New England, and the New York Independent System Operator. Together these markets serve roughly 200 million Americans across more than 30 states and the District of Columbia, and account for close to two-thirds of the electricity demand under FERC's jurisdiction. FERC Chair Laura Swett framed the urgency plainly: "We're holding grid operators accountable with tight, ambitious deadlines because the stakes are high and the country demands urgency."
Why Section 206 Is the Sharp Tool
What makes this action notable is not just the substance but the legal mechanism. Rather than open a traditional notice-and-comment rulemaking, which can take years and invites prolonged litigation, FERC reached for Section 206 of the Federal Power Act. Section 206 lets the commission find that existing rates, terms, or practices may be unjust and unreasonable and order operators to show cause why they should not change them. It puts the burden on the grid operators to defend the status quo rather than on FERC to build a new rule from scratch.
The practical effect is speed and durability. Show-cause orders are harder to challenge and harder to stall than a sweeping rulemaking, and they apply pressure to each operator's specific tariff rather than imposing a single national template. FERC identified roughly five reform areas it wants addressed: transmission service application and study processes, cost-allocation mechanisms that prevent cost-shifting, co-location and behind-the-meter generation arrangements, new flexible transmission services for large loads, and evaluation processes for co-located generating facilities. Each operator must now respond on its own terms within a fixed window.
The Clock Is Already Running
The deadlines are deliberately aggressive. Within 30 days, each grid operator must submit a mandatory reliability or resource adequacy report detailing how it will secure enough generation capacity to serve the large loads queuing up to connect. Within 60 days, operators must either justify why their existing pricing structures can accommodate massive data centers or file tariff reforms to fix them. For organizations accustomed to multi-year stakeholder processes, a 60-day window to rework foundational interconnection economics is a genuine shock to the system.
Those timelines signal that FERC wants visible movement before the end of the summer, not a study cycle that drifts into next year. The commission is betting that the threat of imposed reforms will push operators to propose workable changes themselves. Swett has described the broader aim as setting "the stage for a resilient, reliable, and forward-thinking grid that empowers communities and safeguards consumers" while providing "certainty for investors." That last phrase matters: investors funding gigawatt-scale campuses need predictable interconnection timelines as much as they need power itself.
The Real Fight Is Over Who Pays
Underneath the procedural language sits a distributive question that has become politically radioactive: when a data center demanding hundreds of megawatts connects to the grid, who pays for the transmission upgrades it triggers? Historically, many of those costs have been socialized across the broad base of ratepayers, which means ordinary households and small businesses can end up subsidizing the grid investments that serve hyperscaler campuses. FERC's orders push hard in the other direction, instructing operators to design cost-allocation rules that prevent large loads from shifting their upgrade costs onto existing customers.
The commission was careful to mark the boundary of its own authority. FERC said it "acts today to guard against cost shifting among transmission customers but leaves to the states the responsibility to ensure that there is no cost shifting among retail customers." In other words, FERC will police the wholesale transmission layer, but state regulators still own the retail rate question. For data center developers, the message is unmistakable: the era of assuming the grid will quietly absorb the cost of your connection is ending, and large loads should expect to shoulder more of the bill.
What It Means for Data Center Strategy
For anyone planning a data center build inside these six markets, the orders change the calculus in two directions at once. On the positive side, faster and more standardized interconnection processes could shrink the queue backlog that has stranded projects for years, getting compute online sooner. On the cost side, developers should budget for a larger share of interconnection and upgrade expenses landing on them rather than being spread across ratepayers. The net effect is a market that may move faster but prices the true cost of power more honestly into every project.
This also rewards developers who think creatively about generation. The explicit attention FERC gave to co-location and behind-the-meter arrangements signals that operators bringing their own power, whether through on-site generation, dedicated nuclear contracts, or co-located gas and storage, may find a smoother and cheaper path than those simply demanding grid capacity. The strategic implication for CIOs and infrastructure planners is to treat power sourcing as a first-class design decision, not an afterthought handled by the colocation provider.
The Texas Asterisk and the Road Ahead
There is one enormous gap in the map. Texas, home to the fastest-growing concentration of AI data centers in the country, runs its own grid through ERCOT, which operates outside FERC's federal jurisdiction. None of these orders apply there. That carve-out matters because it leaves the single hottest data center market in America governed by an entirely separate set of rules, and it may push some developers toward Texas precisely to avoid the new federal scrutiny. Any national picture of data center interconnection reform now has a Texas-shaped hole in the middle of it.
For the rest of the country, the orders are a clear signal that the regulatory environment around AI infrastructure is maturing fast. Power, cost allocation, and reliability are moving from background concerns to front-and-center constraints that will shape where and how data centers get built. The grid operators now have weeks, not years, to respond, and their filings will set the template for how the United States connects the next wave of gigawatt-scale compute. Enterprises that depend on that compute should watch these dockets closely, because the outcome will influence both the timeline and the price of the AI capacity they plan to consume.
