Six orders, one theory
On June 18, 2026, the Federal Energy Regulatory Commission opened Section 206 proceedings against every RTO and ISO it regulates simultaneously: PJM (Docket EL26-67), SPP (EL26-68), NYISO (EL26-69), MISO (EL26-70), CAISO (EL26-71) and ISO-NE (EL26-72). Each order makes the same preliminary finding — that the grid operator's existing tariff "may be unjust, unreasonable, or unduly discriminatory or preferential" because it lacks provisions built for how data centers and other very large loads actually behave, from speculative interconnection requests to sudden withdrawal after transmission upgrades are already under construction (CAISO-hosted order text, Docket EL26-71; FERC Commission meeting presentation, Items E-7 to E-12).
Each RTO and its transmission owners have until August 17, 2026 — 60 days from issuance — to either defend the status quo or file Section 205 tariff revisions. A separate 30-day informational filing on generation adequacy was due July 20, and FERC set a September 16 deadline for stakeholder comment (McGuireWoods client alert). FERC's working definition of a "large load" is peak demand above 50 MW connecting above 69 kV — squarely inside the range of a modern AI training campus.
What FERC actually wants changed
The orders group reforms into five buckets: interconnection application and study procedures; cost transparency and cost-shifting protections; co-location arrangements between data centers and generators; flexible-load and behind-the-meter service classes; and terms for loads sited electrically close to existing large loads. The centerpiece is the cost recovery agreement — a mechanism requiring the host utility or the large load itself to post financial security covering a minimum share of network-upgrade costs, so that if a data center developer walks away after a transmission owner has already built the interconnection, the bill doesn't land on residential ratepayers instead (RMI, "Understanding FERC's Large Load Orders").
That's not a hypothetical risk. FERC's own market data shows why the agency moved on all six regions at once rather than waiting for individual dockets: more than 50 GW of data center capacity was operating nationally at the end of 2025, a 24% compound annual growth rate since 2020, and MISO — the region FERC says has grown fastest — saw data center capacity expand at a 43% CAGR over the same period (Utility Dive, MISO/FERC market report). Average facility size more than tripled, from 25 MW in 2020 to roughly 80 MW for projects energized in 2025. Queues built for occasional large industrial loads were never designed for that.
The case for FERC's intervention
The strongest argument for these orders isn't about speed at all — it's about who pays. Interconnection queues nationwide are already clogged with speculative requests, a problem FERC spent years addressing on the generation side through Order No. 2023. Large loads present the mirror image: a developer can file for gigawatt-scale service at several sites simultaneously, trigger expensive network studies and upgrade obligations at each, and then build at only one. Absent standardized cost-recovery terms, transmission owners either eat that cost or spread it across every other customer on the system — including households with no stake in the AI buildout. Treating that as a ratepayer-protection question, not merely a permitting-speed question, is a legitimate exercise of FERC's Federal Power Act mandate to keep tariffs just and reasonable, and MISO itself conceded to FERC that its tariff doesn't "provide a consistent or transparent framework" for evaluating these requests.
Where the orders risk overshooting
The design risk is that six simultaneous, nationwide Section 206 dockets — each inviting intervention, briefing and eventual litigation over rehearing — become exactly the kind of multi-year slog that Order 2023 became on the generation side, at the moment US AI infrastructure competitiveness is most sensitive to interconnection timelines. That risk is already showing: every RTO/ISO has since asked FERC for additional time to respond, reportedly seeking roughly three more months beyond the August 17 deadline (RTO Insider, "Every ISO/RTO Asks FERC for an Extra 3 Months on Large Loads") — a sign the timeline FERC set may already be more aggressive than six regionally distinct grid operators can execute credibly.
FERC should hold the line on the underlying goal — cost-causers pay for the infrastructure they trigger — while resisting the temptation to impose one uniform national template across grids as different as CAISO, with its renewable-heavy, transmission-constrained footprint, and MISO, which is absorbing data center growth nearly twice as fast. The orders explicitly say they won't unwind agreements large loads have already negotiated, which is the right instinct: retroactive cost-shifting onto developers who built under the old rules would chill exactly the investment the country needs. Flexible-load tariffs — curtailable service, on-site generation, demand response — are a better long-run tool than new firm-service categories, because they let data centers absorb some grid risk themselves rather than forcing FERC to draw a single national line between "large load" and everyone else. The August 17 filings will show whether FERC's proportionality survives contact with six utilities each insisting their region is the exception.