The Department of Energy's draft 2026 National Transmission Needs Study makes it official: load growth — led by data centers — is now the main reason the U.S. needs new transmission. Congestion already adds $11 billion a year to wholesale power costs, tens of billions in new lines are approved, and the study feeds directly into federal corridor designations that can override local siting authority. Public comments are open until September 7, 2026 — and the draft is full of numbers communities can use.
Every three years, federal law requires the Department of Energy to tell the country where its electric grid is congested, constrained, or about to be. The draft 2026 National Transmission Needs Study, released for public comment in July, is that report — and this edition reads differently from every one before it.
The headline isn't buried. In DOE's own words, load growth is now the main driver of future transmission need — and the first item on its list of what's driving load growth is "growth in data centers, artificial intelligence (AI), and cryptocurrency mining." The agency devotes a call-out box to data-center projections showing demand reaching more than 400 TWh by 2030 — over 10% of all U.S. electricity.
If you're a resident wondering whether the data center proposed outside town has anything to do with the new 500 kV line proposed through it, the federal government just answered you.
| What the study found | Number |
|---|---|
| Congestion's added cost to wholesale power, 2023 | $11 billion (peaked at $21B in 2022) |
| Share of hours driving most congestion costs | just 5% |
| U.S. demand growth projected by NERC, 2024–2034 | +25% (4,281 → 5,353 TWh) |
| Data-center demand by 2030 (LBNL/EPRI/McKinsey range) | up to 400+ TWh, >10% of U.S. total |
| FERC's own range for data-center growth by 2030 | 13 to 55 GW — a 4x spread |
| Transmission built 2016–2024 | 85,000 circuit-miles — 98% by incumbent utilities |
| Aging-grid replacement need (Brattle est.) | ~4,000 circuit-miles / ~$10B per year |
| Newly approved regional buildouts | MISO $21.8B · SPP $7.7B · Texas 765 kV $33B |
This isn't just a study — it's legal machinery. Under Federal Power Act §216, the Needs Study is the analytical foundation DOE uses to designate National Interest Electric Transmission Corridors (NIETCs). Inside a designated corridor, federal regulators gain backstop siting authority — meaning a transmission project rejected or delayed at the state or local level can still move forward.
Translation for communities: the maps in this draft are an early look at where federal authority may eventually reach. The time to influence that is now, during the comment period — not after a corridor lands on your county.
1. You now have federal validation. When a developer's consultant tells your planning board that a data center has nothing to do with rising delivery charges, you can quote the Department of Energy: load growth, led by data centers, is the principal driver of the country's transmission needs. Print the study's data-center box (p. 80) and bring it.
2. The bill for all of this is unassigned. The study identifies need — it is silent on who pays. Transmission costs flow into the delivery portion of every electric bill by default, and utilities earn a regulated return on every mile they build (which is one reason incumbents built 98% of it). Unless your state adopts large-load tariffs and cost-causation rules, a buildout justified by data-center demand lands on households. That's the fight at your PUC, and the study is your exhibit A.
3. Nobody actually knows how much of this demand is real. FERC's own projection spans 13 to 55 GW by 2030 — the top of the range is four times the bottom. Speculative and duplicate interconnection requests ("phantom load") are inflating forecasts everywhere. If lines get built for load that never materializes, ratepayers hold the bag for decades. The ask: load commitments, minimum-take contracts, and developer-funded upgrades before shovels move — so the party creating the risk carries it.
4. The cheapest fix is flexibility, not always wire. The study notes most congestion costs pile up in just 5% of hours — the same peak hours when data centers could curtail. That finding lines up almost exactly with what Duke University's Nicholas Institute published in February 2025: Curtailment-Enabled Headroom, led by Tyler Norris, found the existing grid could absorb up to 98 GW of new large load if that load agreed to curtail just 0.5% of annual hours — roughly 44 hours a year — avoiding $150 billion or more in new power plants and transmission lines. DOE says congestion concentrates in 5% of hours; Duke says you only need the flexible loads to stand down in 0.5% of them. Texas already wrote a version of this into law (SB 6: ERCOT can curtail large loads in emergencies). Communities and PUCs should demand the same: flexible-load agreements and grid-enhancing technologies evaluated before approving billion-dollar buildouts that peak-shave for a handful of hours a year. (We break down the Duke numbers — and the SLA and rate-design reasons operators resist them — in "Why Your Electric Bill Is Going Up — and What Data Centers Have to Do With It" elsewhere on this blog, with the full research library in the Your utility bill tab.)
One section deserves more attention than it will get: 2.3% of the nation's transmission miles cross Tribal lands, and over 1,000 high-voltage substations sit on or near them — 919 in SPP South alone. Yet Navajo Nation and Hopi Tribe homes still lack electricity, and DOE's Office of Indian Energy found energy costs consume a 28.3% larger share of income for households on Indian land than elsewhere. Infrastructure that crosses a community without serving it is the oldest story in energy — and it's a preview of what data-center-driven buildout does wherever the benefits and burdens land on different people.
The grid is being rebuilt around AI demand either way. The only open question is whether communities shape the terms — and this study, for all its dry federal prose, just handed them the evidence.
Sources: DOE 2026 National Transmission Needs Study — Draft for Consultation and Public Comment, July 2026; Norris, T. et al., Curtailment-Enabled Headroom: How Flexible Large Loads Can Accelerate Decarbonization, Nicholas Institute for Energy, Environment & Sustainability, Duke University, February 2025.