Small Modular Reactor Permit Clears TVA's Clinch River BWRX-300

A regulatory first with the hard part still ahead

On September 29, the Nuclear Regulatory Commission gave the Tennessee Valley Authority a construction permit for Clinch River Nuclear Unit 1. The unit is a 300-megawatt BWRX-300 small modular reactor designed by GE Vernova Hitachi, planned for TVA's site near Oak Ridge, Tennessee.1 It is the first US construction permit for this design. The Breakthrough Institute and the NRC both describe it as the country's first construction authorization for a commercial small modular reactor.23 TVA counts it as only the second advanced-reactor construction permit ever issued in the United States. TerraPower's Natrium permit came first, in March.26

The coverage agrees on the main facts. The permit lets TVA build the reactor, but it does not let TVA load fuel or run it. That requires a separate operating license from the NRC.25 The coverage disagrees on what the milestone means. Tech outlets have presented it as nuclear power arriving for the AI era. Trade and construction reporting describes a utility that still hasn't committed to building anything. The second account is closer to the truth, but the first explains why so much money is lining up behind the project.

A faster regulator, by design

The review speed is the clearest result. The NRC finished in 14 months, four months earlier than it had scheduled.23 Jeremy Bowen, who directs the agency's Office of Advanced Reactors, laid out the steps at the signing. The safety evaluation decision came nine months after TVA applied, staff finished their safety review at 12 months, and the permit was signed just under 14 months in.4 NRC Chairman Ho Nieh presented the result as evidence that a reformed agency can make credible safety decisions without holding major projects back.6

The speed has reasons beyond agency culture. TVA already held an Early Site Permit for Clinch River, granted in 2019, covering a combined 800 MWe. That meant site suitability, environmental impact and emergency planning had largely been settled before this review began.28 The BWRX-300 also builds on the licensing basis of GE Hitachi's larger ESBWR boiling-water design, so regulators faced less new technology than they would with a truly novel reactor.18 One analysis suggests the design's history gave it a real speed advantage over SMR concepts still early in licensing.17

The reports differ on small details. Some sources say TVA filed its application in April 2025, while others say May.13 The NRC formally began its review in July 2025.3 These differences don't change the main point: a US small modular reactor got through construction licensing in about the time a large data center campus takes to plan.

No shovel, no price, no final decision

Construction reporting brings the story back to earth. TVA told ENR that nuclear construction has not started. It also said the permit has not triggered new construction packages, orders for long-lead equipment or expanded contractor work. TVA's board still has to approve a final commitment to build.7 So far the board has authorized $350 million for the project: $200 million in 2022 and another $150 million in August 2024.7 The Department of Energy separately chose TVA in December 2025 for up to $400 million in cost-shared federal funding.7

Those figures are small next to the likely total. Citing Knox News, one account reports that TVA estimated in 2024 that a single SMR would cost about $5.4 billion before interest and tax credits.8 Several outlets note that TVA hasn't disclosed a final cost estimate or set a construction start date.1416 The timeline is also unclear. TVA earlier aimed to finish the first reactor in fiscal 2032, while other reporting cites a projected commercial operation date of December 2033.810

TVA's own wording shows where the risk sits. The utility says it is still negotiating with industry partners over cost sharing. Interim CEO Mike Skaggs said the priority is making sure people in the Tennessee Valley don't pay for a first-of-a-kind technology whose benefits reach well beyond the region.3 Skaggs also said that lining up customers for the reactor's output will matter for deployment.7 In practice, the permit is no longer the obstacle. The question is who signs the checks.

The data center pressure behind the permit

Data center demand explains why the question of who pays is now urgent rather than abstract. TVA spent years planning for almost no load growth. Before its 2019 resource plan it saw flat demand for several years and expected that to continue until 2029.21 After the pandemic, demand jumped and has kept rising.21 Former CEO Don Moul said data centers grew to 18% of TVA's industrial load in 2025, and he expects data center demand in the region to double by 2030.22 TVA's preliminary 2026 integrated resource plan said actual and forecast demand is already outrunning its reference case and nearing the high-growth scenario, mainly because of AI and hyperscale data centers.23

The planning numbers are large. TVA expects to need between 11 and 32 gigawatts of new generating capacity through 2040, both to meet growth and to replace retiring plants.21 Industrial Info Resources tracks about $50 billion in TVA generation projects and says more than half of that is earmarked for SMRs.21 One Clinch River unit, at 300 MW, is a small piece of that range. That gap matters, and it isn't always clear in coverage that presents the permit as an answer to AI power demand.

Natural gas, not nuclear, is doing most of the near-term work. The same resource plan identifies incremental needs of 7 to 26 GW of gas generation between now and 2040.23 Nuclear is the long-term bet, while gas fills the gap now. A 2030s reactor can't serve data centers that are being energized this decade.

Making data centers pay

TVA's rate decisions this summer may matter as much for Clinch River's finances as the NRC permit does. In August, TVA's board voted unanimously to move data centers out of the manufacturing rate class into their own class. Their bills will rise by about 10%. New data centers will also pay upfront capacity commitment charges of roughly $1.5 million per megawatt, spread over three to five years.24 The new rate took effect October 1 and is tied to what TVA calls a Ratepayer Protection Pledge. Under it, hyperscalers are expected to build or buy new power supplies and pay for the grid upgrades their facilities need.25 TVA also said it is open to data center customers bringing their own generation.25

Read together, these decisions point to a likely funding model. This is analysis, not something TVA has announced. A first-of-a-kind reactor that TVA doesn't want households to pay for, and a class of customers now formally expected to pay for new firm capacity, could be a natural match. Skaggs's remark about securing customers for the output points the same way.7 Critics are wary. The Southern Alliance for Clean Energy has argued that decisions justified by expected demand surges could still push costs onto ordinary customers. TVA also hasn't explained how it would deal with stranded assets if a data center closes.

Data center reactor, or utility reactor?

Some tech coverage called Clinch River a "data center-specific" reactor.11 That goes too far. Clinch River is a utility project for TVA's whole seven-state grid. US Senator Marsha Blackburn described it as powering Tennessee homes and businesses.8 Projects built directly around data centers are a different group. Blue Energy filed the first part of an NRC construction permit application in September for BWRX-300 units at a gas-plus-nuclear plant in Victoria, Texas, intended to power data centers.1516 What Clinch River offers developers like Blue Energy is a licensing precedent. TVA says the permit sets out a repeatable path for future BWRX-300 applicants.3

That repeatability is the real economic argument. TVA is considering as many as four units at Clinch River.14 Ontario Power Generation has laid the foundation for the first of four BWRX-300s at Darlington. Fleet projects are also moving forward in Poland and the UK.1418 Supporters expect a shared design to cut costs through common procurement, manufacturing and construction experience.14 Skeptics point out that this hasn't been shown in the US, where recent large nuclear builds ran far behind schedule and over budget.11

The bottom line

The Clinch River permit shows that US nuclear licensing has stopped being the main bottleneck. It doesn't show that small modular reactors can be built on time, at a known price, at the scale data center growth requires. For TVA, the next steps are a board vote and a signed cost-sharing deal, ideally with large power buyers paying their share. Until those happen, the AI boom is the reason Clinch River has momentum, but not yet a source of electricity.