On May 18, 2026 the U.S. Nuclear Regulatory Commission completed the Environmental Assessment for Long Mott Energy LLC—the Dow subsidiary that filed the Construction Permit Application for a four-unit X-energy Xe-100 plant at the Seadrift, Texas petrochemical complex—and issued a Finding of No Significant Impact. The review closed ahead of the 18-month docket schedule that began in May 2025. The Construction Permit itself is now expected in the first quarter of 2027, putting the project on track for first criticality in the early 2030s.
If you read that headline like the rest of the financial press did—320 megawatts electric, four modules, Texas Gulf Coast, another SMR milestone—you missed the actual news. The Xe-100 at Seadrift is not primarily an electricity project. It is the first commercial U.S. nuclear plant whose entire commercial case rests on selling 750°C steam directly to a petrochemical owner who is currently making that steam by burning natural gas. The megawatts are a byproduct. The temperature is the product.
200 MWth Per Module — And Why Thermal Is the Headline Number
Each Xe-100 reactor module is rated at 200 megawatts thermal (MWth) and approximately 80 megawatts electric (MWe). A four-pack delivers 800 MWth and roughly 320 MWe. If you only care about the grid you care about the 320. If you run a hydrocracker, an ethylene cracker, a steam reformer, or a crude unit preheat train, you care about the 800.
Every major refining and petrochemical unit has a temperature envelope. Crude atmospheric heaters run 350–400°C. Vacuum heaters push 400–420°C. Hydrotreater reactor feed effluent exchangers and reformer charge heaters run 480–550°C. Steam methane reformers—the hydrogen factory inside every modern refinery—run tube-skin metal temperatures north of 900°C with process gas at 800–850°C. The Xe-100’s helium coolant exits the core at over 750°C. That single number puts the reactor inside the working envelope of every refinery utility load short of the SMR tube itself. No other commercial U.S. reactor design—not the AP1000, not the BWRX-300, not the Natrium intermediate loop—walks out of the core that hot.
That is the entire commercial thesis of the Seadrift project. Dow’s Union Carbide Seadrift plant currently runs a gas-fired cogeneration unit to supply process steam and on-site electricity. Replacing it with a four-pack of Xe-100s eliminates roughly 440 megawatts of continuous fossil heat duty and the carbon that goes with it. The reactors will deliver process steam at industrial useful conditions—not at the de-rated, post-condensing-turbine conditions that come off a conventional nuclear plant’s extraction steam.
TRISO Fuel and the Walk-Away Safety Case
Co-locating a fission reactor with a tank farm and an ethylene cracker is not a thing the NRC has done before. The Xe-100 can be considered because the fuel form rewrites the accident envelope. Each Xe-100 core contains hundreds of thousands of tennis-ball-sized graphite pebbles. Inside each pebble are roughly 18,000 TRISO particles—poppy-seed-sized uranium kernels enriched to about 20% U-235 (HALEU, high-assay low-enriched uranium) wrapped in three concentric layers of pyrolytic carbon and one of silicon carbide. The silicon carbide layer functions as a containment vessel at the particle scale. It holds together through 1,800°C—well past any credible loss-of-coolant temperature in this reactor.
The consequence is a reactor that cannot melt in the operational sense of the word. In a complete loss of forced helium circulation the pebble bed self-limits through negative temperature reactivity feedback and passive thermal radiation to the reactor cavity cooling system. There is no pressurized water inventory to flash. There is no zirconium to oxidize and generate hydrogen. This is the engineering reason the NRC’s Environmental Assessment came back clean enough for a FONSI, on a site adjacent to a chemicals complex, ahead of schedule.
The Fuel Supply Question Is Already Answered
The historical objection to HTGRs in the United States has been fuel. TRISO at commercial scale has not been a thing here in over fifty years. On February 13, 2026 the NRC issued TRISO-X a Special Nuclear Material license under 10 CFR Part 70 for the TX-1 fuel fabrication facility in Oak Ridge, Tennessee—the first new commercial nuclear fuel facility licensed in the U.S. in more than half a century, and the first ever NRC Category II commercial fuel plant. Vertical construction began on November 17, 2025. TX-1 will produce 5 metric tons of uranium per year as roughly 700,000 finished pebbles—enough fuel for eleven Xe-100 modules annually. Operations begin in 2027. Initial output goes to Seadrift.
This is the part of the project that most readers underweight. A reactor design without a fuel supply chain is a research artifact. The TX-1 license, the construction start, and the Seadrift fuel allocation collectively mean the HTGR fuel question in the United States is closed. The next-cohort question is whether the U.S. can scale HALEU enrichment—currently the Centrus Piketon cascade and DOE downblend allocations—to keep TX-1 fed. That is a separate fight.
Cascade — The Data Center Side of the Same Reactor
The same Xe-100 module that solves Dow’s steam problem also solves Amazon’s power problem. The Cascade Advanced Energy Facility—Energy Northwest’s site outside Richland, Washington, funded by Amazon—will deploy an initial four Xe-100 modules (320 MWe) with an option to expand to twelve (960 MWe) to power AWS data centers. Construction targets late this decade, operations in the early 2030s, same family of reactor, same fuel, same NRC docket templates.
The interesting thing about reading Seadrift and Cascade in parallel is that one reactor design is now being deployed simultaneously into the two largest industrial electricity loads in the country—hyperscale compute and Gulf Coast petrochemicals—under different commercial structures (Dow as steam offtaker, Amazon as power offtaker) and the unit economics work for both. The Xe-100 is the only U.S.-licensed advanced reactor whose temperature, modularity, and fuel form let it serve both a chemical plant and a data center campus with the same module and the same fuel pebble.
What This Means For the Refining Sector
U.S. refineries collectively burn roughly 1.5 quads of fuel gas, refinery off-gas, and natural gas per year to generate process heat and on-site steam. The hydrogen plants alone—steam methane reformers feeding hydrotreating and hydrocracking—account for somewhere between 6–10% of total refinery CO₂ emissions and rise above 30% at hydrocracker-heavy facilities. Replacing the SMR firebox with a nuclear-supplied 750°C steam loop is, on paper, the single largest decarbonization lever available to a refinery without changing its product slate.
The honest counter-arguments are real. Electric steam crackers like the BASF/SABIC/Linde demonstration plant in Ludwigshafen are operating now and hit the same 850°C envelope with grid power. Coolbrook’s RotoDynamic Heater hits process temperatures with electricity and a smaller footprint than a reactor. Both approaches avoid the licensing, security, and public-engagement overhead that comes with co-locating fission and feedstock. But both require firm, low-carbon, baseload electricity at industrial scale—which, in Gulf Coast markets, increasingly means natural gas with carbon capture or, eventually, nuclear at the wholesale level. Seadrift is the bet that the cleanest path is to skip the conversion to electrons entirely and deliver the heat directly.
The Timeline That Actually Matters
The dates worth tracking are tight. NRC Construction Permit issuance: Q1 2027. TX-1 fuel operations: 2027. Operating License Application: anticipated 2028–2029. First criticality: early 2030s. Cascade Phase 1 commercial operation: also early 2030s. The supply chain—HALEU enrichment, TRISO pebble fabrication, Mitsubishi Heavy Industries pressure vessel forging capacity, helium circulator manufacturing—is sized for roughly two to four modules per year through 2032 and then scales with demand. The narrow constraint is HALEU, not vessels and not pebbles.
For refiners watching this from outside the nuclear industry, the practical signal is simpler. If you are scoping a fifteen-to-twenty-year capital plan for a hydrogen plant rebuild, a captive cogen replacement, or a major steam-system overhaul, the next economically rational option after combined-cycle gas with CCUS is no longer hypothetical. It has a docket number, a fuel supply, an industrial customer, and a Q1 2027 construction permit on the calendar.
The Dual-Use Angle
Process-heat reactors and industrial-scale AI infrastructure are the same procurement decision viewed from two sides of the meter. The same Xe-100 four-pack that lets Dow decarbonize a chemicals complex lets a hyperscaler underwrite a campus without negotiating a fifteen-year merchant power deal. The same TRISO fuel facility that supplies Seadrift supplies Cascade. Federal evaluators looking at industrial-decarbonization and AI-infrastructure-resilience as separate problems are going to find—within twenty-four months—that the supply chain that solves one solves the other.
Porritt Inc. builds AI tools for process safety, engineering compliance, and refinery operations. If you are scoping how a process-heat reactor integrates with an OSHA 1910.119 PSM program, a Mechanical Integrity program, a Management of Change workflow, or an industrial steam-system optimization study, that is the conversation we have every week. Reach out at porrittinc.com for a consultation or a NORMEX Standards AI demo.
Timothy Porritt is founder of Porritt Inc., building AI-powered tools for process safety, engineering compliance, and industrial operations. Based in Salt Lake City, Utah.