SEADRIFT, Texas — The steam that leaves a chemical plant usually comes from a gas-fired boiler a few hundred yards away, fed by a pipeline that runs back into the Gulf Coast’s refining complex.
X-energy and Dow want that steam to come from four reactors instead.
On Nov. 9 the Nuclear Regulatory Commission issued a construction permit to a joint venture of the two companies for four Xe-100 units at Dow’s Seadrift site, a chemical complex on the Texas coast. The reactors are designed to produce 320 megawatts of electricity in total, or up to 800 megawatts of industrial steam.
It is the first construction permit the agency has issued for a high-temperature gas-cooled reactor in nearly 60 years. The last commercial gas-cooled plant in the United States shut down in the 1980s.
The permit is a gate, not a finish line. The commission’s process runs in two stages: a construction permit establishes that the design and the site are acceptable, and an operating license, granted later, establishes that the plant can actually run. Between the two sit years of building, testing and inspection.
What makes Seadrift notable is the customer. Most American nuclear projects have been built for utilities that sell electricity into a grid. Dow needs something different: continuous steam at a fixed pressure to drive chemical processes, delivered around the clock, in volumes that a wind farm or a solar array cannot promise.
Industrial heat is one of the harder problems in emissions policy for exactly that reason. Electricity can be decarbonized by building generation and transmission. Process heat usually means burning something on site, which is why chemical plants, refineries and paper mills have moved more slowly than power companies.
The Xe-100 is a pebble-bed design. Its fuel comes in spheres coated in layers of ceramic and carbon, a form engineered to contain radioactive material under accident conditions rather than relying on pumps and operators to prevent one. The reactor is cooled by helium and moderated by graphite, and the design is meant to be built in modules that can be added as demand grows.
That architecture is why the company argues the units suit an industrial site. A chemical plant does not need a gigawatt all at once. It needs firm heat, and it needs the rest of the plant to keep running while one unit is serviced.
The units are small by the standards of the industry, roughly 80 megawatts of electricity each, which is part of the sales pitch: a factory can buy one, learn to run it, and add the next.
A chemical complex of this size runs continuously, and its demand for steam does not fall at night or rise with the weather. That profile is what makes the numbers work in theory: the plant runs at a high utilization rate, so the capital cost is spread across most of the hours in a year, and the fuel bill stays small relative to the output.
The fuel is the other dependency. The coated particles at the center of the design need uranium enriched well above the level used in ordinary power reactors, and domestic production of that material is still being scaled up. Until it is, a reactor design is a drawing with an assumption attached.
Amazon has separately reached an agreement with X-energy for a power project in Washington state, a signal that the buyers of small reactors are increasingly companies that need firm, low-carbon power for data centers and factories rather than utilities serving households.
The obstacles are the ones that have stopped small reactors before. First-of-a-kind construction is expensive, and the cost estimates that matter are the ones produced after the first unit is finished. The supply chain for specialized components is thin. Fuel fabrication capacity has to exist before reactors can be loaded, and it has to be licensed.
There is also the matter of operating experience. The United States has not run a gas-cooled commercial reactor in nearly four decades, which means the workforce that knows how to do it has to be rebuilt at the same time as the machines.
For Dow, the calculation is about price stability as much as emissions. Natural gas prices move, and a plant that buys fuel for decades is exposed to every move. A reactor front-loads the cost into construction and then runs on fuel whose price is a small fraction of the total.
For the regulator, Seadrift is a test of whether a licensing framework built for large light-water plants can handle a different technology without adding years. The commission has been preparing for that shift, and the pace of the review will be watched by every other advanced-reactor developer in the country.
For Texas, the project is a foothold in a model the state has been courting: reactors sited behind the fence of an industrial facility, selling heat and power to the owner rather than to a grid. If it works, the template is portable to refineries, ammonia plants and desalination.
If it does not, the permit becomes a document on a shelf, and the boilers at Seadrift keep burning gas.
The first concrete is expected only after the remaining conditions attached to the permit are satisfied. That is the honest state of the project: approved, not built. Four reactors, one permit, and a plant that will keep making chemicals either way.
