The most consequential object in the advanced-reactor buildout is not a reactor. It is a particle of uranium roughly the size of a poppy seed, wrapped in four engineered coatings, one of which is a microscopic ceramic pressure vessel. It is called TRISO — tri-structural isotropic coated-particle fuel — and almost every high-temperature reactor now being proposed to power data centers, hydrogen plants, and industrial heat loads depends on it. Understand the particle and you understand why the siting conversation has changed.

For sixty years, the rule of nuclear siting was distance. Large light-water reactors are licensed with emergency planning zones measured in miles, because their safety case has to account for fuel that can melt and a coolant that can flash to steam and carry fission products with it. The advanced-reactor pitch — a reactor compact enough to sit beside the load it powers — only works if the fuel itself cannot fail in the ways that drive those zones. That is the claim TRISO is built to make good on.

Five Layers, One Containment Vessel Per Particle

A single TRISO particle starts with a kernel: a sphere of uranium oxide or uranium oxycarbide (UCO) a few hundred microns across — typically around 0.4–0.5 mm. That kernel is where fission happens. Everything built around it exists to keep the products of fission where they were born.

Working outward, the kernel is coated with four layers, each with a distinct job:

• Porous carbon buffer — a spongy inner layer that absorbs recoil energy from fission fragments and gives fission gases somewhere to go as the kernel swells, so internal pressure does not crack the shell.
• Inner pyrolytic carbon (IPyC) — a dense carbon layer that seals the buffer and provides a clean surface for the next coating.
• Silicon carbide (SiC) — the load-bearing layer. This is the miniature pressure vessel: a dense ceramic shell, on the order of 35 microns thick, that holds in gaseous and metallic fission products and gives the particle its strength at temperature.
• Outer pyrolytic carbon (OPyC) — a final carbon layer that protects the SiC during fuel fabrication and bonds the particle into its surrounding matrix.

The finished particle is under a millimeter across. Hundreds of thousands of them are then bound in a graphite matrix and pressed into one of two forms: pebbles — graphite spheres roughly the size of a billiard ball, each holding on the order of tens of thousands of particles — or compacts, cylindrical slugs loaded into the channels of hexagonal graphite blocks for prismatic-core reactors. Pebble-bed designs and block-type designs disagree on geometry but agree on the particle.

The design philosophy is radical in its simplicity: instead of one large containment building backstopping fuel that is assumed to fail, you give every particle its own containment and qualify the population. A reactor core can hold billions of these vessels. The safety case becomes statistical and material, not architectural.

Why 1600 °C Is the Number That Matters

The headline property of TRISO is fission-product retention at temperature. Across the U.S. Department of Energy’s Advanced Gas Reactor (AGR) fuel qualification campaign — irradiation tests run at Idaho National Laboratory and examined over more than a decade — coated particles were driven to high burnup and then heated in post-irradiation safety tests to temperatures well above any credible accident, and the SiC layer continued to hold its inventory. The practical threshold cited for the onset of meaningful coating degradation sits around 1600 °C.

That number is the whole game, because it is compared against a second number the reactor designer controls: the peak temperature the fuel can reach with all active cooling gone. In a high-temperature gas reactor, two design choices keep that peak below the threshold. First, low power density — these cores run at a few megawatts per cubic meter, against roughly a hundred for a light-water reactor — so there is simply less heat per unit volume to get rid of. Second, a tall, slender core geometry that lets decay heat escape by conduction and radiation to the surrounding structure and ground, with no pump, no operator action, and no AC power required.

Put those together and you get the phrase the industry leans on: walk-away safe. It is not marketing if the physics closes. If the maximum achievable fuel temperature in a complete loss of forced cooling stays under the coating’s damage threshold, the fuel does not release its inventory, and the accident that defines large-reactor siting cannot occur in the same form. That is the argument for putting one of these machines near a campus instead of a county away.

Inert Coolant, High Outlet Temperature, and the Heat Nobody Talks About

Most TRISO-fueled designs are cooled by helium or, in the fluoride-salt variants, by molten FLiBe. Helium is chemically inert, single-phase, and transparent to neutrons — it will not burn, will not boil, and will not become radioactive in the way water activation products do. That lets the coolant leave the core hot: outlet temperatures around 750 °C are typical for the gas-cooled pebble designs, and some concepts push higher.

High outlet temperature is the feature the electricity headline tends to bury. A reactor that delivers 750 °C heat is not only more efficient at making power; it can deliver process heat directly — to hydrogen production, chemical synthesis, desalination, and the high-temperature steps of industrial manufacturing that are otherwise stuck burning natural gas. For an AI campus, that heat can be turned into reliable around-the-clock electricity. For a refinery or chemical plant, it can displace combustion at the point in the process where decarbonization is hardest.

This is the part of the story closest to our own work. Porritt Inc.’s Project Genesis — a 25,000-barrel-per-day modular micro-refinery — is designed for an initial operating run on fission heat, with engineered tie-in taps to retrofit fusion heat sources once those are tested. The reason a refinery program cares about coated-particle fuel is the same reason a hyperscaler does: it is the enabling technology for clean, dispatchable, high-grade heat delivered next to the load. The particle that makes the reactor safe to site is the same particle that makes the heat worth siting for.

The Bottleneck Is the Enrichment, Not the Particle

If TRISO is so capable, why is it not everywhere already? The honest answer is fuel supply, not fuel physics. Advanced reactors generally need HALEU — high-assay low-enriched uranium, enriched above the 5% ceiling of conventional power-reactor fuel and up to just under 20%. Higher enrichment is what lets these compact cores run long, efficient cycles. But the commercial supply chain for HALEU has historically been thin, and standing up domestic enrichment and conversion capacity has been the rate-limiting step for the whole class of reactors.

The manufacturing of the particle itself has also had to scale from a laboratory and demonstration footprint to a commercial one — dedicated coated-particle fuel fabrication lines that can turn out the volumes a fleet would consume. Both of these — enrichment and fabrication — are supply problems being worked on actively, and both are areas a reader should check against the latest program status rather than trust to any single article.

What the Particle Means for the Buildout

The strategic point is that the constraint on siting clean firm power next to large loads has shifted. It is no longer primarily a reactor-architecture problem; it is a fuel-supply and licensing-throughput problem. The coated particle has already done the hard physical work — it turns “meltdown” from a system-level failure mode into a per-particle material limit that the designer can engineer against. What remains is industrial: making enough qualified fuel, and moving licensing at the speed the load growth demands.

For anyone evaluating where to put capital or attention in the energy-and-AI convergence, that reframing is the useful one. Watch the fuel. The reactor vendors will compete on geometry and integration, but they are drinking from the same well, and the depth of that well — qualified HALEU and qualified TRISO at volume — sets the ceiling on how fast firm, clean, high-temperature power can show up beside the data centers, refineries, and industrial plants that need it. The poppy-seed-sized vessel is the quiet enabler of all of it.

---

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.