A sodium cooled test reactor in Idaho went critical at 20 minutes past midnight on the fourth of July, and that date had been printed in the groundbreaking release ten months earlier

There is nothing to look at.

A low windowless box on a gravel pad, sagebrush pushing up to the edge of the grading, a bank of insulated pipe leaving one wall and turning into a trench.

No cooling towers. No dome.

The groundbreaking release, issued on the twenty eighth of August the previous year, named the fourth of July as the target for completion and criticality.

The building went up in 36 days, and the whole plot was bare desert a few months before that.

Something inside it started a chain reaction and kept it going.

What it did not do, and was never built to do, is make a single watt.

Why a sodium cooled core can be built this fast

Start with what a conventional plant has to survive.

Water carries the heat away, and to keep that water liquid at operating temperature it has to be held at about 2,200 pounds per square inch.

That happens inside a steel vessel the size of a house, inside concrete thick enough to stop artillery.

That vessel is the schedule.

Forging it, shipping it, setting it and pouring around it is most of the decade that a large reactor takes before a single fuel rod arrives on site.

Sodium removes the pressure.

Liquid sodium boils at roughly 1,621 degrees Fahrenheit at ordinary atmospheric pressure, so the coolant loop never has to be pressurized at all, the vessel gets thin, the containment shrinks and the critical path stops running through a forge.

What the sodium costs you back

It burns on contact with air and reacts violently with water.

That is why the program was split down the middle into two separate buildings on the same site, one nuclear and one not.

The fueled unit proves the physics.

A second facility next door is being built to circulate 60,000 pounds of sodium under real operating conditions and prove out the heat exchangers that would eventually turn a turbine, and it is not commissioned.

The core itself is small and conventional in its materials.

Thirteen fuel assemblies of uranium dioxide enriched to 4.95 percent, in a hexagonal graphite lattice, with the rods assembled at a rate of 540 in two days and change.

What is actually on the record

Criticality at 12:20 in the morning, Mountain Time, on the fourth of July.

Three other advanced reactors had already done it that summer, on the fourth of June, the eighteenth of June and the thirtieth of June.

All four ran under an executive order and a departmental standard rather than a commercial license.

That pathway is the second reason for the speed, because it let the department authorize the test directly instead of routing it through the years long process that has historically stretched novel designs across a decade.

The date itself was not a surprise.

The groundbreaking release, issued on the twenty eighth of August the previous year, named the fourth of July as the target for completion and criticality.

The symbolism was fixed ten months out, and the engineering was scheduled to meet it.

What eight months does not include

The company calls it eight months from breaking ground to a sustained chain reaction.

Ground was broken in late August and criticality came on the fourth of July, which is ten months and change, with heavy site work reported as starting only in the new year.

Either count is fast. Neither is eight.

The larger gap is what criticality means, because this is a zero power physics reactor and the 10 megawatt figure attached to it everywhere belongs to the commercial design rather than to the machine that went critical.

Nothing was generated and nothing was delivered.

Federal commercial authorization sits entirely ahead of the team, the sodium loop next door is unfinished, and the same laboratory is still targeting criticality for its own microreactor around the turn of the year.

What the data center changes

This was never aimed at a municipal grid.

The commercial unit is a 10 megawatt reactor, and five of them share one turbine in a pod rated at 50 megawatts, sized to stand beside a computing campus rather than feed a region.

That is a demand profile a small reactor fits unusually well.

Constant output, compact footprint, no exhaust, no fuel convoy every week, and capacity bought in 50 megawatt steps instead of a full gigawatt commitment made a decade in advance.

The waste argument needs one correction.

This is a graphite moderated design rather than a fast reactor, so the usual fast spectrum spent fuel argument does not apply to it, though the country still has no licensed repository for any class of high level waste.

A commercial scale unit at the same site is already announced for electricity in 2027, with an on site data center attached, and four companies now have criticality on the board.

What happened in Idaho is real and narrower than it reads.

Uranium fission sustained itself in a full scale core, inside a building that did not exist a year earlier.

That proves the schedule and it proves the physics, and a test reactor that makes no power still has every commercial step in front of it.