Germany’s THTR-300 was a thorium high-temperature, helium-cooled pebble-bed reactor rated at roughly 750 MW thermal and 308 MW electric — a machine built on gorgeous theory, confident scale-up assumptions, and mechanical systems that sometimes behaved like a shopping cart with one bad wheel.

Planning kicked off in the 1970s. It reached first criticality in 1983, tied to the grid in 1985, achieved sustained full-power operation in February 1987, and was shut down in September 1989 after a short, bruising operating life with an average capacity factor of 45%. It was the commercial-scale follow-on to the AVR: same pebble-bed philosophy, same coated-particle (TRISO-type) fuel, same helium cooling — just bigger. And scale-up, as reactor history keeps reminding us, is where elegant ideas go to take their professional engineering exam.

Here’s the part most people miss: the fuel wasn’t a neat, static bundle like an LWR. The core held about 670,000 spherical fuel elements, each roughly 6 cm (2.4 inches) in diameter. Think 670,000 graphite billiard balls continuously moving through the core — measured, sorted, routed, and reinserted — with mind-numbing reliability required at every step. FOREVER. That’s not just a reactor. That’s a high-temperature neutron engine welded to a complex 24/7 logistics system.

On paper, TRISO fuel is legitimately tough. That’s real.

But THTR-300’s recurring Achilles’ heel wasn’t the fuel kernel — it was the machinery and instrumentation required to move and track a small planet’s worth of pebbles without jamming, miscounting, or losing traceability. Pebbles didn’t always move as predicted. Measurements didn’t always agree with reality. And if you can’t confidently say where your fuel is and what burnup it has, you’re not operating a reactor — you’re hosting a neutron lottery.

In May 1986, a stuck-pebble event during fuel handling led to a radioactive dust/aerosol release through the handling exhaust path. Subsequent reviews documented procedural deviations and problematic operator reporting around irregularities. Not catastrophic in dose — but reputationally brutal — and the timing couldn’t have been worse, landing in the same year public nuclear confidence was already shattered by the Chernobyl disaster.

Here’s the broader pattern: advanced reactors often suffer from the same optimism — liquid sodium, molten salt, liquid lead — beautiful thermodynamics wrapped in Rube Goldberg machinery. The physics is elegant. The hardware is temperamental.

Physics doesn’t fail first.

Machinery does.