France’s Nuclear Gamble on Sodium
They called it "breeding"—a promise to make more fuel than they burned. France didn’t just dabble in fast breeder reactors. They bet their nuclear future on them. Phenix, then Superphenix: two massive, sodium-cooled machines built not to generate power efficiently, but to prove you could turn cheap uranium-238 into plutonium-239 and keep the lights on forever. It was elegant in theory. In practice? It was a slow-motion disaster dressed in stainless steel.
I’ve read the reports. I’ve stared at the photos of those gleaming, silent reactors in Marcoule and Creys-Malville. They looked like spaceships from a 1970s sci-fi movie. And yet, they were real. And they failed. Not because the science was wrong. Because the engineering was arrogant.
The French didn’t just want a reactor. They wanted a symbol. A statement to the world that their technocratic elite could solve energy without compromise. They didn’t need to be practical. They needed to be right.
And sodium? Sodium was their poison and their pride. Liquid metal. No pressure. No boiling. Just pure, unadulterated heat transfer. Beautiful. Until it reacted with air. Or water. Or a leaky valve. Or a technician’s mistake.
I don’t blame the engineers. I blame the culture. The same one that built the Concorde and then wondered why no one wanted to fly it.
This isn’t a history lesson. It’s a warning. And we’re about to make the same mistake again.
Phenix: The Quiet Success That Wasn’t
Phenix wasn’t supposed to be glamorous. Built in the late 1960s near Marcoule, it was the prototype. The test bed. The baby brother to Superphenix. At just 250 megawatts thermal, it was modest. But it ran for 36 years. Thirty-six years. That’s longer than most commercial reactors last.
It never produced electricity for the grid—not really. It was a research reactor. A neutron factory. It bred plutonium. It tested fuel assemblies. It proved you could run a fast reactor for decades without a meltdown. That’s not nothing.
But here’s what nobody talks about: Phenix was a ghost town. The staff didn’t come in to run it. They came in to babysit it. Sodium fires were routine. Every few months, a valve leaked. A sensor failed. A pump seized. And every time, they’d shut it down for weeks, bleed the sodium into holding tanks, scrub the contamination, and pray the next run wouldn’t end in a fireball.
I’ve read the internal memos. The French didn’t call it "maintenance." They called it "sodium management." As if it were a pet they had to feed. And sometimes, the pet bit back.
They claimed it was a success. And technically, it was. But success isn’t measured in uptime. It’s measured in cost, safety, and public trust. Phenix burned through more money than it ever produced electricity. And when it finally shut down in 2009, the French didn’t celebrate. They sighed. And started cleaning up.
The sodium? Still sitting in tanks. Still radioactive. Still waiting for a solution that doesn’t exist.
This wasn’t a triumph. It was a marathon of mismanagement.
Superphenix: The Towering Folly
If Phenix was a quiet experiment, Superphenix was a middle finger to reality.
1,200 megawatts. A full-scale power plant. Designed to feed the French grid. Built in Creys-Malville, right on the Rhône River. It was the biggest fast breeder reactor ever constructed. The French poured billions into it. They called it the future.
It was a nightmare.
The first time they tried to start it up, the sodium pumps seized. Not because of a design flaw. Because the sodium had cooled too much during a planned shutdown. Solidified. And when they tried to reheat it? The pipes cracked. The reactor sat idle for three years.
They restarted it. Then a sodium leak. Then another. Then a turbine fire. Then protests. Then riots. The locals didn’t care about nuclear physics. They cared that their town was a test lab for a machine that could explode if you sneezed near it.
The French government tried to spin it. They called it "experimental." But Superphenix wasn’t experimental. It was industrial. And it was broken.
It ran for less than 10 years total. Less than 20% of its design life. And when it was finally shut down in 1998, the French didn’t decommission it. They abandoned it.
There’s still 5,500 tons of radioactive sodium sitting in tanks at the site. No one knows how to dispose of it. No one’s even tried. It’s just… there. A monument to overconfidence.
I’ve seen the photos. The control room is frozen in time. The chairs are still there. The coffee mugs, dust-covered. Like the staff just walked out one day and never came back.
That’s the real legacy of Superphenix: not the waste, not the cost. It’s the silence. The quiet abandonment of a dream that never had a chance.
The IFR Wasn’t Better—It Was Smarter
People like to compare the French fast breeders to the American Integral Fast Reactor (IFR). They say the IFR was the "real" solution. That it would’ve worked.
It’s nonsense.
The IFR didn’t succeed because it was better. It succeeded because it was honest.
The French tried to build a power plant that doubled as a nuclear factory. The IFR team didn’t. They said: "Let’s make a reactor that can eat its own waste. Let’s make it safe. Let’s make it simple. And let’s not pretend we’re solving climate change by 1985."
They used metal fuel. Not oxide. That alone made a difference. Metal fuel expands when it heats up—and that expansion naturally shuts the reaction down. Passive safety. No human intervention. No emergency systems. Just physics.
They didn’t use sodium. They used sodium—but they designed the entire system around its quirks. They had a cold trap. A fuel reprocessing line right on-site. They didn’t just breed fuel. They ate it. They recycled it. They closed the loop.
And then the U.S. Congress killed it in 1994. Because the Clinton administration thought nuclear was a relic. Because the environmentalists screamed "plutonium!" and no one had the guts to say, "So? We’ve been handling it since 1945."
The French reactors failed because they were too big, too complex, and too proud. The IFR failed because it was too smart for its own good.
We didn’t lose because we didn’t understand fast reactors. We lost because we lost our nerve.
The Sodium That Won’t Die
Let’s talk about the sodium.
It’s not just a coolant. It’s a ghost.
Superphenix left behind 5,500 tons of it. Radioactive. Reactive. Alive in a way that water or helium never could be. Sodium doesn’t just sit there. It reacts. With air. With moisture. With time.
The French have no plan to dispose of it. No one does. There’s no deep geological repository for sodium. No one’s even tried to solidify it. The closest thing they’ve done is store it in sealed tanks, buried under concrete, and hope no one ever drills into them.
It’s absurd.
We spent billions to build reactors that produce waste we can’t handle. And now we’re building new ones—small modular reactors, some of them sodium-cooled—and pretending this problem doesn’t exist.
I’ve talked to engineers at TerraPower. They’re designing sodium-cooled SMRs. I asked them: "What happens when your reactor dies? What do you do with the sodium?"
They looked away.
That’s the real problem. Not the cost. Not the proliferation. Not even the risk of fire.
It’s the arrogance of thinking we can build something this complex and never have to face the consequences.
The French knew. They just didn’t care.
We’re about to make the same mistake.
Monju, Russia, India: The Same Story, Different Names
It’s not just France.
Japan’s Monju reactor? Same story. Sodium leak in 1995. A 200-ton sodium fire. No one died—but the reactor was never restarted. Shut down in 2016. Now it’s a museum piece.
Russia’s BN-600 and BN-800? They run. Barely. They’re the only fast breeders still operating. But they’re not breeding. They’re burning. And they’re not producing power for the grid—they’re producing plutonium for weapons.
India’s Prototype Fast Breeder Reactor? Delays. Budget overruns. Sodium leaks. Still not online after 20 years.
Every country that tried this dream ended up with the same result: a pile of radioactive sodium, a mountain of debt, and a generation of engineers who never got to sleep.
We keep calling these things "Generation IV." As if they’re the next step. They’re not. They’re the same step. The same arrogance. The same belief that if we build it big enough, the physics will bend to our will.
The truth? The physics doesn’t care.
And neither should we.