
Humanoids Cost Less Than Funerals
Posted May 15, 2026
Chris Campbell
March 2011. The wave hits Fukushima.
Cooling systems? Gone. Reactors? Melting.
Tokyo Electric Power Company—TEPCO—needs people inside the plant. The kind who can turn valves, climb stairs, and read dials at eye level.
Brave Japanese workers volunteered by the hundreds.
The problem was one of dosage.
Radiation inside the reactor buildings was so high a worker could spend minutes—not hours—before hitting their annual legal exposure limit. Burn through your whole crew in a day and you've still barely started.
So the company sent in the robots they had—tracked units, wheeled crawlers, industrial machines. But the robots failed.
They couldn't open the doors. Couldn't turn the valves. Couldn't climb to the rooms that mattered.
The plant was built for humans. Every valve is sized for human hands. Every staircase pitched for human legs. Every control panel hung at human height. A machine on treads could survive the radiation, but it couldn't reach the dials.
Fifteen years later, the cleanup is barely past the beginning. TEPCO just pushed the start of full debris extraction from one reactor to 2037. The government's 2051 completion target is already considered unreachable. Eight hundred and eighty tons of melted fuel still sit inside the reactors.
Partly because the right machine didn't exist in 2011. Partly because it still doesn't.
And it’s one angle of the humanoid boom that barely anyone’s talking about.
Tyranny of the Doorknob
Most see Tesla's Optimus or Figure AI and think factory floor. Cheap labor. Repetitive tasks. A few hundred grand saved per shift.
That’s a real market. But it’s also not the one that’ll spark the humanoid revolution.
The same robot built to stack boxes at BMW—two legs, two arms, eye-level cameras, hands that grip—is the exact body built to walk into a meltdown, a collapsed building, a deep-sea rig, or a battlefield.
Every nuclear plant, hospital, oil platform, and disaster zone we've built in the last two centuries was designed around the human gait. Human-shaped infrastructure needs human-shaped replacements.
Consider where it already makes the most sense…
Currently waiting for an entry. Not waiting on the thesis.
