The Robotics Economy.
A map of 125 public robotics companies.

Robots aren't new. Factories have run them for fifty years. But they were dumb — bolted in place, looping one motion, useless the second anything changed. The body was never the problem. The brain was, and nobody could build it. Then the same AI that runs ChatGPT turned out to run robots too. Now you show one a task a few times and it handles the rest, in a room it's never seen. Goldman Sachs looked at that and raised its forecast for the market sixfold. The surprise, they said, wasn't the hardware. It was how fast the AI got good.

And the demand for them is real. The working-age population is shrinking in China, Japan, and most of Europe — fewer people every year to do the work. At the same time companies are pulling factories back home, where there aren't enough workers to staff them. Half a million US manufacturing jobs sit open right now, and the sector is on track to need 3.8 million workers by 2033, with nearly 1.9 million of those expected to go unfilled. That gap doesn't close on its own. Robots are what fills it.

One catch worth knowing before treating the theme as simple: China already makes about 70% of the parts inside these robots. Strip the Chinese suppliers out of a Tesla Optimus and it costs roughly three times as much to build. That's the real risk in the theme — and it's why the few US-listed companies that make this stuff are worth finding. Finding them is the point of the map below.

For scale: a humanoid cost about $200,000 to build in 2024. At volume, makers are targeting $20,000 — less than a year's wages. Morgan Stanley puts the market at $5 trillion a year by 2050.