The humanoid robot company count is now 15 and

I made a spreadsheet. I do this sometimes when a trend starts moving fast enough that I can’t keep it all in my head. The spreadsheet tracks every company actively building a humanoid robot with demonstrated walking capability.

Fifteen. As of January 2024, fifteen companies.

Here they are.

Figure AI. Tesla Optimus. Boston Dynamics Atlas. Agility Robotics Digit. Unitree H1. 1X Technologies NEO. Apptronik Apollo. Fourier Intelligence GR-1. Sanctuary AI Phoenix. Xiaomi CyberOne. UBTECH Walker X. Engineered Arts Ameca. Mentee Robotics. Kepler Forerunner. LimX Dynamics.

Some of these I hadn’t heard of until I started digging. That’s the interesting part. The ones you hear about (Figure, Tesla, Boston Dynamics) are maybe half the story. The other half is happening in Shenzhen, in Tel Aviv, in labs that haven’t given interviews yet.

The automobile parallel

In 1900, there were about 30 companies making cars in the United States alone. Most of them are names nobody remembers. The Duryea Motor Wagon Company. The Winton Motor Carriage Company. Pope Manufacturing. They were building something that most people considered a novelty. A toy for the rich. Unreliable. Dangerous. Slow.

By 1920, cars were industry. The novelty companies had either grown into empires (Ford, GM) or disappeared. The transition from “interesting experiment” to “essential infrastructure” took roughly 20 years.

I think humanoid robots will make that transition in 10.

The reason is simple: the technology stack is more mature. Cars in 1900 required inventing the internal combustion engine, the transmission, the tire, the road, the gas station, and the traffic system. Humanoid robots in 2024 inherit decades of progress in batteries, electric motors, sensors, computer vision, and machine learning. They don’t need new physics. They need new integration.

What’s different about this moment

Three years ago, I wrote about Tesla Optimus walking on stage and it was honestly a little embarrassing to watch. A robot shuffling across a stage like it had just woken up from general anesthesia.

Now I’m tracking fifteen companies with robots that walk, carry objects, and in some cases hold conversations. The progression isn’t linear. It’s a wave. Companies are learning from each other, recruiting from each other, and being funded by investors who can see the same spreadsheet I made.

Figure AI raised capital from Microsoft, NVIDIA, and Jeff Bezos. Tesla has the manufacturing capability to build millions of anything. Unitree is selling robots for under $100,000. Agility is already deploying in Amazon warehouses.

Five of these fifteen companies have announced plans for commercial deployment by 2025 or 2026. Not demos. Deployment. Robots in factories, warehouses, and eventually homes.

Who survives?

Most of them won’t. That’s the honest answer. Of the fifteen, maybe three or four will become major companies. The rest will be acquired, pivot, or run out of money. That’s how every hardware industry works. The car industry started with dozens of manufacturers and consolidated to a handful. The same will happen here.

The question is which handful.

My guess (and it’s only a guess, I’ve been wrong about these things before) is that the survivors will be the ones who solve the hands problem, the cost problem, and the deployment problem. Not the walking problem. Walking is basically solved. Every company on my list can build a robot that walks. The differentiator will be: Can it do useful work? Can you build it for under $20,000? Can you deploy thousands of them?

I don’t know who wins. But I know the race is real, the funding is real, and the pace is accelerating. Fifteen companies. More next month. This is what the start of a new industry looks like.

I kept the spreadsheet as a living document. I’ll update it quarterly. By the end of 2024, I expect the count to be north of twenty.

The feeling

I grew up watching Star Wars and Asimov and I, Robot and dreaming about a world with humanoid machines. For most of my life, that world felt distant. A generational away. Someday. Now I have a spreadsheet with fifteen companies and a timeline that says maybe this decade.

I’m probably too optimistic. I’m often too optimistic about robots. But the spreadsheet doesn’t lie. Fifteen companies, building, walking, talking, grasping.

The industry started.

Related thinking: