Today in Humanoid Daily: $1.7 billion for robots at Atoms, Tesla Optimus, and robot surgery

Biggest story today: it’s got to be Atoms and Travis Kalanick

It’s the robotics startup founded by Uber founder Travis Kalanick. They just raised $1.7 billion in one of the largest robotics funding rounds we’ve ever seen. It’s led by Andreessen Horowitz. Uber also invested, and Andreessen Horowitz co-founder Ben Horowitz is taking a board seat. 

That’s a pretty strong signal that there’s something going on here beyond just an investment. This is a big deal that he cares about.

What’s kind of remarkable ... even very remarkable: Atoms hasn’t even released a product yet: a robot, a humanoid, anything. I don’t know, maybe we’re back to raising money on scratches on napkins. We haven’t seen that since the dot-com era. Maybe only if you’re the founder of Uber!

This kind of capital changes the landscape overnight. It means Atoms can hire super aggressively, pick people up from competitors, move fast, and build manufacturing partnerships. And when Travis Kalanick built Uber into a global transportation giant, he showed he has the ability to do it once. He can probably do it again.

Second story: Ropedia has raised $22 million to solve one of robotics’ biggest bottlenecks: training data. 

We all know training data is a problem. Training data is necessary. More, more, more is necessary. We’ve trained LLMs on internet-scale data. We don’t have that for robotics or for human motion.

So this system, called Homie, is a lightweight headset with four cameras that records human actions from a first-person perspective. Instead of teleoperating a robot, you simply do stuff. You perform tasks naturally, and you generate rich data sets that AI can learn from.

As embodied AI races ahead, and as we need more and more of it for robots and humanoids, high-quality data is becoming super valuable for training better and better models.

Third story, in healthcare: Medtronic has previewed Touch Surgery AI

We’re seeing more and more of these. I got pitched on a different one earlier this week as well. This is an AI platform designed to run directly inside an operating room.

Rather than analyzing surgeries after they’re finished, it performs real-time AI inference on surgical video and instrument data. 

And guess what? It’s another sign that surgical robotics is becoming as much about AI software as robotic hardware.

You probably saw the story a couple of weeks ago about the teleoperation of those AgiBot robots that performed surgery on a live subject. It happened to be a pig, not a human. 

We’re getting more and more of this, and we’re getting more and more surgical robots as well.

Finally, China has approved its first national standard for evaluating humanoid robot data sets

It’s actually kind of a big deal. It’s a little technical, but standards really matter. If everybody trains and evaluates robots the same way, it becomes much easier to share models, compare performance, and scale development.

It’s an example of building the infrastructure around humanoid robots, not just the robots themselves.

Finally, today’s number of the day is 46

46 is the number of days that Tesla needed to completely dismantle its Model S and Model X production lines in Fremont, California, to make room for, you guessed it, Optimus Gen 3 production.

Today, that job is done. 

Those lines had been building cars for 14 years and 11 years, respectively. Replacing them with humanoid robot manufacturing is yet another sign, if we needed any more, that Tesla sees its future in humanoids, not just EVs ... maybe even more in humanoids than EVs.

That’s it for today. If you enjoyed this episode, like, subscribe, and there’s much, much more at humanoidmegahub.com.