Agility Robotics will open a 60,000-square-foot facility to train humanoid robots in Fremont, California, just up the highway from the factory where Tesla is expected to begin manufacturing its Optimus robot this year.
Tesla is increasing its bet on Optimus. Elon Musk recently said he expects it to be the “biggest product ever” “if it’s available beyond Tesla within the next year.”
Agility doesn’t have Tesla’s capital, but it already has a robot, Digit, that can help in the real world. The robot is already generating revenue by transporting totes and bins in manufacturing and warehouse environments for customers such as Amazon, GXO, Schaeffler, and Toyota Motor Manufacturing Canada. The company announced it has secured $300 million in robot contracts.
“It’s great that[Tesla]is in the same space as us, because actually for a long time agility has been there on its own, but it’s good to have other people in the humanoid space,” CEO Peggy Johnson told TechCrunch. “We’ve commercialized. We now know what it takes to go into these facilities, meet safety bars, regulatory bars, compliance, connect to IT infrastructure, connect to warehouse management systems.”
Agility has not disclosed how many Digits it has built or deployed, but outside observers estimate that dozens were in pilot or revenue-generating deployments. The company announced, for example, that Digits moved 100,000 packages at its GXO logistics facility.
Johnson is currently leading Agility through a reverse merger, which is expected to become the first pure-use humanoid robot company to hit the public market later this year. Founded in 2015 by a group of researchers who developed new technology that allows robots to safely walk on two legs, Agility is looking to capitalize on its lead over a new generation of AI-inspired robot startups such as Figure, 1X, The Bot Company, and Sunday Robotics.
While the advent of transformer-based neural networks, which helped birth LLM, also promises major advances in robot behavior, Agility takes a pragmatic approach to autonomy.
“When you think about self-driving cars, you know that as a non-humanoid example, you definitely don’t want to put an anti-lock brake controller under AI control,” Damion Shelton, co-founder and chairman of Agility, told TechCrunch. “The analogy with humanoids is all the safety stuff that you have to go through that isn’t a generative AI path, right? You don’t want to get creative with the safety stack.”
However, what AI actually does is promise scale.
“The first time[Quicktime inventor Bruce Leake, who is on Agility’s board of directors]asked me how you go about coding applications for robots, I didn’t have a good answer,” Shelton said. “The number of things you can imagine a robot to do is far greater than the number of engineers who can program it. And generative AI will definitively answer that question.”
The new facility is designed to accelerate the company’s robotics deployment. Johnson said more than 30 customers are in talks with the company to implement Digit, and the new facility will be a place for 6-foot-tall robots to learn new skills in an environment similar to what they would experience in the field.
Unlike many new entrants into the humanoid space, Agility has no plans to offer home humanoid robots anytime soon. This view is consistent with that of most independent robotics experts, who believe that today’s most powerful robots are not safe enough for consumer use. Digit currently operates in a space without humans, but version 5, scheduled to be released this fall, will have the ability to detect humans, eliminating the need for it to be kept in a robot-only zone.
Co-founder and chief robotics officer Jonathan Hurst said there’s plenty of work to keep Agility busy in manufacturing and logistics alone.
“Start with bins and totes, then pick and kit,” Hurst told TechCrunch. “Then let’s start working on cardboard, which is really hard. Loading and unloading things like tractor trailers. Okay, now we have 100 million robots. You know, it’s a trillion dollar company.”
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
