Agility Robotics is establishing a 60,000-square-foot training facility for its humanoid robots in Fremont, California — a location that places it just up the highway from the factory where Tesla is expected to begin manufacturing its Optimus robots this year. The move signals intensifying competition in the humanoid robotics space, even as Agility positions itself as a company with a head start in commercial deployment.
Tesla has been ramping up its investment in Optimus, with CEO Elon Musk recently stating he expects it to be "the biggest product ever" once it becomes "useful outside of Tesla sometime next year." While Tesla brings enormous capital to the table, Agility counters with a robot that is already working — and earning money — in real-world environments.
Digit Already Generating Revenue in Real-World Settings
Agility's humanoid robot, Digit, is already carrying totes and bins in manufacturing and warehouse facilities for a roster of customers that includes Amazon, GXO, Schaeffler, and Toyota Motor Manufacturing Canada. The company reports that it has secured $300 million in contract orders for its robots.
Although Agility has not disclosed exactly how many Digit units it has built or deployed, outside observers estimate that dozens are currently working in pilot or revenue-generating roles. The company has noted that Digits have moved 100,000 totes at a GXO logistics facility alone.
CEO Peggy Johnson told TechCrunch that having Tesla nearby is a positive development for the industry. For a long time, Agility was operating largely alone in the humanoid space, and having other companies enter the field is welcome. Johnson emphasized that Agility has already commercialized its technology and understands what it takes to integrate robots into existing facilities — meeting safety standards, regulatory requirements, and compliance benchmarks, while connecting to customers' IT infrastructure and warehouse management systems.
A Practical Approach to AI and Autonomy
While transformer-based neural networks — the same technology underlying large language models — promise significant advances in robotic behavior, Agility is taking a measured approach to artificial intelligence. Co-founder and chairman Damion Shelton drew a comparison to self-driving cars, noting that critical safety systems like anti-lock brake controllers should not be placed under AI control. Similarly, he argued, the safety stack in humanoid robots must follow a path that is not governed by generative AI, because creativity is not desirable when it comes to safety.
Where AI does prove invaluable, according to Shelton, is in scaling robot applications. He recalled that when board member Bruce Leak — the inventor of QuickTime — first asked how the company planned to code applications for the robot, Agility did not have a strong answer. The range of tasks a robot could potentially perform far exceeds the number of engineers available to program them, and Shelton said generative AI definitively answers that question.