Agility Robotics opens training facility near Tesla's Fremont plant as humanoid robot race heats up

Agility Robotics is launching a 60,000-square-foot training center in Fremont, California—strategically positioned near Tesla's manufacturing facility where Optimus robots are set to begin production this year. The move underscores an intensifying competition in the humanoid robotics sector between established players and tech giants entering the space.

A working robot vs. ambitious promises#

Unlike newer entrants still chasing the humanoid dream, Agility's Digit robot is already deployed across real industrial sites. The six-foot-tall machine handles materials handling and bin transport at facilities operated by major customers including Amazon, GXO Logistics, Schaeffler, and Toyota Motor Manufacturing Canada. The company has locked in $300 million in contract orders and reports that Digit has moved more than 100,000 totes at one logistics hub alone.

CEO Peggy Johnson emphasized the importance of practical expertise: the startup has proven it can navigate safety protocols, regulatory requirements, IT integration, and warehouse management systems—barriers many newer competitors have yet to clear. More than 30 potential customers are currently discussing Digit deployments with the company.

The training hub advantage#

The new facility serves as an innovation center where Digit learns new tasks in controlled environments that mirror real-world warehouse and factory conditions. This approach allows Agility to expand its robot's capabilities before field deployment. Version 5 of the robot, arriving this fall, will be enhanced with human-detection sensors, allowing it to operate in shared spaces rather than robot-only zones.

Public markets and generative AI integration#

Agility is moving toward becoming the first pure-play humanoid robotics company to list publicly later this year through a reverse merger. The company is leveraging generative AI to address a core bottleneck: the shortage of engineers who can program robotic behaviors. Co-founder Damion Shelton explained that while AI can't govern safety-critical systems—those must bypass generative models entirely—it can unlock scalability by enabling non-specialists to code new robot applications.