Figure AI says it’s building robots that will build more robots.
The rising Silicon Valley startup just unveiled its BotQ smart manufacturing facility, designed to produce 12,000 humanoids a year. Figure’s plans call for 100,000 robots within the next four years.
The robot factory is on Figure’s new San Jose campus, where the company intends to launch new facilities every 90 days. Figure is building out the high-rate manufacturing facility as it develops the next-generation version of its flagship humanoid, which will presumably be called the Figure 03.
The company shared a stylized video showing industrial robots putting the finishing touches on a Figure 03 head.
On the X social media platform, Figure founder Brett Adcock said BotQ will be the world’s largest humanoid robot factory by volume.
“This will be a critical year as we scale up to unprecedented humanoid robots,” Adcock said.
It will be the second facility for mass producing humanoids in the United States. Agility Robotics opened the first plant in late 2024 in Salem, OR, where the company plans to scale to 10,000 units a year. Like Figure, Agility Robotics is also busy developing the next-generation version of its humanoid, Digit.
Meanwhile, Tesla plans to manufacture 10,000 Optimus humanoid robots by the end of 2025, though Elon Musk has admitted the target is aspirational.
The perceived leaders in the emerging market are racing to mass produce their AI robots amid an escalating technological arms race with China. Shanghai-based AgiBot says it’s already started manufacturing humanoids at scale, with a thousand robots coming off its production lines in 2024.
According to Figure, the third-generation robot is designed from the ground up for high-rate manufacturing. In a blog post, the company predicted that Figure 03 robots will be working on the production floor this year.
A small fleet of Figure 2 humanoids were deployed at the BMW plant in Spartanburg, SC in January 2025 after a pilot program the prior year. The robots have also been put to work for another major unnamed customer for logistics.
They’re powered by Figure’s self-developed AI system called Helix. Adcock is so confident in the end-to-end neural network that he’s pushed home testing forward two years. Figure’s marketing has recently started emphasizing household use cases after initially focusing on industrial applications.
For the consumer market, Figure faces competition from another well-funded startup, 1X Technologies, which plans to test its NEO Gamma humanoids in select homes this year.
Figure’s high-volume factory is designed to expand quickly. Engineers spent six months architecting a manufacturing execution system (MES) that tracks every step of the production process. Info from Figure’s website, blog posts, and job listings indicate they use Dassault Systèmes’ Enovia for product lifecycle management (PLM), Oracle NetSuite for enterprise resource planning (ERP), and a warehouse management system (WMS) integration for tracking inventory.
They’ve switched from CNC machining to injection molding and die casting to speed up production. Parts that previously took a week to make now take under 20 seconds. Figure says it’s making parts in-house due to supply chain challenges.