Hardware

Figure Destroys F.02 Humanoid Robots in Molten Steel

Robotics startup Figure has decommissioned its F.02 humanoid robot fleet by having them autonomously jump into a vat of molten steel to protect its proprietary hardware.

Superhuman AI3 days agoHardware
Image: Superhuman AI

Figure AI has retired its fleet of F.02 humanoid robots in a highly unusual manner, sending the machines to jump autonomously into a 75-ton vat of molten steel at a foundry in Finland. The dramatic decommissioning stunt, which was inspired by a suggestion from actor Arnold Schwarzenegger, aimed to destroy the older hardware securely to prevent proprietary technology from falling into the hands of competitors. The viral video of the event has already accumulated over 1.8 million views on social media as Figure transitions its focus to its expanding F.03 fleet.

This dramatic retirement comes amid a broader wave of commercial robotics developments. DoorDash recently secured an FAA Part 135 air carrier certification for its DoorDash Air drone fleet, which will handle mid-distance deliveries of 3 to 5 miles. In 2025, these drones achieved an average delivery time of 25 minutes. Meanwhile, a new report from Anthropic highlights the economic hurdles facing the industry. While existing robots can perform tasks representing 34 percent of US working hours, they are currently more cost-effective than human workers for only 0.3 percent of those tasks. Anthropic estimates it could take 40 years for that economic viability figure to reach 10 percent.

Other notable industry updates include Amazon planning to invest over $100 million in an Indiana manufacturing hub by 2028, and Tesla reportedly building hundreds of Optimus robots weekly, though most remain in internal testing. In software and hardware optimization, developers are utilizing tools like the CompactifAI API, which offers GLM 5.3 at $1.10 and $3.50 per 1 million input and output tokens, and the Quasar 438B model at approximately 59 percent cheaper than average. For robotics practitioners, these developments highlight the rapid pace of hardware iteration and the critical need to secure proprietary designs as the industry scales.

This is our own summary of reporting by Superhuman AI

More in Hardware