Figure sorts 250,000 packages autonomously

- Figure AI said on May 24 its F.03 humanoid robots completed a 200-hour autonomous warehouse run, sorting, scanning and placing more than 250,000 packages. - The company said the robots handled roughly one package every three seconds without human supervision, extending a livestreamed warehouse test into a multi-day stress run. - Figure’s latest public benchmark follows its April 29 production update and earlier Helix logistics posts on the company’s news site.

Figure AI said its F.03 humanoid robots completed more than 200 hours of fully autonomous package sorting in a live warehouse run, processing more than 250,000 packages without human supervision. The company said the robots picked parcels, oriented labels for barcode scans and placed items onto conveyors at roughly one package every three seconds. The result was shared publicly over the weekend through a post on X and follows a series of Figure updates aimed at showing its humanoids can move from lab demos to sustained logistics work. Figure has not disclosed the warehouse operator or released an independently audited breakdown of error rates, downtime or labor substitution. ### How much did the robots actually do? Figure said the run crossed 200 hours and 250,000 packages, which puts the system near continuous operation for more than eight days. At the stated pace of about one package every three seconds, the throughput is broadly consistent with the package total the company reported over that span. A May 2026 report on the company’s livestreamed test said the demonstration began as an eight-hour session and continued after the robots kept operating without what Figure described as a failure. Figure presented the exercise as a stress test for reliability and endurance in a narrow warehouse workflow. ### What task were the humanoids performing? Figure’s warehouse task centers on small-package handling: grasping mixed parcels, turning labels toward scanners and placing items onto a conveyor. (cryptobriefing.com) In a June 2025 logistics update, the company said its Helix system had expanded from rigid boxes to poly bags, padded envelopes and other deformable parcels, while improving throughput to about 4.05 seconds per package and orienting labels correctly for scanning about 95% of the time. (rockingrobots.com) That earlier post matters because it shows the warehouse run was not a new task invented for a one-off clip. Figure had already been publishing package-handling benchmarks and describing logistics as an early commercial use case for Helix, its vision-language-action control system. ### Why is Figure emphasizing a 200-hour run? Figure has been trying to show that humanoid robotics is moving from short demonstrations to repeatable industrial operation. (figure.ai) On April 29, the company said it had delivered more than 350 third-generation robots from its BotQ manufacturing facility and increased production from one Figure 03 per day to one per hour in under 120 days. That production update tied fleet growth directly to data collection and commercial deployment. (figure.ai) Figure said robots shipped from BotQ were being allocated to research, data collection and commercial use-case development, suggesting the warehouse run is part of a broader effort to prove both manufacturing scale and field reliability. ### What does Figure say is different about F.03? Figure introduced Figure 03 in October 2025 as its third-generation humanoid, saying the machine had a redesigned sensory suite and hand system built for Helix, lower manufacturing cost and features intended for high-volume production. (figure.ai) The company said the robot’s camera architecture doubled frame rate, cut latency and widened field of view, while embedded palm cameras and new tactile sensing were designed to improve grasping and manipulation. Those hardware changes are relevant to warehouse work because package sorting depends on close-range perception, quick adjustment to different parcel shapes and stable grasping when labels are wrinkled or partly obscured. That is an inference from Figure’s published specifications and logistics benchmarks. ### What is still missing from the public record? Figure’s public materials describe package totals, runtime and task type, but they do not provide a third-party audit of mis-sorts, barcode read failures, intervention rates or total system economics. (figure.ai) The company also has not publicly identified the warehouse customer associated with the latest run. The next public markers are likely to come from Figure’s own update cycle. (figure.ai) The company’s news page shows recent posts on Figure 03 production and Helix deployments, and its April 29 manufacturing note said robots were already being directed into commercial use-case development beyond headquarters testing. (figure.ai) (cryptobriefing.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.