Figure

What Figure is building

Figure describes itself as an AI robotics company focused on bringing a commercially viable general-purpose humanoid to market.

Its hardware has moved through three public generations. Figure 01 established the early platform, Figure 02 was used in workforce deployments, and Figure 03 is the current generation positioned for both home and commercial environments.

The company’s strategy increasingly links four pieces: the humanoid body, the Helix AI system, the Index human-behavior dataset and manufacturing at BotQ.

From prototypes to a fleet

The important company-level question is no longer whether Figure can build a single humanoid prototype. It is whether the company can manufacture, train and operate a growing fleet consistently.

In April 2026, Figure reported that BotQ had delivered more than 350 Figure 03 robots. The company also said it had increased its production rate from one Figure 03 per day to one per hour in under 120 days.

Those are company-reported production figures. Tomorrow, Tested. has not independently audited BotQ throughput.

Industrial experience

Figure’s public deployment story includes BMW Group Plant Spartanburg.

According to Figure, Figure 02 contributed to production of 30,000 cars during 2025. In June 2026, the company showed Figure 03 returning to BMW for a sequencing logistics workflow that required manipulation while stepping and repositioning.

This is useful evidence that the platform has moved beyond staged household demos, although the operational performance figures remain primarily disclosed by Figure itself.

The move toward home robotics

Figure 03 was redesigned with home use in mind: softer exterior materials, changes to the sensory system and hands, wireless charging, audio improvements and battery-safety work.

The difficult part is not merely making a robot that fits physically inside a home. Homes are unstructured, crowded and different from one another. Figure’s recent Helix work therefore focuses heavily on transferring learned behavior into unseen environments.

What to watch

For Figure, the biggest evidence gap is shifting from “can the robot perform this task?” to “how reliably, at what cost and across how many uncontrolled environments can it perform it?”

Manufacturing scale, independent deployment data, service requirements, safety performance and eventual customer pricing will matter as much as visually impressive demonstrations.

Related Tomorrow, Tested. pages track Figure 03 hardware and the latest Helix results separately so company claims, robot specifications and AI evaluations do not get collapsed into one story.