Figure 02 Humanoid Robot

A dated reading of what is claimed, reported, and independently verified in the current evidence.

Current reading

Readiness band and full integer triple

AI-assisted assembly · derived results

Readiness band

Watch

Primary summary from verified readiness

Evidence strength Weak

Decision

Human editorial judgment · 2026-07-23

Proceed with caution

Why: The BMW run genuinely proves ~11 months of real production throughput at a fixed task, but the CEO's public 'fleet' framing did not match BMW's confirmed single-robot/single-task reality, OpenAI's language/vision layer was dropped after ~1 year, and a since-filed lawsuit contests whether the safety program (e-stop certification) was actually delivered — so the throughput case is proven but the safety-governance case is not.
Next: Treat Figure 02 as a closed one-robot/one-task case study proving narrow-task industrial viability, not a template for fleet-scale deployment; evaluate the successor generation (Figure 03 / Helix 02) directly and require independent verification of e-stop certification and force-limiting before any deployment near humans.

Constraints

No named blocker is present in the current public projection.

Evidence summary

AI-assisted assembly

Counts and dates only. Raw signals, private excerpts, trust records, and internal corpus material are not published here.

Publication record

Initial public reading