Synthesis ยท evidence map

US-China robotics: two evidence curves, not one horse race

China industrial deployment scale, US frontier humanoid / AI concentration, Tesla capacity evidence, Figure deployment KPI and Leaderdrive supplier filings belong on one evidence ladder โ€” not in a single winner ranking.

Updated 2026-06-15. Evidence map only; no trade recommendation.

Four blocks to keep separate

China scale is industrial-deployment scale

IFR reported 295,000 industrial robot installations in China in 2024, 54.4% of global installations, operational stock above 2.027m and domestic supplier share rising to 57%. This is a real deployment flywheel, not humanoid economics proof.

๐ŸŸข/๐ŸŸ 

US frontier is AI / humanoid / mega-round concentration

Figure raised US$675m at a US$2.6bn valuation, while Tesla, Figure, Apptronik, Agility, NVIDIA, Physical Intelligence and Skild represent a concentrated frontier stack. Capital and AI stack are signals, not S5 economics.

๐ŸŸข/๐ŸŸก

Tesla and Figure answer different questions

Tesla Optimus is a filing-backed capacity-intent and vertical-manufacturing curve; Figure is a customer-site deployment KPI plus manufacturing-ramp curve. Capacity is not production; deployment KPI is not full economics.

๐ŸŸข/๐ŸŸ 

Leaderdrive is a supplier-signal case, not a CATL/NVIDIA proof

Leaderdrive's 2025 revenue, profit, reducer unit sales and robot-relevant revenue line are strong filing-backed S4 signals. Missing before the stronger analogy: named humanoid design wins, humanoid-only revenue, order volume, per-robot value content and margin durability.

๐ŸŸข/๐ŸŸ 

How to read the comparison

China has a larger industrial robot deployment flywheel and a stronger domestic supplier penetration signal in factory automation. That does not automatically prove humanoid customer economics.

The US has unusually concentrated frontier signals around AI/model/compute stack, private humanoid mega-rounds and company-level S4 evidence. That does not mean Optimus designed capacity is current production or that Figure BMW / BotQ KPI proves full unit economics.

Leaderdrive analogy test

็ปฟ็š„่ฐๆณข is best treated as a strong filing-backed component-volume signal. The CATL/NVIDIA analogy needs additional proof: named humanoid design wins, humanoid-only revenue, architecture durability, margin durability and a platform-like expansion path. Product fit is not named customer proof.

Footer

Evidence map only. No winner ranking. No trade recommendation. Industrial robot scale โ‰  humanoid economics; designed capacity โ‰  production; supplier product fit โ‰  named customer proof.

Sources