China builds 85% of the world’s humanoids robots for cheap at scale, but finding buyers is tricky

China’s humanoid robot market is scaling quickly, with more than 13,000 robots shipped in 2025 and Morgan Stanley expecting sales to more than double to about 28,000 units this year. The article highlights strong state support, improving production capacity, and lower costs, but also notes that demand, real-world functionality, and training data remain major constraints. Chinese firms like Unitree, AGIBOT, and Matrix Robotics are gaining share, while U.S. rivals still trail in shipment volumes.

Analysis

The near-term equity winner is less the robot OEMs themselves than the industrial ecosystem that can monetize experimentation: high-end component suppliers, motion-control, sensors, and manufacturing automation software. China’s advantage is not just lower unit cost; it is the ability to compress iteration cycles by co-locating factories, testing, and data collection, which should widen the gap versus U.S. startups that still depend on higher-cost outsourced hardware and slower deployment feedback loops.

The second-order risk is that headline shipment growth can mask weak economic utility. If buyers are mainly SOEs, labs, and demo-driven venues, reported orders can stay buoyant while actual productivity ROI remains unattractive, leading to a classic “deployment without adoption” phase. That matters because the market will likely re-rate on evidence of repeat orders into messy environments, not on one-off pilots; the inflection is likely months-to-years, not weeks.

For TSLA, the risk/reward is asymmetric but indirect: humanoid progress keeps option value alive on Optimus, yet China’s faster commercialization raises the bar for its eventual launch and pricing. The bigger competitive threat is not a single robot vendor but a Chinese supply-chain flywheel that could commoditize subassemblies and pressure margins across global robotics once volumes scale. Conversely, if the software/data problem proves harder than the hardware problem, current enthusiasm will fade quickly and the market could unwind over the next 1-2 quarters as investors realize most robots are still pre-commercial.

Key Decisions for Investors

• Go long a China robotics supply-chain basket vs short a U.S. humanoid pure-play basket for 3-6 months; the trade benefits if commercialization and unit economics in China continue to outpace narrative-only U.S. peers.
• Use TSLA as a volatility expression: buy medium-dated call spreads only on pullbacks, since humanoid upside is real but already embedded in long-duration optionality; keep downside limited if execution slips over the next 2 quarters.
• Add exposure to industrial automation enablers tied to logistics and factory integration on weakness; these names monetize real adoption even if humanoid unit sales disappoint, giving a better risk-adjusted path than OEMs.
• Short low-quality private-market robotics proxies via public comps if available, focusing on companies reliant on repeated funding rounds rather than recurring service revenue; the bubble risk should surface over 6-12 months if pilot conversion stays weak.
• Watch for a catalyst window around Chinese policy releases and SOE procurement data; if orders are still concentrated in entertainment/demo use cases by the next reporting cycle, fade the momentum trade aggressively.