CNBC's The China Connection newsletter: China ships more humanoid robots than the U.S. as investors diverge on AI bets

Chinese humanoid robot startups are shipping into factories and malls and are already taking the top six spots in Omdia's 2025 global robot-shipment rankings, while U.S. peers such as Figure are valued much higher at $39 billion. AI2 Robotics said it is valued at 20 billion yuan ($2.93 billion), and Galbot is described as the highest-valued privately held Chinese company in the sector, but cross-border capital remains constrained by geopolitics. The article also notes China's Q1 GDP grew 5%, March retail sales rose 1.7%, and exports slowed to 2.5%.

Analysis

The investable gap here is less about robotics progress than about who owns the commercialization bottleneck. China is already producing deployment-scale supply chains, so the marginal winner is likely not the “best” humanoid model but the ecosystem that can deliver actuators, reducers, sensors, batteries, and integration services at acceptable cost. That favors Chinese industrial hardware suppliers and contract manufacturers with humanoid exposure, while U.S. names that are priced as broad AI platforms will need to prove software monetization faster than physical deployments can scale.

The second-order effect is on sourcing power: if overseas buyers keep assembling robots from Shenzhen components plus Western software, value capture bifurcates. China keeps the hardware margin and learning curve, while U.S. firms retain premium software economics but lose bargaining leverage on bill-of-materials costs. That can compress gross margins for U.S. humanoid startups over the next 12-24 months if Chinese component suppliers standardize enough to become the default stack.

For TSLA, the issue is not near-term unit sales but whether Optimus remains a valuation call option or becomes a procurement reality. If China’s deployment cadence forces a benchmark on price/performance, Tesla may face a more expensive path to defend a software-plus-robot platform narrative, especially if its hardware sourcing is slower to localize. For LI, the direct read-through is limited, but its EV manufacturing base benefits from any capital reallocation toward hard-tech industrialization and from China’s broader credibility in high-volume physical manufacturing.

Contrarianly, the market may be underestimating how quickly foreign capital can reprice this segment once a few marquee factory wins are visible. The risk to chasing the China humanoid theme now is geopolitical: policy friction can block cross-border ownership even if product adoption keeps rising. The best asymmetry is to own the enablers of production scale rather than the headline model winners, because deployment data will compound faster than brand prestige.

Key Decisions for Investors

• Initiate a tactical long in TSLA via 6-9 month call spreads only on pullbacks, targeting a catalyst window around the next product/demo cycle; thesis is optionality on humanoids, but cap downside because monetization remains unproven and valuation is already rich.
• Build a basket long of China industrial automation/robotics enablers through liquid proxies where possible; preferred setup is a 3-12 month long versus short U.S. humanoid pure-plays, betting that deployment scale will matter more than platform rhetoric.
• Pair trade: long LI / short a U.S. high-multiple AI-robotics proxy on any sector-wide rerating; LI offers exposure to China manufacturing credibility without paying for a pure humanoid narrative, while the short leg hedges sentiment-driven multiple expansion.
• Avoid chasing Chinese humanoid private-market names indirectly via venture-style vehicles unless terms include supply-chain ownership; the best risk/reward is in picks-and-shovels, not late-stage valuation momentum.
• Set a 60-90 day catalyst watch on foreign orders and factory deployment announcements; if overseas procurement shifts to Chinese hardware becomes visible, add to hardware beneficiaries and fade U.S. platform multiple expansion.