Chinese humanoid robots' biggest obstacle: humans are still (mostly) better

Humanoid robotics remains constrained by performance—robots are not yet as efficient as humans and need time to learn new skills, creating an industry bottleneck. Unitree’s IPO-day surge (+460% as referenced) was followed by a sharp -18.7% share drop, while U.S. FCC import restrictions on foreign-made advanced robotic devices (including humanoids) add a regulatory overhang. Despite the headwinds, Keenon reported shipping 100,000+ robots and expects to exceed 150,000 units by year-end next year, highlighting continued commercialization but with execution risk toward near-perfect (99.9%) task completion.

Analysis

This looks less like a breakthrough moment for humanoids and more like a rotation back to industrial reality. When end-users say they need uptime and a solved workflow, capital shifts to the vendors that can sell integration, service, and measurable labor savings today; that favors incumbents in factory automation, AMRs, and cobots over stand-alone humanoid stories that still require heavy customer education. The first-order loser is the high-multiple narrative around “general-purpose” robots; the second-order loser is any supplier ecosystem priced off a fast adoption curve that can slip by years.

The regulatory overhang matters most as a friction tax, not a volume killer. Even modest import restriction risk pushes foreign robotics players toward localization, U.S. partners, and slower customer rollout, which improves the relative position of domestic automation platforms with existing certifications and field support. Over 6-18 months, this should widen the gap between companies shipping proven systems and those selling demos; the market will likely pay up for reliability and recurring service revenue while discounting speculative unit growth.

Contrarian view: consensus may still be underestimating how hard 99.9% task completion is in messy real-world environments. If pilot programs can’t prove payback versus a cobot/AMR stack, humanoid timelines get pushed out another few quarters and the reset in expectations can be sharp. The main bullish surprise would be a documented enterprise deployment with repeatable economics; absent that, the burden of proof stays squarely on the humanoid vendors.

Key Decisions for Investors

• Buy ABB on any post-news weakness over the next 1-2 weeks; 3-6 month thesis is relative outperformance as buyers favor deployed automation over aspiration. Falsify if order growth or service margins roll over.
• Pair trade: long ABB / short 9880 HK (UBTECH) for 1-3 months. Best risk/reward if humanoid hype fades while industrial automation budgets stay intact; thesis breaks on disclosed multi-site customer wins or faster-than-expected revenue conversion.
• Use a small TSLA put spread as a sentiment hedge over the next 1-2 quarters if you think humanoid enthusiasm is embedded in the multiple. Keep size modest because auto delivery remains the larger driver; close if Optimus-related milestones become commercially credible.