China's Humanoid Robot Dominance and Unitree's 'ChatGPT Moment' Debate

China's humanoid robot makers captured more than 97% of worldwide shipments in the first half of 2026, according to market researcher Smart Analytics Global, with total global volume reaching roughly 19,100 units, up 272% year over year ^[1]. But the headline dominance number hides a reshuffle happening inside China itself: Shanghai-based AgiBot overtook Unitree for the first time to become the world's largest humanoid vendor, shipping roughly 8,400 units (44% share, up 562% year over year) versus Unitree's approximately 5,900 units (31% share) ^[2]. The two companies aren't even competing the same way - AgiBot spread its bets across a full-size bipedal A-series, a compact X-series, and wheeled G-series robots, while Unitree's volume stayed concentrated almost entirely in its flagship G1 ^[2]. Forecasters expect the China-versus-rest-of-world gap to keep widening: TrendForce projects up to 94% output growth for China's sector in 2026, with Unitree and AgiBot alone accounting for nearly 80% of total shipments ^[3]. The story here isn't simply 'China wins' - it's that inside China's own runaway lead, the market leader just changed hands in six months.

The person best positioned to know how close humanoid robots are to a breakthrough is Unitree founder Wang Xingxing - and his own public statements pull in two directions at once. He has said the industry's 'ChatGPT moment' could arrive in as little as 2 to 3 years under optimistic conditions, or take 5 to 10 years if progress is slower ^[4]. He defines that moment precisely: a robot placed in an unfamiliar home or environment that can complete roughly 80% of tasks using only voice or text commands, with no retraining ^[4]. That bar matters because, by his own account, Unitree's robots aren't close to clearing it yet - they must be retrained from scratch for every new task and remain less efficient than human workers ^[4]. He also named the specific technical wall: not locomotion or gross movement, but tactile feedback during manipulation. 'When a robot reaches for an object, it looks like it has almost got hold of it - but that last bit of tactile feedback, that final small margin of error, it cannot correct,' he said ^[5]. A founder saying 'years away' and describing an 'important tipping point' in the same breath is less contradiction than candor about a genuinely open technical problem - but it sits awkwardly next to a stock chart that just went vertical.

Nothing captures the gap between spectacle and utility better than the World Robot Conference itself. Chinese firms showcased roughly 3,000 products in Beijing over five days, and Unitree's robots boxed, danced, and played ping pong against a human opponent ^[6]. Forbes contributor John Koetsier argues that the shipment numbers driving headlines reflect strategy and subsidized scale, not necessarily an engineering lead - framing Western firms' slower rollout as deliberate caution rather than falling behind: 'The gap in the sales chart isn't a gap in engineering... they're holding back until the product is genuinely reliable' ^[7]. That skepticism was borne out on the same show floor: multiple robots publicly malfunctioned on opening day, and one robot failed to fold a shirt after several minutes of repeated attempts ^[6][8]. Put together, the picture is a market that is unambiguously scaling in volume - and, per Smart Analytics Global, shifting toward genuine industrial and commercial deployment rather than demos, with that segment rising from about 50% to over 70% of shipments year over year ^[7]- while the hardest last-mile manipulation problems remain visibly unsolved in public.

Two forces are converging on Unitree at once: capital markets and geopolitics. Unitree's Shanghai STAR Market debut on August 19, 2026 sent its stock up as much as 460-629% intraday, pushing its peak valuation to roughly $66 billion on an IPO that raised about $900 million at a $9 billion starting valuation ^[4]. The euphoria didn't hold: shares fell roughly 19-20% the next trading day and kept slipping through the end of the week ^[9]. That volatility landed just weeks after the U.S. FCC banned imports of new foreign-made humanoid and quadruped robots effective July 29, 2026, explicitly on national-security grounds and explicitly aimed at a Chinese supply base that the ban's own reasoning pegs at roughly 85% of the global market ^[10]. So in the same month, Unitree's home stock market handed it a near-record single-day rally while its largest potential export market closed its door - a split-screen that says more about how differently Beijing and Washington are pricing this technology than any single shipment statistic.