China's humanoid robot market is growing fast enough to embarrass recent forecasts.
Morgan Stanley has revised its China humanoid robot shipment estimate upward for the second time, landing at 50,000 units for 2026. The bank's reasoning is straightforward: the machines are no longer performing for audiences at trade shows. They are showing up in factories, shops, and restaurants doing actual work. That is a meaningful change in kind, not just in scale.
The repeated doubling matters because it signals that early deployment data is outrunning analyst models. When a major bank revises the same number twice in a short window, it usually means the original forecast was built on demo-era assumptions that real-world adoption has already made obsolete. China's combination of manufacturing scale, state-backed investment, and a willingness to deploy unproven hardware in live environments gives it a structural edge in moving robots from prototype to production floor faster than most Western rivals.
Fifty thousand units is still a small number relative to China's industrial workforce, and Morgan Stanley's enthusiasm for the sector is worth treating with some skepticism — banks have been wrong about robotics timelines before, and demand from restaurants is not the same as demand from automotive lines. But if the trajectory holds, the more interesting question is not whether humanoid robots work, but which Chinese manufacturers end up owning the category before anyone else gets to scale.