Unitree Technology (宇树科技) issued a clarification on January 22, 2026, to correct online speculation about its 2025 sales. The company said that actual shipments of its pure humanoid robots in 2025 exceeded 5,500 units — defined as devices sold and delivered to end customers — and that more than 6,500 robot bodies had rolled off its mass-production lines. Unitree stressed these figures refer only to its humanoid models and exclude other forms such as two-armed wheeled platforms.
The distinction between ‘‘actual shipments’’ and orders is important: Unitree explicitly noted that orders remain higher than delivered volumes. By separating body production from delivered, customer-ready units and by excluding non-humanoid platforms, the company sought to push back against misstated aggregate tallies circulating online that mixed different robot types and gave a misleading picture of its scale.
If independently verifiable, the numbers mark a notable moment for the robotics industry. Tens of thousands of research demonstrations have given way, in a few Chinese firms, to assembly-line output measured in the thousands — a quantitative jump that suggests Unitree is moving from prototype and pilot stages toward volume manufacturing and routine commercial deployments.
The potential buyers for these machines span retail, events, hospitality, education and light commercial services, areas where humanoid form factors can provide a more human-facing presence than wheeled or static robots. High production volumes also imply supply-chain scaling, greater component sourcing, and — crucially — pressure on software, maintenance and after-sales services to turn hardware shipments into recurring revenue.
The company’s clarification itself reveals two forces at work: intense market curiosity about robotics milestones, and the reputational risk of imprecise reporting. Unitree’s note that it had never previously disclosed 2025 sales suggests the figures were released strategically to correct the record and to signal capability to customers and investors without inviting exaggerated headline numbers based on mixed product counts.
There are important caveats. The claims have not been independently audited in public filings, and shipped units do not alone indicate commercial success if those units are heavily discounted, returned, or deployed in limited, highly supervised settings. Moreover, the rapid growth of physical output raises questions about reliability, safety certification, and the depth of the software ecosystem needed to make humanoids useful beyond controlled demos.
For global audiences, the development matters because it highlights China’s ability to move beyond isolated breakthroughs to industrial-scale manufacturing in robotics. That shift alters competitive dynamics — not merely who builds the most advanced prototype but who can cost-effectively manufacture, maintain and monetize large fleets of embodied robots. The coming year will test whether volume translates into sustainable business models and wider adoption.