AGIBOT crossed 10,000 robots on March 30. Three months ago it had 5,000.
The production ramp is the story, not the headline number. Two years to 1,000 units. One year to 5,000. Three months to 10,000. Shanghai-based AGIBOT is accelerating into a market that globally shipped fewer than 14,000 humanoid robots all of last year, and the velocity is what separates it from the pack.
Founded in February 2023 by former Huawei engineers Deng Taihua and Peng Zhihui, AGIBOT ranked second globally in humanoid robot shipments in 2025 with 5,168 units, according to Rest of World. Its closest competitor, Unitree, later disclosed 5,500 units in its March 2026 IPO prospectus — making it the top seller by the most current official filing. Tesla, Figure AI, and Agility Robotics combined shipped fewer than 500 units in 2025. AGIBOT's 10,000-unit cumulative total was achieved roughly three months after year-end 2025 — which means the factory has been running at a pace that makes the 2025 numbers look like a warm-up act.
The CTO, Peng Zhihui, was a Huawei Genius Youth recruit pulling a reported annual salary of roughly 2 million yuan — around $275,000 — before he left in December 2022 to build robots full time. His departure was announced on Bilibili. The comment section asked the same thing factory managers are asking now: can this actually work on a production line?
At Joyson Electronics, a major automotive parts supplier in China, AGIBOT's G2 robot is deployed across production lines under a procurement contract worth more than 100 million yuan (about $14 million), according to jiemian.com. Cycle times are as quick as 12.97 seconds per task with a success rate above 99 percent. More than 1,000 workstations have been identified as suitable for humanoid robot deployment at Joyson alone. AGIBOT says it has surpassed 1 billion yuan (roughly $144 million) in annual revenue. These are real numbers from a real deployment, and they matter because most humanoid robot announcements are not attached to a factory floor.
But the current output still needs context. AGIBOT's robots are operating at roughly 50 percent of a human worker's productivity, the company acknowledged. It expects that to reach 70 to 80 percent by the end of 2026. A robot that works half as fast as a person — even at a lower ongoing cost — is an incremental improvement, not an inflection point, on a line where a single worker might earn $8,000 to $12,000 a year. The economic case closes faster in high-wage markets and regions with labor shortages. In lower-cost environments, the payback calculation runs longer.
China's structural advantage in building the things is harder to argue with. China controls roughly 26 percent of the global actuator market versus approximately 5 percent for the United States, according to Visual Capitalist. Actuators — the motors that move a robot's joints — are to humanoids what chips are to AI: the limiting reagent. That market share did not come from robotics. It came from electric vehicles. Companies like BYD and CATL built out the battery, motor, and sensor supply chain at a scale that made China the manufacturing base for the entire global EV industry. Humanoid robots use many of the same components. The factory that already existed is the one AGIBOT is drawing from.
"China has a more robust hardware supply chain — much of it built up through the EV sector, from sensors to batteries — and the world's strongest manufacturing base, allowing companies to iterate far faster than Western competitors," Selina Xu, China AI policy lead at a major research firm, told TechCrunch. The iteration speed AGIBOT is demonstrating — doubling its output every few months — is a direct consequence of that supply chain.
The Chinese government is making sure the advantage compounds. Shanghai released a plan in August 2025 aimed at building a 50 billion yuan (about $7 billion) embodied AI industry cluster by 2027, with subsidies covering up to 30 percent of project investment, capped at 50 million yuan per project. That is direct capital aimed at keeping the supply chain local and the factories running.
AGIBOT is targeting a Hong Kong IPO in the third quarter of 2026, with a valuation range of $5.1 billion to $6.4 billion, the company has said. CICC and CITIC Securities are leading the offering; Morgan Stanley joined more recently. At the revenue figures AGIBOT is citing, the multiple will say a lot about what public markets think the next 18 months look like.
The global market is expected to grow significantly: Omdia projects humanoid robot shipments will reach 2.6 million units annually by 2035, up from roughly 13,000 last year. That is a doubling-every-year trajectory, and AGIBOT is betting its manufacturing velocity puts it at the front of it.
What to watch next is simple: the 10,000 robots still have to show up at customer sites and hit 70 to 80 percent of human efficiency by year-end. If they do, the IPO story writes itself. If they don't, the production ramp will be the headline and the deployment reality will be the footnote. Right now the company is asking buyers to take both on faith — and the early factory data from Joyson suggests the faith is earned, but not yet proven at the scale the numbers imply. China built the supply chain faster than anyone expected. The question is whether the robots that run on it can close the gap with the humans they are meant to replace.
