Thomas English
AgiBot, a Chinese robotics firm, is beginning to embed homegrown AI systems like DeepSeek into humanoid robots at scale — touting a machine that can fold T-shirts today and, Beijing hopes, help rebuild China’s factory floor tomorrow, Reuters reported Monday.
Beijing is pouring billions into firms like AgiBot to marry its low-cost hardware supply chain with domestic AI, betting that legions of general-purpose bots can shore up an aging workforce and blunt U.S. tariff pressure, according to the outlet. That effort sets the stage for a direct collision with Tesla’s Optimus project, which aims to deploy thousands of humanoid robots in American factories by year’s end.
“Just imagine that one day in our own robot factory, our robots are assembling themselves,” Yao Maoqing, a partner at AgiBot, told Reuters. “The robots in five or 10 years could organize a resident’s room, pick up a package or even transfer people from a bed to a washroom.”
Inside AgiBot’s suburban Shanghai warehouse, around 100 robots run 17-hour shifts repeatedly folding shirts, assembling sandwiches and opening doors, according to the outlet. Meanwhile, some 200 human operators collect data and footage of the mechanical workers and feed it into an embodied AI model, allowing the machines to learn new tasks in hours instead of months — an approach executives reportedly say DeepSeek’s reasoning engine accelerates by chewing through terabytes of motion data.
That learning loop is greased by state cash: more than $20 billion in combined subsidies already allocated, a $137 billion innovation fund on the way, free warehouse space from city governments. Ming Hsun Lee, head of Greater China automotive and industrial research at Bank of America Securities, estimated the materials cost of a Chinese humanoid at about $35,000 — a figure expected to fall to $17,000 by 2030, versus $50,000-$60,000 for an Optimus built outside China, according to Morgan Stanley estimates.
Tesla CEO Elon Musk, however, suggested the advantage of his Optimus project is real-world deployment at speed and scale.
“This year we hopefully will be able to make about 5,000 Optimus robots,” Musk told investors at an all-hands meeting in April. “We’re aiming for enough parts to make 10,000, maybe 12,000. But since it’s a totally new product with totally new — like, everything is totally new — I’ll say we’re succeeding if we get to half of the 10,000. But even 5,000 robots — that’s the size of a Roman legion, FYI.”
Musk then outlined his plan to scale Optimus’ manufacturing exponentially, beginning production with “at least one legion” in 2025 and expanding to “probably 10 legions” in 2026.
“In a very short period of time, Optimus has gone from being an idea to the most sophisticated humanoid robot on Earth,” he continued. “There’s nothing even close to the level of sophistication of Optimus.”
DeepSeek’s open-source R-series models, released under permissive licenses and tuned for vision and language control, jolted Silicon Valley in January by matching Western performance at a fraction of training costs — a leap that lets AgiBot upload fresh factory footage each night and retrain its policy network without tearing through Nvidia clusters China can’t easily import.
China is also approaching the frontier technology with showmanship. Twenty-one biped bots lined up for Beijing’s first robot half-marathon in April; only six finished, some face-planting as batteries died, Reuters separately reported.
China’s Ministry of Industry and Information Technology called humanoid robotics “an important new engine of economic growth” in 2023, according to the South China Morning Post, setting targets for scaled deployment in logistics, healthcare and domestic services by decade’s end.
Neither AgiBot, DeepSeek nor Tesla immediately responded to the Daily Caller News Foundation’s requests for comment.
All content created by the Daily Caller News Foundation, an independent and nonpartisan newswire service, is available without charge to any legitimate news publisher that can provide a large audience. All republished articles must include our logo, our reporter’s byline and their DCNF affiliation. For any questions about our guidelines or partnering with us, please contact licensing@dailycallernewsfoundation.org.
