Technology 4 min read By Arthur Ellington
AGIBOT Co-Founder Predicts Humanoid Robots Will Reach GPT-3.5 Moment Within Five Years
Yao Maoqing, co-founder of Chinese robotics firm AGIBOT, says embodied AI will achieve a breakthrough comparable to ChatGPT's GPT-3.5 within three to five years, as the company leads global humanoid shipments and eyes a Hong Kong IPO.
Humanoid robots will achieve their equivalent of the GPT-3.5 breakthrough within three to five years, according to Yao Maoqing, co-founder of Chinese robotics firm AGIBOT. Speaking at the Fortune Leaders Forum in Macau on 8 September, Yao said embodied AI is following a similar trajectory to large language models and will soon reach a level where robots can perform everyday tasks with 80 to 90 per cent reliability.
Yao's reference to GPT-3.5 points to the model that powered ChatGPT, which marked the first time an AI service could handle common, general-purpose tasks rather than only specialised ones. «As models keep maturing, embodied AI will reach GPT-3.5,» he said. «People differ somewhat on timing, but overall it's within the 3-to-5-year window.»
AGIBOT has emerged as a leading player in China's rapidly expanding robotics sector. According to data from Counterpoint Research released in late August, the company shipped 9,700 humanoid units during the first six months of the year, making it the top seller globally. The startup is now considering an initial public offering in Hong Kong.
Despite the optimism, Yao acknowledged that building robots capable of operating reliably in industrial settings remains a significant challenge. «It's genuinely difficult for today's humanoid firms to develop a robot that can run stably inside a factory,» he said. «Industry players are brutally honest: they care about four metrics — success rate, cycle time, stability and cost — rather than what technology is used. From that standpoint, we absolutely have to train our robots to 100 per cent efficacy first.»
In late June, AGIBOT ran a six-day livestream demonstrating its robots completing more than 64,000 manufacturing tasks with a 99.99 per cent success rate. Reaching that milestone required eight-hour sessions in the middle of the night for a month, Yao said on stage.
While robot dance performances and boxing matches attract headlines, manufacturers are urgently searching for real-world applications. Shanghai-headquartered Keenon Robotics has rolled out robots aimed at automating hotel services. Wan Bin, Keenon's chief operating officer, noted that one Shanghai hotel is using five robots to greet guests, deliver room service, clean rooms and manage the restaurant floor. «This scene is one that will become increasingly common in the future,» he said.
Other manufacturers are targeting industrial applications. «We've invested very heavily in developing robots for the industrial sector,» said Jianxin Pang, UBTech's vice president and vice dean of research. «But first, we're starting with relatively common scenarios with a big enough market — that helps us ensure return on our investment.»
Yao remains confident that embodied AI will follow the same scaling law that transformed large language models. «Most industry teams believe that embodied AI will follow the exponential scaling law we discovered in digital intelligence and large language models,» he said. «As data volume rises and your model's parameter count grows, there will definitely be a step-up in intelligence.»
The comments come as China's robotics startups race to commercialise humanoid technology, balancing ambitious timelines against the practical demands of industrial customers who prioritise reliability and cost over novelty. AGIBOT's planned Hong Kong listing would provide fresh capital as the sector intensifies competition.
