In a bold move shaking up the robotics industry, Shanghai-based start-up Agibot, founded by former Huawei Technologies “Genius Youth” recruits, is gearing up to manufacture up to 5,000 humanoid robots this year. According to a senior executive, this ambitious target rivals Elon Musk’s Optimus production plans, signalling fierce competition in the race for robotic innovation.
According to Yao Maoqing, a partner at Agibot and president of its embodied intelligence unit, this ambitious growth reflects China’s rapidly advancing robotics sector. The industry has seen remarkable expansion, with industrial robot production surging 27% year-over-year to 91,088 units in the first two months of the year, while service robots witnessed a 36% increase, reaching 1.5 million units, as per government data.
Scaling Up for Market Dominance
Established in 2023, Agibot has swiftly become a vital participant in the burgeoning robotics ecosystem of Shanghai. At present, the company uses a facility located in the Lingang New Area, which is preparing to add another manufacturing plant in Pudong producing more than 400 robots in a month.
Yao told the South China Morning Post.
“This year, we aim to deploy new products in industrial scenarios, replacing humans in specific tasks to create tangible customer value,”
While Abbot’s focus remains on industrial applications, Yao predicts it will take another five years before humanoid robots become a common feature in households.
China’s Competitive Edge in Robotics
Agibot was co-founded by Peng Zhihui, a past prodigy of Huawei’s “Genius Youth” program nurturing the country’s best technology talents. Together with the likes of Wang Xingxing from Unitree Robotics, Peng is pushing China’s advancement in the global robotics race.
Yao, who previously worked with Google’s Waymo and electric vehicle maker Nio, believes China has a natural advantage in humanoid robotics. The country’s well-established hardware supply chain and strong artificial intelligence talent pool provide an edge over global competitors.
Yao said,
“The humanoid robotics industry is just getting started, with many applications yet to be developed, and production costs are still high,”
However, he expects significant cost reductions as production scales up, making essential components like motors, reducers, and modules more affordable through mass manufacturing and improved efficiency.
He estimates that humanoid robots will become mainstream once production costs drop to 50,000 yuan (US$6,900) per unit.
A Diverse Product Lineup
Agibot operates three core product lines:
- Yuanzheng: A bipedal humanoid robot designed for commercial use.
- Genie: A dual-armed robot on wheels.
- Lingxi: A compact humanoid robot built for developers and consumers.
In January, the company reached a major milestone, surpassing 1,000 robots produced, including 269 wheeled units and 731 bipedal models. Peng currently leads the development of Lingxi, which recently launched its latest X2 model.
Strategic Investments and Valuation
Agibot has completed at least eight funding rounds, attracting investments from top venture capital firms such as GL Ventures, Lanchi Ventures, HongShan Capital Group, and CAS Star, backed by the Chinese Academy of Sciences. Industry giants like BYD and Beijing Automotive Group (BAIC) have also invested, alongside government-supported funds like Lingang’s Sci-Tech Fund, according to corporate data tracker Tianyancha.
With this strong financial backing, Agibot’s current valuation stands at approximately 10 billion yuan.
The Challenge of Data for AI-Powered Robots
One of the biggest hurdles in the robotics industry is data scarcity. Unlike large language models that can be trained using vast amounts of online text, humanoid robots require multimodal data to navigate and interact with the physical world effectively.
Agibot’s Shanghai Smart Factory showcases humanoid robots seamlessly collaborating with humans, shaping the future of digital employees in manufacturing. Where around 100 robots will operate in a 1,000-square-meter facility. These robots generate approximately 30,000-50,000 high-quality motion data records daily, with each record containing tens of thousands of tokens.
Yao said.
“To achieve general intelligence, robots require significantly more data so that we can verify scaling laws and expect emerging capabilities once data reaches billions of tokens,”
He emphasized Agibot’s commitment to open-source datasets, collaboration with industry peers and users, and using simulations to generate additional training data.
Introducing “Genie Operator-1”
Earlier this month, Agibot introduced “Genie Operator-1”, a groundbreaking foundation model designed to enhance robots’ multitasking abilities, making them more efficient and adaptable in various environments.
Shanghai, China March 10, 2025, AgiBot has unveiled Genie Operator-1 (GO-1), a groundbreaking AI model designed to improve how robots see, understand, and interact with the world.
GO-1 introduces a new Vision-Language-Latent-Action (ViLLA) framework, combining two advanced technologies: a Vision-Language Model (VLM) and a Mixture of Experts (MoE).
- VLM helps robots recognize objects and scenes using vast amounts of internet data.
- MoE has two key parts:
- Latent Planner learns from human actions and different robot types to develop a general understanding of movement.
- Action Expert is trained in over a million real-world robot demonstrations to perform fast and precise tasks.
As Agibot continues its rapid expansion, it is cementing its place as a formidable challenger in the robotics industry, proving that China’s technological ecosystem is ready to take on the world’s biggest players including Elon Musk’s Optimus.
- Latent Planner learns from human actions and different robot types to develop a general understanding of movement.
