4 hrs ago
Chinese Startup Targets GPT-3.0 Milestone for Humanoid Robot Brains
Spirit AI is a Chinese company trying to build better “brains” for humanoid robots.
It wants robots to understand ordinary spoken instructions and then perform several physical actions.
The company predicts this could happen by the middle of 2027.
Workers wear sensors and repeat everyday activities so the robots can learn from real human movements.
The robots can already perform some simple tasks in carefully arranged rooms.
They are being tested on production lines and in retail settings.
Doing complicated movements, such as unscrewing a bottle cap, is still difficult.
Using robots safely in busy homes and public places will require more progress.
Spirit AI expects humanoid robots to reach a “GPT-3.0 milestone” by mid-2027, enabling natural-language instructions and multi-step physical actions.
The Beijing-based startup says its robots already achieve a 90% success rate on simple tasks in structured living-room environments.
Spirit AI employs about 1,000 contractors to collect real-world movement data and has deployed Moz1 robots at CATL and JD.com.
The company has raised more than $670 million since its 2024 founding and is valued at 20 billion yuan, or about $2.9 billion.
Spirit AI says industrial and commercial uses may expand within two years, while home use remains more difficult because of fine-motor and safety challenges.
- Who
- Spirit AI, led in this article by co-founder and chief scientist Gao Yang, is developing the robot-brain technology.
- What
- The company predicts a software breakthrough allowing humanoid robots to follow natural-language instructions and perform sequences of physical actions.
- Where
- Spirit AI is headquartered in Beijing, China, with robots also deployed at facilities operated by CATL and JD.com.
- When
- Gao said the GPT-3.0 milestone could arrive by mid-2027; the Reuters visit was described as taking place on September 17, 2026.
- Why
- The company is trying to improve embodied AI so robots can perform more economically useful tasks beyond isolated movements.
Real-World Training
Simulation-Based Training
How robots should learn
Real-World Training
Spirit AI relies mainly on real-world data, arguing that physical interactions are necessary for learning tasks involving objects such as flexible electric cables.
Simulation-Based Training
Many competitors use virtual simulators to reduce model-training costs; the article does not identify a specific competitor or provide a direct response to Spirit AI’s criticism.
Safety outlook
Real-World Training
Gao said robot software is currently too immature for rogue-AI risks to be the most pressing concern, while emphasizing emergency braking and whole-body force control.
Simulation-Based Training
The article reports that AI companies in the United States have called for slowing advanced-model development after hacking incidents, and says safety will become more important as robots enter crowded commercial and home environments.
Key facts
- Company
- Spirit AI
- Headquarters
- Beijing, China
- Projected milestone
- GPT-3.0 capability by mid-2027
- Current task success
- 90% for simple tasks in structured living-room environments
- Data workforce
- Around 1,000 contractors nationwide
- Funding
- More than $670 million raised since the company was founded in 2024
- Valuation
- 20 billion yuan, approximately $2.9 billion
- Current deployments
- Tens of Moz1 wheeled humanoid robots at CATL and JD.com production lines
Quotes
Gao Yang
Co-founder and chief scientist of Spirit AI, and an assistant professor of robotics at Tsinghua University
“The next one to two years mark the initial window for industrial applications. Two years from now, we'll see robots deployed in commercial service settings doing simpler tasks. Entering homes is far harder than both.”
telegraphindia.com
“We anticipate reaching the GPT-3.0 milestone by mid-2027. You will be able to speak to a robot in natural language, and it will execute a series of reasonable physical actions to attempt the task.”
telegraphindia.com
Sources
Related news
China’s Humanoid Robot Boom Races Ahead of Real-World Readiness
China Pushes Humanoid Robots From Spectacle Toward Industrial Workforce
Unitree CEO Predicts ChatGPT Moment for Humanoid Robots
Unitree’s Humanoid Robot Boom Faces Long Road to Usefulness
Chinese Humanoid Robot Takes Center Stage at Ujjain Shop Opening
