Sudo AI company profile
Sudo AI develops foundation models and developer tools for robotic manipulation. Its work focuses on learning transferable robot skills and deploying them across physical hardware, with Sudo R1 targeting simulation-to-hardware robot control.
Facts
- Segment: Foundation Models
- Country: China
- Founded: 2025
- Website: https://www.sudo.ai/
Profile sections
- Summary
- Market context
- Products
- Facts
- Context / Market Position
- Events
- Team
- Record / Articles
- Sources
Summary
Sudo AI develops foundation models and developer tools for robotic manipulation. Its work focuses on learning transferable robot skills and deploying them across physical hardware, with Sudo R1 targeting simulation-to-hardware robot control.
Market context
The company is one of the clearer simulation-first robot-learning startups in China, with an early public claim of zero-real-world-data transfer for contact-rich manipulation. Its focus sits between robotics foundation models, manipulation policy learning and mobile manipulation developer infrastructure. Sudo AI develops robot manipulation models and self-developed hardware-software systems for testing and deployment.
- Capability: Simulation-trained robot policies. Object-picking foundation model development. Sim-to-real deployment on physical robot hardware. Closed-loop visual manipulation control
- Strategic position: Robotics intelligence layer focused on making manipulation skills scalable through simulation rather than real-world data collection. Positioned around developer-facing robot intelligence rather than…
Facts
- Website: https://www.sudo.ai/
- HQ: Shanghai, Shanghai, China
- Founded: 2025
- Segment: Foundation Models
Simulation-first robot manipulation company
Sudo AI develops simulation-trained robot manipulation systems, led by Sudo R1 for object picking and manipulation foundation-model research.
- Target environment: Robot manipulation labs, warehouse picking tests, embodied AI research, physical AI development and robot learning environments.
- Deployment model: Manipulation-focused model and hardware development using simulation-to-hardware training workflows.
- Customer context: Robot developers, AI labs, warehouse automation teams and researchers building manipulation systems.
- Workflow context: Object picking, simulation training, closed-loop manipulation, robot policy learning and manipulation foundation-model evaluation.
- Commercial maturity: Early-stage robotics AI company with Sudo R1 and simulation-trained manipulation work. Scaling depends on sim-to-real transfer, object diversity, hardware compatibility, speed, safety, and productio…
- Market position: Robotics AI company focused on simulation-trained manipulation rather than large real-world teleoperation datasets.
- Adoption constraints: Adoption depends on transfer reliability, object diversity, hardware compatibility, safety, speed and proof across real production workflows.
- Adjacent context: Adjacent context includes manipulation foundation models, simulation-to-real robot learning, warehouse picking and physical AI systems.
- Source confidence: high
Events
- 2026-04-20 - Capability - Sudo AI publicly reveals sudo R1 simulation-trained manipulation system - Sudo AI introduced sudo R1 as a fully integrated robot system for object picking, built with self-developed hardware an…
Record / Articles
- 2026-04-20 - Sudo AI reveals sudo R1, a manipulation system trained entirely in simulation - Sudo AI's first public system uses no real-world demonstrations; a closed-loop policy running at 15 to 25 Hz picks unseen obje…
Sudo AI canonical Korthos profile