Democratizing Humanoids Mini Pi Plus, an Affordable and Easy to Set Up Full Stack Ecosystem

In this talk

1. 1:31

   Physical AI needs shared infrastructure

   Han argues that physical AI needs a dependable foundation before robots can scale into homes, factories, or companion roles. He points to costly hardware, fragmented components, and the lack of a common software pipeline as barriers to progress.

2. 5:03

   A three-layer foundation

   Hightorque’s proposed stack combines physical components such as actuators and batteries, a standardized development platform, and applications. Han says the platform should be stable and open so developers can contribute algorithms and build use cases.

3. 7:27

   Affordable hardware and open software

   Han gives three criteria for lowering the entry barrier: modular parts, lower cost, and accessible development tools. He says the robots start at $3,500 and that the company provides open hardware and software, including ROS-based locomotion, perception, and whole-body control algorithms.

4. 12:29

   Designed for developers and deployment

   Han highlights sim-to-real transfer and says building key components in-house lets Hightorque tune production hardware for closer alignment with simulation. He also describes documentation and materials for developers, plus demonstrations of stable movement, terrain-aware locomotion, and teleoperation.

5. 16:07

   Research, education, and commercial uses

   Han says more than 100 clients worldwide use the platform, including research labs and startups, and mentions robot competitions. He describes research and education as early uses, alongside commercial collaborations such as interactive characters and film projects.