5 Million Humanoids by 2030 // David Reger (NEURA Robotics)

In this talk

1. 2:57

   Why target five million humanoids

   Reger points to aging populations and shrinking workforces as the need behind NEURA’s 2030 target. He cites China’s goal of having intelligent robots account for 5% of its workforce, which he estimates would mean 40 million robots, and projects 100 million fewer workers across Europe, China and Japan by 2030.

2. 4:42

   Physical AI needs new data

   Reger says humanoids depend on physical AI, but the data needed to train it is not readily available as text and image data is for language models. Robots need to generate this data through physical interactions, which he identifies as a major development hurdle.

3. 5:36

   Starting with industrial tasks

   NEURA plans to begin with industrial work, aiming to cover many common human tasks. Reger gives examples such as pressing buttons, turning switches, and moving objects into position.

4. 8:38

   NEURA’s platform and hardware approach

   Reger says NEURA’s approach combines its existing experience with robot arms and the data those robots produce, along with additional sensors intended to help bridge simulation and real-world use. He describes a layered platform that lets outside contributors build on it, and points to sensor design and high-bandwidth data streaming as differentiators.

5. 12:17

   Robots will coexist through a transition

   Reger expects humanoids to work alongside robot arms and mobile robots, with wheeled platforms useful where stability and efficiency matter. He believes robots may eventually take on all physical tasks, while people shift to other work; he compares that transition to the move from farm labor to tractors. At home, NEURA’s MiPA wheeled platform is still used mainly for testing, with more applications and partnerships needed for daily use.