Chinese startup DexForce announced a novel approach to training humanoid robots by leveraging large-scale virtual datasets instead of physical trials. The method generates synthetic sensor inputs and environmental scenarios, allowing robots to learn complex tasks in a simulated environment before deployment. This reduces the need for expensive hardware iterations and accelerates development cycles.

Industry analysts say the technique could reshape manufacturing and logistics by making robot adoption more affordable for mid‑size firms. DexForce's investors see the cost savings as a competitive edge against rivals still reliant on costly physical testing, potentially spurring broader AI‑driven automation across sectors.

This story was written by a language model from reporting by The Economic Times. Check the original before quoting it.