Google launches advanced AI models for complete humanoid robot control

Google released Gemini Robotics 2, Gemini Robotics ER 2, and On-Device 2, vision-language-action tools that enable humanoid robots to reason and act with full-body control. The development marks a significant advancement in enabling robots to perform a wide range of tasks instead of single repetitive movements. This initiative aligns with a broader trend toward generalist models that allow robots to learn and execute multiple functions. The launch follows the FCC's decision to ban advanced foreign-made humanoid robots, particularly from China. Experts point to a symbiotic relationship between AI and robotics: AI models generate synthetic data to train robots, creating a feedback loop that could accelerate mass adoption of these technologies.