Nvidia is betting big on making physical AI safer, from robotaxis to humanoid robots. Their Halos system is designed to help developers implement safety measures in autonomous vehicles and other robots, aiming to prevent accidents.

The company launched Halos for Robotics in 2025, targeting warehouses, factories, and even surgical settings. CEO Jensen Huang highlighted its $10 billion annual revenue contribution in 2026.

Nvidia's full-stack safety system includes hardware like the IGX Thor computing module, dedicated to safety-related tasks. This ensures that functional and safety systems run on the same silicon, crucial for physical AI.

Nvidia has also invested in and partnered with various robotics companies, including Agility Robotics and Figure AI, to develop and enhance these safety features.

As AI becomes more integrated into our lives, ensuring safety will be key. Nvidia's multi-pronged approach could be a significant step towards safer autonomous technologies.