Nvidia has placed one of its biggest bets beyond AI data centers: a full-stack safety system for physical AI technologies including robotics and self-driving vehicles. The company's Halos system, which launched in 2025 for autonomous vehicles, has since expanded to cover a broader range of machines that interact with the physical world.
Nvidia Halos for Robotics, announced in June 2026, extends the safety architecture to autonomous mobile robots in warehouses, humanoid robots operating inside factories, and even surgical robots. The system combines hardware and software tools designed to help developers implement guardrails that prevent machines from harming nearby people.
The investment reflects Nvidia's recognition that the next frontier of AI is not just running models faster but deploying them safely in environments where failures have physical consequences. As AI models grow more capable and robot hardware improves, the gap between a software bug and real-world harm narrows.
Robotics companies are already building on the Halos platform. Nvidia says the system addresses a market need: as autonomous machines move from controlled factory floors into shared human spaces, the safety requirements become far more complex. A warehouse robot that collides with a worker or a surgical robot that deviates from its plan carries risks fundamentally different from a chatbot generating incorrect text.
Industry observers see the move as Nvidia positioning itself not just as the company that makes AI chips but as the company that defines how AI interacts with the physical world. The safety layer could become as critical as the compute layer in determining which robotics platforms gain enterprise adoption.