Nvidia is offering software and hardware to make humanoid robots safer around humans by letting them make split-second decisions rather than stopping at every potential contact.

What happened

Nvidia announced Halos, an operating system based on self-driving vehicle technology, that will run on its IGX Thor hardware to give humanoid robots better awareness of their surroundings. The software will allow robots to analyze what is happening around them and make their own decisions, rather than relying solely on sensors that force them to stop when a human is nearby. Nvidia is providing this technology for robots like Agility Robotics' Digit.

Why it matters

Current safety systems require robots to stop or slow down whenever there is a chance of contact with a human, which makes them less productive and prevents cooperative tasks like handing items or helping lift objects. Humanoid robotics is expected to generate $200 billion（約32兆円） in revenue by 2035 according to Barclays estimates, so solving the safety-versus-productivity trade-off is key to unlocking the market. Nvidia sees robotics as the next major AI opportunity alongside its existing data-center business.

What to watch

Nvidia has created a lab where robot makers and customers can run safety tests before approaching regulators, suggesting the company expects to move quickly from development to real-world deployment. The company is gradually expanding beyond chips for AI data centers into other fields, often offering technology for free to broaden AI adoption.