Nvidia seeks to make humanoid AI Robots safer around humans - Los Angeles Times

NVIDIA is advancing technologies to enhance the safety of humanoid robots in human environments, enabling more natural collaboration between robots and people. By offering software and chip solutions ... Read original →

NVIDIA’s push into humanoid robot safety isn’t just about robotics—it’s a strategic lever to anchor its next-gen AI chip dominance in smart hardware. By integrating 3nm EUV-based IGX Thor with the Halos OS, it forces upstream sensor and edge-compute vendors to upgrade for real-time situational awareness, while downstream software must adapt to ultra-low-latency architectures. Regulatory-wise, NVIDIA’s in-house certification labs preempt fragmented EU/US safety standards, avoiding the liability quagmire that plagued L3 autonomy. Against rivals like Tesla Optimus and Figure AI, NVIDIA’s open-stack approach locks in developers via a ‘chip-toolchain-certification’ moat. Over the next 18 months, warehouses will see scaled deployment, but complex domains like healthcare will expose current AI’s poor generalization—ironically fueling demand for bespoke AI accelerators and reinforcing NVIDIA’s pricing power in high-end semiconductors.