Nvidia seeks to make humanoid AI robots safer around humans - The Japan Times

NVIDIA is working to enhance the safety of humanoid robots to enable safer collaboration in human environments. The company argues that humanoid robots must be capable of making split-second decisions... Read original →

NVIDIA’s migration of autonomous driving safety logic to humanoid robots is a strategic expansion of its AI chip ecosystem from automotive to industrial domains. This move pressures upstream 3nm EUV capacity toward the IGX Thor platform and accelerates standardization of sensor-fused ROS architectures. Regulatory-wise, with the EU and U.S. fast-tracking AI robot safety certifications, NVIDIA’s in-house lab aims to dominate standard-setting and lock in chip-centric partnerships. Competitors like Tesla Optimus and Figure AI may be forced to open their perception stacks prematurely. TSMC in Taiwan, China, emerges as the critical foundry enabler. Over the next 18 months, warehouses will serve as validation grounds, but mass adoption hinges on whether healthcare and retail can achieve localized, safety-certified data loops—determining if humanoids evolve from costly automation into universal productivity tools.