At CES 2026, NVIDIA has officially launched its Vera Rubin platform, featuring six new AI chips designed to significantly enhance performance and efficiency. The Rubin architecture promises a remarkable 10x reduction in inference costs compared to its predecessor, Blackwell, alongside up to 50 petaflops of AI performance and innovative rack-scale manufacturing solutions. CEO Jensen Huang highlighted that the new chips are already in full production and are pivotal in shaping the future of AI infrastructure. The Vera Rubin NVL72 supercomputer is expected to deliver five times greater inference performance, marking a substantial leap for NVIDIA in the competitive landscape of AI technology.