Nvidia's focus on rack-scale AI systems is a portent for the year to come — Rubin points the way forward for company, as data center business booms

Nvidia's CES keynote focused on the Vera Rubin platform and the NVL72 AI supercomputer, marking a shift towards selling entire AI systems rather than individual GPUs. The Vera Rubin platform integrates GPUs, CPUs, interconnects, DPUs, and switches into rack-scale computing systems designed for AI workloads. These systems aim to significantly reduce the cost of inference and improve performance for large language models. Nvidia's emphasis on pre-integrated systems reflects the trend of large customers buying hardware in standardized blocks. The company's decision to spotlight Vera Rubin at CES signals a strategic shift towards system-level gains in the data center market over traditional GPU announcements.