NVIDIA's Vera Rubin Shipment Accelerates Amid Rising Profit Margins
NVIDIA's Vera Rubin shipment appears to be accelerating, according to OpenAI, which has deployed its first batch of NVIDIA Vera Rubin racks. The deployment supports pre-training for OpenAI's next-generation frontier AI.
The racks are part of a joint effort by both parties to build accelerated computing infrastructure that advances the development of frontier AI. CoreWeave executives confirmed that profit margins on recently signed contracts have exceeded previous levels by 5-10%. The Vera Rubin SKU has been expanding its profit margins since its launch, with prices for the previous generation SKU remaining flat or higher than several years ago.
NVIDIA's Vera Rubin racks are being used in data centers of partners including CoreWeave, Google Cloud, Microsoft Intelligent Cloud, and Oracle Cloud Infrastructure. The platform covers 30 countries globally and over 350 factory nodes, forming the largest rack-level AI infrastructure supply chain system currently in existence.