Nvidia and Oracle plan multi-gigawatt Vera Rubin cloud capacity

The collaboration pairs GPU-heavy training capacity with CPU-heavy agent work, and it lets customers scale each resource on its own. Oracle Cloud Infrastructure will also offer a new eight-GPU server configuration, giving buyers more ways to deploy as models change. The announcement builds on Google Cloud's general availability of Vera Rubin infrastructure, announced October 21, so the platform is reaching major clouds in quick succession. The multi-gigawatt framing shows that power and data center space now set how fast AI capacity can come online, not chip supply alone. Oracle's expansion also extends its existing Nvidia relationship instead of starting a new one. Pricing, delivery dates and named customers were not part of the summary available for this item, so the scale should be read as a stated target rather than a contracted volume.