Google Cloud Plans Deployment of Up to 960k NVIDIA Vera Rubin GPUs
Google Cloud announced plans to expand its infrastructure based on the NVIDIA Vera Rubin platform across multiple sites, deploying up to 960,000 Rubin GPUs. The A5X instance, built on NVIDIA Vera Rubin NVL72 rack-scale systems, is now live, with a version featuring ConnectX-9 SuperNIC and Virgo networking launching next month. Additionally, Google Cloud will offer rack-scale confidential computing services on bare-metal instances.
Full Report
Google Cloud Plans to Deploy Up to 960,000 GPUs on NVIDIA's Vera Rubin Platform
On October 22, it was reported that Google Cloud announced on October 21 (local time) its plan to expand infrastructure based on the NVIDIA Vera Rubin platform across multiple compute sites, with capacity to deploy up to 960,000 Rubin GPUs.
Google Cloud also stated on the same day that its A5X instances, built on the NVIDIA Vera Rubin NVL72 rack-scale systems, are now generally available. These instances feature NVIDIA ConnectX-9 SuperNICs, Google Cloud’s Virgo networking, and fourth-generation advanced liquid cooling technology, and will be officially launched next month.
In other developments, Google Cloud will become one of the first cloud service providers to offer rack-scale NVIDIA confidential computing services on bare-metal instances. Additionally, Google Cloud will launch confidential computing services on G4 virtual machines equipped with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs.