Google Cloud and Nvidia target 960,000 Vera Rubin GPUs as A5X instances near launch
As seen on the 24/7 Wall St. homepage on October 21, 2026.
A 960,000-GPU target puts a hard number on Vera Rubin demand before the chip generation even ships broadly, with A5X instances generally available next month and Europe next in line.
GOOGLE CLOUD & NVIDIA PLAN TO SCALE VERA RUBIN INFRASTRUCTURE TO UP TO 960,000 GPUS Google Cloud announced that its new A5X instances, powered by NVIDIA Vera Rubin NVL72 systems, will become generally available next month, with plans to expand availability to Europe. The new https://t.co/iXI9T62epX
- Replies9
- Reposts17
- Likes130
Continue ReadingShow less
Google Cloud's A5X instances, built on Nvidia's Vera Rubin NVL72 systems, are set to become generally available next month, and the scale of the ambition behind them is striking: the two companies are planning infrastructure that could reach up to 960,000 GPUs.
That headline number lands before Vera Rubin has even shipped broadly, which means the demand signal is being locked in at the infrastructure layer ahead of general market availability. For investors watching the AI compute race, this is an early look at how aggressively Google Cloud intends to compete for workloads that require the latest silicon.
Europe is next in line for availability after the initial rollout, suggesting Google Cloud is moving quickly to give hyperscale customers a Vera Rubin option across major regions. The geographic expansion matters because enterprise and sovereign AI customers in Europe have been a contested market for all the major cloud providers.
A 960,000-GPU ceiling is the kind of commitment that typically requires significant supply arrangements, putting a spotlight on Nvidia's ability to deliver NVL72 systems at that scale in the near term.