Google Cloud, NVIDIA launch Vera Rubin instances

Google Cloud and NVIDIA have announced general availability of NVIDIA Vera Rubin infrastructure, giving cloud customers access to next-generation hardware built for demanding AI workloads. The instances are scheduled to become available next month, with a planned expansion into Europe later. The launch marks another step in the race to supply enough compute for training, inference and large-scale reinforcement learning. The offering, described as A5X instances, combines NVIDIA Vera Rubin NVL72 systems with ConnectX-9 networking. That integrated design is intended to help customers run large AI jobs without assembling every layer of infrastructure themselves. Instead of buying and tuning separate components, cloud users can access a platform that pairs high-performance accelerators with the networking needed to keep massive workloads moving efficiently. The timing matters because AI capacity remains one of the biggest constraints on enterprise adoption. Companies want to train larger models, serve more inference requests and run complex reinforcement-learning experiments, but those tasks require enormous compute and reliable interconnects. By making Vera Rubin available through Google Cloud, the two companies are positioning themselves for customers that need scale without building their own data-center footprint. The planned European expansion also has competitive and policy implications. Demand for AI infrastructure is growing across Europe, and cloud providers are under pressure to offer advanced capacity locally. Bringing Vera Rubin instances to the region could support local AI development, data residency requirements and sovereign AI initiatives. At the same time, the launch sharpens competition among cloud providers and chipmakers, all of whom are trying to win the next wave of AI spending. For now, the message is straightforward: Google Cloud is adding NVIDIA's newest infrastructure to its lineup, and customers will be able to test whether the combined platform delivers the performance and scale that modern AI workloads require.