Google Cloud to open Vera Rubin-based A5X instances to all customers next month
Google Cloud will open Vera Rubin-based A5X instances to all customers next month. Early tests by London advertising group WPP show that AI agent workloads that previously required multiple server clusters can run on a single A5X instance. Externally, racks are interconnected via ConnectX-9 network cards and Google's Virgo network, scaling across multiple sites to 960K Rubin GPUs; the design targets reinforcement learning, where faster CPU-to-GPU and GPU-to-GPU links can shorten each training round.
Why read this
AI score 70
The source gives the availability timing for A5X instances and early test results from WPP, offering insight into cloud providers' hardware design direction for reinforcement learning workloads.
Topics
Body
Google Cloud will open its Vera Rubin-based A5X instances to all customers next month.
In early tests by London advertising group WPP, AI agent workloads that previously needed a cluster of several servers ran on a single A5X instance.
Outside each rack, ConnectX-9 network cards and Google’s Virgo network tie racks together, letting Google Cloud scale the setup to 960K Rubin GPUs across multiple sites.
The design targets reinforcement learning, where models repeatedly generate experience, score results and retrain, so faster CPU-to-GPU and GPU-to-GPU links shorten each round.