Weka partners with Andromeda to standardize AI data storage

AI data‑infrastructure company Weka and high‑performance computing platform Andromeda announced a strategic partnership. Weka’s NeuralMesh platform will serve as the core AI data‑storage layer for Andromeda’s managed GPU clusters, available through dedicated NeuralMesh deployments or the GPU‑native NeuralMesh Axon method.

The integration lets customers run workloads across more than 50 global compute providers with consistent storage performance, rapid deployment (as fast as 10 minutes) and reduced storage costs. Reported throughput exceeds 400 GB per second per cluster and more than 6 million IOPS, enabling a biotech client to transfer metadata for a billion‑file catalog in minutes. Andromeda CEO Will Mushe said the goal is to make compute globally fluid, ensuring idle GPUs are fully utilized.

The partnership also adds fault‑isolation capabilities, protecting workloads from cluster‑wide failures while maintaining high availability.