CoreWeave Launches Managed PostgreSQL Service Co-Located with AI Workloads
CoreWeave Database is now generally available, allowing users to run managed PostgreSQL within their CKS clusters to eliminate cross-cloud latency.
US-based AI cloud computing company CoreWeave announced on November 11, 2026, that its managed database service, CoreWeave Database, is now generally available. The service provides fully PostgreSQL-compatible database instances running on CPU nodes directly within the user's CoreWeave Kubernetes Service (CKS) cluster, alongside training and inference workloads.
Previously, teams using the CoreWeave platform had to self-manage PostgreSQL or keep databases in another cloud, requiring applications to connect across cloud boundaries. The new service uses in-cluster endpoints, eliminating cross-cloud hops, egress path management, and additional security configurations. Data is stored on CoreWeave Distributed File Storage (DFS), supporting automatic failover, encryption, and point-in-time recovery.
Built on the open-source CloudNativePG Kubernetes operator, the service handles operations, backups, and security patches via CoreWeave. Users can migrate existing databases from other clouds or self-managed deployments using native Postgres tools.