CoreWeave launches managed PostgreSQL inside AI Kubernetes clusters
CoreWeave Database is now generally available, giving teams managed PostgreSQL inside their CoreWeave Kubernetes Service clusters. The service places databases alongside AI training and inference workloads, rather than requiring teams to run PostgreSQL themselves or use another cloud.
- CoreWeave Database is generally available.
- Managed PostgreSQL runs inside CoreWeave Kubernetes Service clusters.
- Databases use primary instances, read replicas, and continuous backups.
- The service supports migration with native PostgreSQL tools.
- CoreWeave Database uses CloudNativePG and includes pgvector.
PostgreSQL designed for AI workloads
CoreWeave Database is a PostgreSQL-compatible database service. Each database runs on CPU nodes in a customer’s CoreWeave Kubernetes Service cluster, with a primary instance and read replicas for high availability. Live data is stored on CoreWeave Distributed File Storage.
Because the service is compatible with PostgreSQL, teams can migrate databases from another cloud or from self-managed deployments using native Postgres tools. CoreWeave says the service is intended for application state, workflow metadata, and experiment tracking, as well as workloads such as agents, training jobs, and retrieval-augmented generation services.
The service includes the pgvector extension. CoreWeave says this allows RAG services to query embeddings in the same cluster as the applications and workloads using them.
Managed operations, backups, and access controls
CoreWeave Database takes responsibility for production database operations including replication, failover, patching, and backups. Customers retain control of their data, schemas, queries, and database users.
Every database is continuously backed up by default, with point-in-time recovery to any moment within the retention window. Customers can schedule backups, create an on-demand backup before a schema change or upgrade, and restore a full database to a new cluster to verify its validity.
CoreWeave says databases are private to the customer’s cluster by default and connect through in-cluster endpoints. Data is encrypted at rest and in transit, while CoreWeave manages TLS certificates and security patches. Backups stream to CoreWeave-managed storage without credentials for customers to handle. Native PostgreSQL roles, Kubernetes network policies, and audit logs provide additional controls over access and database activity.
Built on CloudNativePG
CoreWeave Database uses CloudNativePG, an open-source Kubernetes operator for PostgreSQL. The operator handles orchestration, including primary and replica management and failover. CoreWeave adds the managed service around it, including Console and Kubernetes workflows, CoreWeave Distributed File Storage, managed backups, monitoring, and ongoing operations.
CoreWeave also says it contributes to the PostgreSQL ecosystem. Its recent work includes S3 compatibility improvements in Barman, an open-source PostgreSQL backup and recovery tool, and connection configuration improvements in pgexporter, a Prometheus exporter for PostgreSQL metrics.
For teams running AI applications on CoreWeave, the launch consolidates database and compute operations within the same Kubernetes environment. CoreWeave says applications can connect through an in-cluster Kubernetes Service, avoiding a link to another provider’s database, a separate egress path to manage, and a second cloud security boundary.
CoreWeave: CoreWeave Database Is Here: Managed PostgreSQL Where Your AI Runs
