Most AI stacks depend on a relational database for application state, workflow metadata, and experiment tracking. Until now, teams on CoreWeave either ran PostgreSQL themselves or kept it in another cloud. Today, CoreWeave Database is generally available: managed PostgreSQL that runs inside your CoreWeave Kubernetes Service (CKS) cluster, next to your training and inference workloads.
Bring your existing PostgreSQL databases to CoreWeave
CoreWeave Database is a PostgreSQL-compatible database service. Each database runs on CPU nodes in your cluster, with a primary instance and read replicas for high availability, and stores live data on CoreWeave Distributed File Storage (DFS). Because CoreWeave Database is fully compatible with PostgreSQL, you can migrate an existing database from another cloud or a self-managed deployment using native Postgres tools.
“Migrating our biomedical literature database from another cloud to CoreWeave Database was straightforward, and the CoreWeave team helped us get to production quickly. Our managed PostgreSQL now runs alongside the applications that depend on it, and we’re confident our data is encrypted, private, and backed up.” Mike Mattacola, CEO, Advaita Bio
Keep your data next to your GPUs
Running CoreWeave Database on CPU nodes in the same CKS cluster as your training and inference workloads gives you:
- No cross-cloud hop: Applications connect through an in-cluster Kubernetes Service, with no link to another provider's database, no egress path to manage, and no second cloud to secure.
- One platform to operate: Manage the database with the same tools as the rest of your cluster: the CoreWeave Cloud Console or kubectl and GitOps, the same monitoring, and the same security boundary.
- Fewer round trips for chatty workloads: Agents reading and writing state between tool calls, training jobs logging experiment metadata, and RAG services querying embeddings with the included pgvector extension all hit the database on every step. In the cluster, none of those calls cross to another cloud.
Spend your time on improving AI models, not database operations
Running PostgreSQL in production means owning replication, failover, patching, and backups. With CoreWeave Database, we offer this as a managed service, so you can focus on improving your AI models and applications. Your team owns your data, schemas, queries, and database users.
Data protection is built in. Every database is continuously backed up by default, with point-in-time recovery to any moment within the retention window. You can schedule your backups and take an on-demand backup before a schema change or upgrade, and restore a full database to a new cluster to verify its validity.
Run production workloads on a highly available, secure database
Every database comes with high availability and security controls in place:
- Automatic failover: Each database runs a primary and replicas spread across different hosts. If the primary goes down, a replica takes over automatically.
- Private to your cluster: Applications connect through in-cluster endpoints, and you decide whether to expose the database beyond that.
- Encryption and patching handled: Data is encrypted at rest and in transit, and we manage TLS certificates and apply security patches.
- Backups without credentials: Backups stream to CoreWeave-managed storage, with no credentials for you to handle.
- Access on your terms: Control who connects with native PostgreSQL roles and your own Kubernetes network policies.
- Audit logging: Audit logs give you a record of database activity.
CoreWeave is committed to CloudNativePG and the open-source PostgreSQL ecosystem
CoreWeave Database leverages CloudNativePG (CNPG), the open-source Kubernetes operator for PostgreSQL, which handles orchestration including primary and replica management and failover. We add the managed service around it: Console and Kubernetes workflows, CoreWeave DFS, managed backups, monitoring, and ongoing operations.
We also contribute to the PostgreSQL ecosystem that CoreWeave Database depends on. Recent work includes S3 compatibility improvements in Barman, the open-source backup and recovery tool for PostgreSQL, and connection configuration improvements in pgexporter, a Prometheus exporter for PostgreSQL metrics.
Try CoreWeave Database today
Create your first database with the CoreWeave Database documentation.
