Block Volumes

What it does

Attach a block volume to any compute pod and it behaves like a regular disk: mount it, write files, persist data across pod restarts. Each volume is replicated across peers so a single node failure doesn't lose data. Useful for stateful workloads like databases, build caches.

Block Volumes are scoped to a region of the peer network so latency stays low. For most workloads that just need a writable disk that doesn't disappear, it's exactly right, at a meaningful discount versus managed block storage from hyperscalers.