Supabase Unifies Analytics and Vector Storage

Supabase is moving from being a place to store app data to being a place to run adjacent data workloads without leaving the product. Analytics Buckets let teams replicate transactional Postgres data into Iceberg tables on S3 for dashboarding and large scans, while Vector Buckets give teams cheaper long term storage for embeddings with search attached. That turns storage into a higher value control point inside the same dashboard developers already use for database, auth, and logs.

• Analytics Buckets are built for a different job than the core Postgres database. Postgres is for app reads and writes. Iceberg on S3 is for wide table scans, historical analysis, and feeding BI tools. Supabase is packaging that split so developers do not have to stitch together a separate warehouse stack themselves.
  1 sacra 3 supabase 4 amazon
• Vector Buckets extend the same idea to AI. Instead of keeping every embedding in hot Postgres storage, teams can push older vectors into colder object storage and still query them. That matters for RAG and agent workloads where embedding volume grows fast, but only a slice needs low latency access.
  1 sacra 3 supabase
• This also sharpens Supabase’s position versus database peers like Neon. Neon is strongest around serverless Postgres, autoscaling, and branching. Supabase is broadening the product surface around that core, bundling database, file storage, observability integrations like Sentry log drains, analytics storage, and AI storage into one operating console.
  1 sacra 2 github 5 neon 6 neon

The next step is a fuller developer data stack inside Supabase, where app transactions, logs, analytics tables, and vector data all land in one system and can be wired together from the dashboard. If that product holds together, Supabase captures more of the spend that would otherwise leak to separate warehouses, vector databases, and observability tools.