Prompt
What PostgreSQL should I use for a new product database?
Latest observation
Oct 7, 2026
OpenAI APIWeb search: off
For a new product database, you usually want the latest stable PostgreSQL major version that your hosting/platform supports—unless you have a dependency that requires an older one.
Practical recommendation
- Use PostgreSQL 16 or 17 if available on your stack.
- If you want the safest conservative choice with broad ecosystem support, PostgreSQL 16 is a very solid default.
- Avoid starting a new product on an outdated major version unless there’s a specific compatibility reason.
How to choose
Pick based on these factors:
- Hosting support
- Your cloud provider / managed DB service may not yet support the newest major release.
- Library/framework compatibility
- Some ORMs, drivers, or extensions may lag behind.
- Extensions you need
- If you rely on PostGIS, TimescaleDB, pgvector, etc., verify version support.
- Operational stability
- Newer major versions are generally fine, but if your team values maximum maturity, one version behind the newest is often a good compromise.
Rule of thumb
- Greenfield SaaS / product app: PostgreSQL 16
- If fully supported and you want newest features: PostgreSQL 17
- If you need maximum compatibility today: PostgreSQL 15 or 16, depending on environment support
Also consider
For a product database, PostgreSQL is a great default because it gives you:
- strong ACID transactions
- flexible JSON support
- good indexing options
- reliable scaling for most early-stage products
If you want, I can also suggest a PostgreSQL setup checklist for a new product database (schema, extensions, backups, indexes, and connection pooling).