The quietly relentless winner
In 1995 PostgreSQL was an academic side project at Berkeley. MySQL had momentum, Oracle had budget, and "Postgres" was the database that nobody outside research labs used. Three decades later it's the most-loved database in every developer survey, the default offering on every major cloud, and the engine behind companies from Apple to Instagram to Notion. Nothing flashy happened — Postgres just kept being correct, kept being extensible, kept being free, and kept shipping.
Three structural reasons it won
- Correctness over shortcuts. When the SQL standard says X, Postgres does X. MySQL has decades of "actually we silently truncate" surprises; Postgres does not.
- Extensibility as a first-class API. You can add custom data types, operators, index methods, and entire query languages without forking the engine. PostGIS, TimescaleDB, pgvector — all extensions, none requiring a special build.
- Community ownership. No single company can hold the project hostage with relicensing. The same governance model has produced 30 years of yearly major releases.
The yearly release cadence
A new major version ships every September. Each release adds material features (logical replication, parallel queries, JSONB, MERGE, virtual generated columns, uuidv7, and so on) without breaking existing applications. You can usually upgrade in an afternoon — the cost of not upgrading compounds.