Postgres Rewritten in Rust — pgrust Now Passes 100% of Official Regression Tests
pgrust — a from-scratch Postgres rewrite in Rust — passes all 46,066 official regression tests. Here's what this milestone means for the future of databases.
What Is pgrust?
In July 2026, the pgrust project — a from-scratch rewrite of PostgreSQL in Rust — achieved an unprecedented milestone: passing 100% of the official Postgres regression tests. All 46,066 tests, from basic SQL queries to complex transactional workloads.
The achievement is the result of years of work by Mauricio "Mali" Spera, who positions this milestone not as an endpoint but as the key to changing Postgres from the inside: "Keep Postgres' behavior, leave its tests as the oracle, and use Rust with AI-assisted programming to explore deeper changes."
Why Rust for Databases?
Rust is the ideal language for rewriting databases. Its safety features — ownership, borrowing, lifetimes — guarantee no memory errors or data races. For a system as complex as Postgres (30+ years of development), this is transformative. The pgrust author tested against Postgres' full test suite — 46,066 tests — and every single one passed. The significance is that pgrust isn't a fork — it's behaviorally identical to Postgres while benefiting from Rust's guarantees.
Performance Benchmarks
The initial benchmarks are promising: pgrust shows approximately 50% faster OLTP (online transaction processing) performance and roughly 300x faster OLAP (analytical) queries in certain workloads. These gains come from memory safety eliminating entire categories of runtime checks that Postgres C code requires, and from Rust's more aggressive compiler optimizations.
What This Means for Georgian Developers and SiTech
For Georgian developers and technology companies using Supabase/Postgres — including us at SiTech, where Postgres is central to our stack — this development is particularly exciting. If pgrust matures, it could mean better performance and deeper Rust integration while maintaining full Postgres compatibility.
This is especially beneficial for data-intensive applications, real-time analytics, and AI/ML projects. The pgrust project reminds us that database innovation is far from over — and that today's critical infrastructure can be rebuilt better with tomorrow's tools and languages.