I list the features not in any specific order other than my own coolness factor.
Temporal constraints
In the age of AI-driven data, temporal constraints define rules based on time—such as before, after, during, and until—that govern overall system behavior.
With PostgreSQL 18, temporal constraints can now be used to enforce data integrity across time, maintaining foreign key relationships over defined time periods. This allows you to create primary key and foreign key constraints that are valid for specific date or timestamp ranges.
PL/Rust
Modern application users expect high performance and instant responses, shaped by the immediate feedback they experience on social media platforms. This demand for speed is one reason Rust gained significant popularity in 2025.
From blockchain systems to cloud infrastructure, Rust stands out for its memory safety and efficiency. PL/Rust is a loadable procedural language that allows you to write PostgreSQL functions in Rust. Because these functions are compiled to native machine code, they can deliver significantly better performance than most interpreted procedural languages.
Vectorchord
With the rapid rise of AI, discussions around pgvector are everywhere. However, after reading a blog post by vectorchord.ai describing how to build a 100M vector indexing in just 20 minutes, VectorChord clearly stood out. It is an extension I plan to follow and explore further.
VectorChord is a pgvector-compatible extension for similarity search, designed to deliver high performance while remaining disk-efficient.
Virtual generated columns
PostgreSQL 18 introduced virtual generated columns. In a recent discussion with a peer, we noted that generated columns have existed since PostgreSQL 12, but the key difference in version 18 is how they are stored.
With virtual generated columns, values are computed at query time and not stored on disk. This approach reduces storage overhead while still providing strong query performance by calculating the values on the fly when needed.
PostGIS
PostGIS is one of the most underrated PostgreSQL extensions. I recently had the opportunity to present at FOSS4G North America, where learning about real-world use cases and seeing another side of the community made me a true believer.
I plan to further explore and test PostGIS’s spatial capabilities, particularly its powerful support for querying geospatial and location-based data.
pg_stat_statements
Rounding out my list is an all-time favorite: pg_stat_statements. While it isn’t new, it’s a tool every PostgreSQL user should be familiar with and use regularly.
Beyond identifying slow-running queries and performance bottlenecks, pg_stat_statements should be your go-to tool for understanding overall database behavior. By reviewing execution statistics and I/O metrics over time, you can identify trends and see how your database performs during both peak and non-peak hours.
Conclusion
Narrowing this list down to just six resolutions was no easy task, as it meant leaving out many other excellent PostgreSQL extensions and features that are equally worthy of attention. This, however, is a testament to the strength of the PostgreSQL ecosystem—rich, mature, and constantly evolving to meet modern data challenges.
As you explore new capabilities, it’s important to approach them thoughtfully. Take the time to research and evaluate each feature or extension before introducing it into a production environment. Clearly define your requirements, test thoroughly, and ensure that any additions align with your performance, stability, and operational goals. A well-chosen extension can unlock significant value, but only when it fits your workload and usage patterns.
PostgreSQL continues to thrive because of its flexibility, innovation, and vibrant community. Whether your focus for 2026 is performance, observability, extensibility, or new workloads, there has never been a better time to deepen your PostgreSQL expertise. Happy New Year—and may your 2026 be productive, performant, and truly Postgres-prosperous.
