PostgreSQL Charges Ahead with AWS, Databricks, IBM, Microsoft, Snowflake
The industry's most-popular, most-compatible database now has the most momentum, too.
Welcome to the Cloud Database Report. I’m John Foley, a long-time tech journalist who has also worked in strategic comms at Oracle, IBM, and MongoDB. Now I’m an independent tech writer.
Oracle, move over. MongoDB, stand back. BigQuery, no big deal.
The database world now belongs to PostgreSQL.
Good ol’ PostgreSQL, with roots going back 50 years to Ingres (Postgres = Post Ingres), is charging ahead in a big way. In mid-May, Databricks announced plans to acquire Neon, a Postgres cloud database, for $1 billion. Two weeks later, Snowflake said it’s buying Crunchy Data, another Postgres startup.
So the two hottest database companies have thrown their weight behind PostgreSQL, or Postgres, or just PG — it goes by all three monikers. The common theme is that Databricks and Snowflake needed to beef up their transaction capabilities, as reported by Lindsay Clark, but there’s more to the spurt of activity.
Along with those two big moves, there have been a handful of others that add up to momentum and, potentially, a market shift.
Over the past six weeks…
May 7 - CockroachDB announced support for IBM’s s390x processor architecture, which means enterprises can run Cockroach’s Postgres-compatible database on IBM LinuxONE and Z mainframe systems. Here’s why that matters: 70% of global banking transactions run on IBM Z mainframes. Cockroach on an IBM mainframe gives new meaning to the term “computer bug.”
May 16 - IBM introduced PostgreSQL 17 on IBM Cloud, featuring optimized memory management, faster query execution and bulk operations, and lower resource consumption. It comes with a 99.99% uptime SLA. At the time of the announcement, IBM was offering a 30-day, $250 credit “while supplies last” — lol.
May 19 - Microsoft announced the preview of a new “IDE for PostgreSQL in VS code.” In plain English that means a PostgreSQL extension for Visual Studio Code, an integrated development environment, to simplify database management and development. More specifically, the PostgreSQL extension integrates Postgres database tools with the @pgsql GitHub Copilot agent, which is a coding assistant. Separately, Microsoft also introduced DiskANN — a vector indexing algorithm developed by Microsoft Research — on Azure Database for PostgreSQL. It can be used to build scalable, low-latency GenAI apps that, according to Microsoft, outperform pgvector index types.
May 20 - Yugabyte fired a shot across the bow of MongoDB with its support for the new Postgres extension called DocumentDB, an open source document database engine released by Microsoft. Yugabyte says the extension gives developers the flexibility “to replace MongoDB workloads with YugabyteDB.”
May 27 - AWS announced availability of Amazon Aurora DSQL, a serverless, distributed SQL database that’s PostgreSQL-compatible. AWS points to Aurora DSQL’s capabilities as a distributed database thats spans regions: “We’ve fundamentally reimagined distributed database architecture to enable customers to build applications with virtually unlimited scalability and zero operational overhead, while maintaining the strict consistency their businesses demand,” says AWS VP G2 Krishnamoorthy. (I like the name & title, “I’m AWS VP G2….”) AWS touts DSQL’s five nines availability across regions.
June 11 - Databricks launched Lakebase, described as a “first-of-its-kind, fully-managed Postgres database built for AI.” The system, in public preview, is focused on speed and scale for AI agents. “The new Lakebase, powered by Neon technology, brings operational data to the lakehouse (storing data in low-cost lakes) with continuous autoscaling of compute to support agent workloads and unifies operational and analytical data,” says Databricks.
June 17 - EnterpriseDB continued its world tour with the announcement of EDB Postgres AI in Singapore, ushering in a “new era of sovereign data and AI for enterprises, industries, and nations.” A few weeks earlier, EnterpriseDB was making news with NVIDIA in Paris and, before that, in Saudi Arabia with Saudi Business Machines Ltd. EnterpriseDB is looking to distinguish itself with its support for data sovereignty via a hybrid model.
PG 18 in beta
A lot is happening in this space, and there’s every reason to believe it’s going to continue. AWS, Google Cloud, Microsoft, and Oracle all have Postgres offerings.
In the weeks ahead, we will learn more about Snowflake Postgres, which the company says will be “a secure, compliant and fully managed Postgres database that empowers developers with familiar tools designed to build production-ready AI agents and apps, while also satisfying the most rigorous enterprise demands.”
And PostgreSQL 18 is in beta and due for release this fall. It will package improvements for OLTP, analytics, and data warehousing, according to an article by Jelani Harper in The New Stack. “Not only is the latest edition of the engine faster than ever before, it’s becoming increasingly applicable to high-volume deployments,” Harper writes.
A still-rising star
Postgres didn’t burst onto the scene from nowhere. Maintained by the PostgreSQL Global Development Group, it was database of the year in 2017, 2018, 2020, and 2023, according to DB-Engines. And it was the most popular database in Stack Overflow’s developer survey in 2023 and 2024.
Also, because Postgres is one of the original open source databases, there’s a far-reaching ecosystem, including more than 700 contributors. Postgres is at the top of Carnegie Mellon’s Database of Databases leaders board for Compatibility, with 41 Postgres-compatible databases and 33 databases derived from Postgres. See the lists here.
What’s happening now is that Postgres development is shifting into higher gear because many of the things it’s good at — extensibility, ACID properties, multi-modal data, concurrency, stability, reliability, and of course SQL — are more important than ever as IT teams try to get their data estates into better shape for AI.
What’s your view? Leave a comment here or reach me on LinkedIn.
