According to TheRegister.com, Snowflake has made its PostgreSQL extensions open source under the Apache license, allowing developers and data engineers to integrate PostgreSQL directly with its lakehouse system. The extensions, called pg_lake, enable reading and writing directly to Apache Iceberg tables from PostgreSQL, eliminating the need to extract and move data. This follows Snowflake’s $250 million acquisition of PostgreSQL specialist startup Crunchy Data in June this year, which originally developed the technology. Snowflake executive vice president Christian Kleinerman said the move lets PostgreSQL become an interface for managing open lakehouse systems. The company also announced general availability of Snowflake Intelligence, an AI agent for natural language queries. Analyst Robert Kramer from Moor Insights & Strategy noted this gives PostgreSQL users a direct path to Snowflake’s lakehouse and AI capabilities without architectural disruption.
Sponsored content — provided for informational and promotional purposes.
PostgreSQL meets lakehouse
Here’s the thing about this move: it’s actually pretty clever. Most companies aren’t about to rip out PostgreSQL—it’s everywhere, it’s reliable, and developers know it inside out. So instead of trying to replace it, Snowflake‘s saying “hey, keep using what you’re comfortable with, but here’s a bridge to our world.” The pg_lake extensions basically let PostgreSQL act as the catalog for Iceberg tables, which is a neat trick. You can query raw data files, external tables, even geospatial formats—all from familiar PostgreSQL interfaces.
The acquisition play
That $250 million acquisition of Crunchy Data back in June is starting to make a lot more sense now, isn’t it? Snowflake didn’t just buy a company—they bought expertise and technology specifically tailored to the PostgreSQL ecosystem. And they’re open-sourcing it, which is interesting. They could have kept it proprietary, but making it open source lowers the barrier for adoption. It’s like they’re saying “come on in, the water’s fine” rather than building walls around their technology. The company’s blog post frames this as part of their broader enterprise AI strategy, but really it’s about meeting developers where they are.
AI and the platform wars
Now, let’s talk about the bigger picture. Snowflake’s also rolling out Snowflake Intelligence, their AI agent for natural language queries. But here’s the real challenge Kramer pointed out: buyers are struggling to differentiate between Snowflake, Databricks, and other cloud platforms. Everyone’s talking AI, everyone’s talking lakehouse—so what actually makes Snowflake different? According to the analyst, Snowflake’s positioning itself as the platform where AI works reliably for real operations, not just experimentation. That’s a subtle but important distinction. The engineering blog details how this fits into their broader integration strategy, but the message is clear: we’re the safe choice for production AI.
Incremental adoption strategy
What I find most interesting here is the incremental approach. Snowflake isn’t demanding that companies go all-in on their platform overnight. Instead, they’re enabling gradual adoption—letting PostgreSQL teams dip their toes in for high-value analytics while keeping their operational databases intact. That’s smart because let’s be honest, most organizations move slowly when it comes to data architecture changes. They want to blend operational databases with governed AI execution, as Kramer put it. So instead of a revolution, we’re looking at an evolution. And in the enterprise world, evolution usually wins over revolution every time.
