According to TheRegister.com, Snowflake has made its PostgreSQL extensions open source under the Apache license in a bid to help developers integrate the popular database with its lakehouse system. The company acquired PostgreSQL specialist startup Crunchy Data for $250 million back in June 2024, and now they’re releasing the pg_lake extensions that allow direct reading and writing to Apache Iceberg tables from PostgreSQL. Christian Kleinerman, Snowflake’s executive vice president of product, explained this lets developers use PostgreSQL as an interface to manage an open lakehouse. The extensions eliminate the need to extract and move data between systems, making analytics data immediately available. Snowflake also announced general availability of Snowflake Intelligence, its AI agent for natural language queries, alongside additions to its Horizon data catalog.
Sponsored content — provided for informational and promotional purposes.
Snowflake‘s PostgreSQL gambit
This is actually pretty clever when you think about it. Most companies aren’t about to rip out PostgreSQL – it’s everywhere. So instead of fighting that reality, Snowflake is building bridges. The pg_lake extensions basically let PostgreSQL teams gradually adopt Snowflake for analytics without treating it as an all-or-nothing decision.
And here’s the thing – this isn’t just about being nice to developers. It’s about survival in the AI era. Analyst Robert Kramer from Moor Insights & Strategy nailed it when he said buyers are struggling to differentiate between Snowflake, Databricks, and other cloud platforms. When everyone’s shouting about AI capabilities, you need to make adoption as frictionless as possible.
The lakehouse wars heat up
Remember that Snowflake didn’t invent the lakehouse concept – that was Databricks five years ago. But now everyone’s fighting for that same territory. What’s interesting is how Snowflake is positioning itself as the “reliable and responsible” AI platform rather than just a testing ground.
The pg_lake GitHub repository shows they’re serious about this being a proper open source play, not just marketing. And their engineering blog post explains the technical details pretty transparently. But will developers trust Snowflake with their PostgreSQL workflows? That’s the billion-dollar question.
AI reality check
Snowflake Intelligence sounds impressive – natural language queries, insights at every employee’s fingertips. But Kramer rightly points out that Snowflake still needs to prove itself on scale, monitoring, and real-world costs for agent workloads.
Basically, everyone’s racing to build the AI platform that enterprises will actually use in production, not just for experiments. Snowflake’s bet seems to be that by making their platform accessible through tools developers already use (like PostgreSQL), they can win that race. Their broader blog post about PostgreSQL and enterprise AI shows they’re thinking strategically about this.
So what’s the bottom line? Snowflake is playing the long game here. They’re not expecting companies to abandon everything and move to Snowflake overnight. Instead, they’re creating on-ramps. Smart move? Probably. But in the hyperscale AI platform wars, clever technical moves only get you so far. Execution and proving real business value will determine who wins.
