As previously reported by AIM, the Snowflake-Databricks rivalry has entered the Postgres space, a move Databricks made official with its new product announcements at the Data + AI Summit 2025.

The company launched Lakebase, a fully managed Postgres database integrated with its Data Intelligence Platform and built for AI-driven application development. Now in public preview, Lakebase adds an operational layer to the lakehouse ecosystem, enabling developers and enterprises to build AI applications and agents on a single, multi-cloud platform.

“Postgres has essentially won as the lingua franca for open source databases,” said Databricks chief Ali Ghodsi while announcing the product. He added that the best thing about Postgres is the extensions that the community has built over the years. 

“It’s like a massive ecosystem of extensions that are available, that let you do anything, whether it’s vector processing, key-value store, or whatever you want to do — it’s all part of Postgres open source extensions.”

According to Stack Overflow’s 2023 and 2024 Developer Surveys, PostgreSQL has surged in popularity, surpassing MySQL as the most preferred database among developers.

Databricks is positioning Lakebase as a new entrant in the operational database market, targeting a space worth over $100 billion. The database is powered by Neon’s architecture, which separates compute from storage to support independent scaling. This design offers low latency, high concurrency, and high availability for transactional needs.

“We’re creating a new category in the database market: a modern Postgres database, deeply integrated with the lakehouse and today’s development stacks. With Lakebase, we’re giving [enterprises] a database built for the demands of the AI era,” said Ghodsi. 

The company said traditional operational databases are based on legacy architectures and fall short of modern AI use cases, which require rapid access to real-time data and integration between analytical and operational systems. Lakebase addresses this by syncing automatically with lakehouse tables, supporting AI model serving via an online feature store, and integrating with Databricks Apps and Unity Catalogue.

Lakebase is built on open-source Postgres, which benefits from a large developer community and is compatible with existing tools and extensions. It includes a branching feature that allows for copy-on-write clones of databases, useful for agent-based development and testing. The system launches in under a second and supports usage-based pricing.

Better Than Legacy Databases 

“OLTP databases today, whether proprietary like Oracle or open source like MySQL and Postgres, still look largely the same as they did in the 90s, showing little architectural evolution,” said Databricks co-founder Reynold Xin. 

He explained that Databricks Lakebase separates compute and storage into three layers to meet OLTP demands. Data is stored in scalable object stores, with a caching layer in between to reduce latency and manage WAL efficiently. On top, ephemeral Postgres nodes handle read-write operations, enabling fast, elastic compute. 

Xin added that Lakebase can guarantee customers single-digit millisecond latency at scale.

Nikita Shamgunov, founder of Neon, believes we are entering the age of AI-generated software, where AI agents, not humans, will increasingly create applications, including the databases they rely on. “In a couple of years, I think 99% of all the databases on the platform will be created by AI agents. I think we’re at the dawn of the AI software revolution, and every engineer is becoming an AI engineer.”

He further explained how Neon’s architecture, like branching and isolation, makes it ideal for AI agents, who often make mistakes and need safe, disposable environments. Shamguvnov expresses excitement about building the best OLTP system, fully integrated with data and AI platforms, bridging the gap between analytical and transactional use cases.

“Lakebase removes the operational burden of managing transactional databases,” said Anjan Kundavaram, chief product officer at Fivetran. “Our customers can focus on building applications instead of worrying about provisioning, tuning and scaling.”

Similarly, Jelle Van Etten, head of global data platform at Heineken, said, “Our analytical data platform is now evolving to be an operational AI data platform and needs to deliver those insights to applications at low latency.”

Databricks announced that numerous enterprises have participated in Lakebase’s private preview. The applications included retail personalisation, healthcare workflows, and agent-based experiences.

Better Than Snowflake? 

This new product from Databricks comes just days after Snowflake acquired Crunchy Data. Snowflake called Postgres a top choice for developers due to its flexibility, cost-efficiency, and native AI features such as vector support (pgvector).

Snowflake Postgres builds on Snowflake’s earlier move into transactional data with Unistore, which combines transactional and analytical workloads in one system.  Building on native PostgreSQL support, Snowflake Postgres extends that vision, offering enterprises a production-ready solution for transactional applications that require Postgres compatibility.

“We’re tackling a massive $350 billion market opportunity and a real need for our customers to bring Postgres to the Snowflake AI Data Cloud,” said Vivek Raghunathan, Snowflake’s SVP of engineering.

Everyone seems to be moving to PostgreSQL, and Snowflake and Databricks are joining in by acquiring smaller players. It’s not just about adding more databases, it’s about getting ready for AI, real-time data and growing business demands.

The post ‘Postgres Has Won’ appeared first on Analytics India Magazine.