Snowflake acquisition of Crunchy Data adds Postgres database

Additive capabilities

As enterprises have increased their investments in AI development over the past few years, many data management vendors, including Snowflake and Databricks, have expanded by adding environments for customers to develop, deploy and manage AI models and applications.

While Databricks has aggressively acquired other companies to build up its AI development environment, Snowflake has largely added capabilities with its own product development.

Snowflake acquired Neeva in May 2023 to kickstart its expansion into AI development -- and bring current CEO Sridhar Ramaswamy into the company. Since then, however, Snowflake has internally created its Cortex platform.

Now, Snowflake is adding PostgreSQL database capabilities in a move that, like Mohan, Stephen Catanzano, an analyst at Enterprise Strategy Group, now part of Omdia, called significant.

"It's a smart and strategic move," he said. "By acquiring Crunchy Data, Snowflake strengthens its ability to support structured, transactional data workloads critical for powering AI agents and applications that rely on a mix of operational -- Postgres -- and analytical data. This allows Snowflake to serve enterprises that demand governance, security and compliance for production AI solutions."

PostgreSQL is now the most popular database, according to the 2024 Developer Survey by Stack Overflow, outpacing the open source MySQL format and databases from Microsoft, MongoDB and Redis.

Flexibility and versatility, created by a vibrant community, are two of the main characteristics that make it so popular, according to Mohan. Beyond traditional relational database capabilities, PostgreSQL databases can handle geospatial, time series, JSON and vector database workloads.

"PostgreSQL stands out as a shining star in open source transactional databases, largely due to its vibrant community," Mohan said.

Of particular value is its vector database capabilities, Mohan continued. Vector search and storage are now a critical aspect of AI development, enabling users to give structure to unstructured data to make it searchable and letting users run similarity searches to discover relevant data.

"The acquisition will help Snowflake handle real-time, transactional data that feeds AI applications and agents and allows real-time vector embedding creation," Mohan said. "This was a weakness for Snowflake."

Catanzano, meanwhile, noted that PostgreSQL databases also help users remain compliant with industry standards and manage model metadata.

"As organizations build AI solutions, they need a database that can handle transactional data at scale," he said. "Postgres fits this role perfectly."

Snowflake acquired Crunchy Data specifically to improve its transactional data management capabilities, according to Ramaswamy.

"Transactional data systems are very important, [and] we think this is an important acquisition in helping users have access to a transactional store right within Snowflake," he said during a media conference Monday.

In addition, providing customers with options was a motivating factor for the acquisition, according to Christian Kleinerman, Snowflake's executive vice president of product.

Crunchy Data's PostgreSQL engine will not work in concert with Snowflake's existing SQL engine. However, PostgreSQL capabilities are open source and more easily interoperable with a wide variety of third-party applications, including agentic AI capabilities, than Snowflake's existing SQL capabilities.

"We are very committed to meeting customers where they are, giving them choice without overwhelming them," Kleinerman said during the press conference.

Beyond providing customers with flexibility, Snowflake's acquisition of Crunchy Data will help it compete with Databricks, according to Catanzano.

However, Catanzano and Mohan each noted that although Databricks and Snowflake have now both acquired a PostgreSQL database vendor, the capabilities Neon and Crunchy Data provide are not the same.

Neon provides a cloud-native, serverless database that enables users to separate compute and storage to make it efficient for testing and development workloads. Crunchy Data, meanwhile, focuses on providing production-ready capabilities that enable deployments at scale in Kubernetes environments.

"Their approaches and target markets diverge significantly," Mohan said.

Looking ahead

With Snowflake set to add PostgreSQL database capabilities, the acquisition of Crunchy Data should serve as a building block, according to Mohan.

If Snowflake can fully integrate Crunchy Data's capabilities, it has the potential to enable developers to build, train and deploy AI models -- including agents -- directly on transactional data within Snowflake as well as execute hybrid queries on transactional and analytical data. In addition, the acquisition could aid Snowflake's governance capabilities.

"They should leverage Crunchy's exceptional security capabilities to enhance their own governance initiatives," Mohan said.

Catanzano likewise suggested that Snowflake make integrating PostgreSQL capabilities into its AI development ecosystem a priority. In addition, adding new partnerships and more acquisitions that improve its AI development capabilities would be beneficial, he continued.

"Snowflake should invest in partnerships and acquisitions that extend its capabilities in open source AI, unstructured data management and hybrid workloads, ensuring it can serve the full range of AI and data-driven applications that modern enterprises demand," Catanzano said.

Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.