Snowflake and Anthropic deepen Claude integration in Cortex

Snowflake and Anthropic deepen Claude integration in Cortex

Snowflake and Anthropic expanded their partnership to embed Claude across Snowflake Cortex, letting enterprises build AI agents and run Claude-powered Cortex Code and Snowflake Intelligence on governed data.

Snowflake and Anthropic expanded their strategic partnership to integrate Anthropic’s Claude models more deeply into Snowflake’s Cortex platform. The update follows an initial agreement in December that first brought Claude into Cortex to let customers query Snowflake data with Anthropic’s models.

The latest phase adds Claude to Snowflake Cortex Code and Snowflake Intelligence and makes the model available through the Cortex Agents framework. Enterprises can use the integration to deploy custom agents and production-grade applications that access and act on data stored and governed in Snowflake. Use cases the companies highlighted include customer support, data analysis and core operations.

Christian Kleinerman, Snowflake’s executive vice president of product, said Cortex Code has become the fastest-growing product in the company’s history and that the partnership with Anthropic is helping customers move projects from experimentation into production faster.

Block provided an example of the integration in practice. Arnaud Weber, Block’s engineering lead, commented that the company built an ‘AI-native operating layer’ connecting data across Cash App, Afterpay and Square. Weber added that combining Claude with Snowflake’s governed data platform has helped Block investigate compliance and security issues in real time, trace controls and requirements, surface operational insights, and automate workflows using trusted enterprise data.

Industry analysis shows many enterprises struggle to move AI pilots into full production. An early 2025 analysis found about 38% of companies had transitioned AI projects into production. A January 2025 study reported roughly half of AI agent projects remained at proof-of-concept, with more than half of IT leaders citing governance, security, privacy and compliance as major obstacles.

Snowflake and Anthropic stated the integration pairs model capabilities with Snowflake’s data governance, access controls and enterprise performance. The companies pointed to safety and governance features in Cortex and described ways to trace and audit agent actions that work with regulated data.

Snowflake noted potential enterprise applications beyond support and analytics, including security investigations, operational automation and developer productivity. The expanded integration makes Claude available to teams that need to develop AI applications that operate on governed data without moving it outside Snowflake.

Articles by this author