Snowflake and Anthropic have signed a $US200 million agreement to embed AI agents into enterprise data platforms. The multi-year deal integrates Anthropic’s Claude models directly into Snowflake’s governed environment.

This partnership targets CIOs struggling to move generative AI from experimental pilots to secure production workflows.

This partnership competes directly with Salesforce’s "Agentforce" and Microsoft’s Copilot Studio, which also promise secure, data-grounded AI agents. The unique selling proposition here is data gravity: bringing the model to where the data already lives (Snowflake), rather than moving data to the model.

The agreement establishes a joint go-to-market initiative focused on deploying AI agents within large global enterprises.

Snowflake Intelligence now uses Claude Sonnet 4.5 to analyse structured and unstructured data via natural language.

The integration aims to solve data leakage concerns by keeping AI reasoning within Snowflake’s security perimeter.

Snowflake Horizon Catalog provides the governance controls necessary for regulated industries like finance and healthcare.

Enterprises can deploy agents that understand customer data context while maintaining strict compliance standards.

"Enterprises have spent years building secure, trusted data environments, and now they want AI that can work within those environments without compromise," said Dario Amodei, CEO and Co-Founder of Anthropic. "This partnership brings Claude directly into Snowflake, where that data already lives. It's a meaningful step toward making frontier AI genuinely useful for businesses."

Technical Integration and Capabilities

The partnership introduces capabilities for multimodal analysis across text, images, audio, and tabular data.

Customers can access these features using SQL through Snowflake Cortex AI Functions.

Snowflake Cortex Agents allow organisations to build custom, production-ready data agents.

These agents are designed to retrieve information and reason over data with verifiable accuracy.

Snowflake claims Claude is the only frontier model available across Amazon Bedrock, Google Cloud Vertex AI, and Microsoft Azure.

The vendor states Claude achieves over 90 per cent accuracy on complex text-to-SQL tasks.

This figure relies on Snowflake’s internal benchmarks and has not been independently verified.

Snowflake CEO Sridhar Ramaswamy noted the partnership is measured by the "depth of innovation" created for customers.

"Snowflake should have an easier time selling its platform when the LLM [large language model (LLM)] part is backed by a top frontier AI vendor who will work with them on integration and make sure they have early access to new models," Torsten Volk, an analyst at Omdia told AI Business.

"In return, Anthropic receives access to Snowflake's large customer base and receives guaranteed revenue at the same time."