Today Snowflake and Anthropic announce a major expansion of their strategic partnership: a multiyear $200 million agreement to bring Claude models directly into the Snowflake platform and deploy AI agents at large enterprises worldwide.
What the partnership announces
The multiyear $200M deal not only makes Anthropic's models available in Snowflake for more than 12,600 customers, it also creates a joint go-to-market initiative to bring agents to large enterprises. Claude will be accessible in Snowflake and, at the same time, across the three major cloud services: Amazon Bedrock, Google Cloud Vertex AI and Microsoft Azure.
The core idea is simple: combine Claude's reasoning with Snowflake's governed data so business teams can get answers from structured and unstructured data while keeping enterprise-grade security controls.
Why it matters for your company
Can you imagine asking in plain language and having the AI search, combine and explain data without you writing SQL or moving data out of the controlled environment? That's the goal.
Snowflake already uses Claude internally: Claude Code for developer productivity and a GTM assistant that helps speed up sales cycles. Thousands of customers already process billions of Claude tokens per month via Snowflake Cortex AI.
The next phase is deploying agents capable of multi-step analysis and advanced reasoning. According to Snowflake's internal benchmarks, Claude exceeds 90% accuracy on complex text-to-SQL tasks, which helps move projects from tests to pilots and then to production, even in regulated industries like finance, health and life sciences.
For Anthropic, this means bringing frontier models to where the data already lives. For Snowflake, it means raising intelligence over governed data, with a focus on scale and customer value.
Key benefits delivered by the partnership
-
Enterprise intelligence with Claude:
Claude Sonnet 4.5powers Snowflake Intelligence, an enterprise agent that answers from structured and unstructured data in natural language. -
Multimodal analysis across all data types: via
Snowflake Cortex AI Functions, companies can query text, images, audio and traditional tables, even using SQL when convenient. -
Building multi-agent solutions:
Snowflake Cortex Agentslets you create production data agents that retrieve and reason over diverse data with built-in precision controls. -
Integrated governance and observability:
Snowflake Horizon Catalogprovides end-to-end control and responsible AI tools to take agents from pilot to production with confidence.
Real cases and practical examples
-
Simon Data uses Claude in Snowflake to surface patterns that were previously hidden, while keeping governance standards intact.
-
Intercom powers its Fin AI Agent via
Snowflake Cortex AI, which allowed them to automate higher volumes of support and better serve large customers. -
A wealth management firm can build an agent that synthesizes client positions, market data and compliance rules to generate personalized recommendations inside Snowflake's secure perimeter.
These examples show it's not just about single answers, but about agents that "show their work" and support complex processes.
How to get started
Interested companies can begin with the quickstart guide and Anthropic's Enterprise page to deploy Claude in Snowflake. One notable point: Claude is currently the only frontier model available across the three major cloud services, which makes multi-cloud deployment options easier.
Final thought
This alliance is a clear step toward making advanced models useful inside the data environments companies have already invested in. If you work in a regulated organization or handle critical data, the combination of Claude's reasoning and Snowflake's governance could speed the jump from pilots to production. Ready to think about agents that work with your data instead of forcing you to adapt your processes to AI?
Original source
https://www.anthropic.com/news/snowflake-anthropic-expanded-partnership