Snowflake Deepens Its AWS Bet

Snowflake has signed a five-year, $6 billion infrastructure commitment with Amazon Web Services, expanding a cloud relationship that already underpins a large share of Snowflake's platform. The companies framed the agreement around enterprise agentic AI, not just ordinary cloud capacity.

The official announcement says Snowflake will spend on AWS Graviton compute and AI infrastructure over the next five years. Snowflake also said it has passed $7 billion in lifetime sales through AWS Marketplace, while 2025 marketplace sales alone exceeded $2 billion.

The deal matters because Snowflake is trying to keep AI workloads close to the governed enterprise data it already stores. Its Cortex AI tools support tasks such as text-to-SQL, summarization, sentiment analysis, and entity extraction inside Snowflake environments, giving customers a way to build AI applications without moving sensitive datasets into separate systems.

For AWS, the commitment is another data point in a broader push to make custom silicon part of the AI infrastructure stack. The most conservative reading is that large enterprise AI deployments are becoming cloud spend commitments tied to specific processor families and managed services, rather than isolated software subscriptions.

Snowflake did not announce a new foundation model or consumer AI product with the deal. The practical change is capacity, integration, and go-to-market alignment for companies already running Snowflake on AWS.