Snowflake has signed a multi-year strategic collaboration agreement with Amazon Web Services, and it is backing the partnership with real money: a $6 billion commitment to AWS infrastructure over the next five years. It is the largest infrastructure spend the data-cloud company has ever committed to, and a clear signal of how much compute the industry thinks enterprise AI is about to demand.
The deal deepens a relationship that goes back to Snowflake’s earliest days — the company has run on AWS since it launched — but the framing this time is squarely about AI. Snowflake and AWS say the expanded agreement is designed to accelerate enterprise adoption of generative and “agentic” AI, the industry’s term of the moment for systems that do not just answer questions but take actions across workflows.
What the money buys
The $6 billion figure covers AWS infrastructure spend, and it comes alongside expanded joint investment in customer migrations, industry-specific solutions and the AWS Marketplace, where Snowflake says it has now surpassed $7 billion in lifetime sales. Part of the commitment involves running more Snowflake workloads on AWS’s Graviton chips, the custom Arm-based processors AWS pitches as cheaper and more power-efficient than standard silicon for data and AI jobs.
The AI hook is Snowflake Cortex AI, the company’s managed layer for building and deploying AI applications directly on governed data. Snowflake’s pitch is that enterprises can run tasks like text-to-SQL, summarization, sentiment analysis and entity extraction without moving sensitive data out of their secure environment — a governance argument that has become central to how vendors sell AI to nervous enterprise buyers. Snowflake points to customers including Fetch and Hex as already building on the stack.
“We are moving into the era of the agentic enterprise, where AI systems don’t just answer questions, but help organizations reason over trusted data, coordinate workflows, and drive real business outcomes.”
Sridhar Ramaswamy, CEO of Snowflake
AWS CEO Matt Garman, for his part, leaned on the performance-and-cost angle, noting Snowflake’s “deepened commitment to run on Graviton” and framing it as the way to deliver the flexibility and savings customers want for data and AI at scale.
Why it matters
Strip away the language and this is a capacity bet. A $6 billion, five-year commitment is the kind of number a company signs when it expects AI and data workloads to keep climbing steeply — and when it wants guaranteed access to the infrastructure to serve them. It also tightens Snowflake’s alignment with a single cloud provider at a moment when plenty of enterprises say they want multi-cloud flexibility, a tension worth watching as agentic-AI hype meets real deployment budgets.
For now, the message from both sides is confidence: the data is governed, the chips are cheaper, and the agents, they insist, are finally ready to ship.
