An Agreement That Defines the Direction of Enterprise AI
On May 27th, Snowflake and Amazon Web Services announced a USD 6 billion strategic agreement spread over five years. This deal is more than financial: it represents a concrete commitment from both companies toward agentic AI as critical infrastructure for the future of enterprise computing.
The figure speaks for itself: over 11 years since its founding, Snowflake sold approximately USD 7 billion through the AWS Marketplace. This new contract nearly matches that entire history in a single five-year agreement.
Graviton: The Chip Built for Agentic AI
At the heart of this agreement are Amazon's Graviton processors, chips specifically designed for the workloads that agentic AI demands. Unlike traditional processors optimized for computational spikes, Graviton was built to sustain steady, general-purpose computing power — the key characteristic agents need to orchestrate multiple operations simultaneously without interruption.
This specificity is intentional. Agentic AI doesn't function like conventional workloads. It requires the ability to maintain context, execute chained decisions, and coordinate multiple sub-processes without the overhead of architectures designed for other tasks. AWS is betting that Graviton will become the de facto standard for this new class of applications.
Growth as Proof: From Experimental to Strategic
The growth numbers are already validating the bet. In 2025, Snowflake customers' spending on AWS doubled, reaching USD 2 billion annually. That growth was driven directly by Cortex AI, Snowflake's tool that lets enterprises query their own data in natural language without requiring technical expertise.
This combination — data accessible in natural language, processed on infrastructure optimized for AI — sits at the core of how artificial intelligence begins to deliver concrete value in enterprise operations. It's not science fiction or a prototype: it's real money flowing into these platforms because they're demonstrating return on investment.
What It Means for Technology Leaders
For any chief technology officer or CIO mapping their infrastructure roadmap, this agreement sends an unmistakable message: agentic AI has stopped being an experimental expense or isolated innovation initiative. It's a strategic bet worth billions that industry-leading companies (like AWS and Snowflake) are making now.
The implications are profound:
- Infrastructure as competitive advantage: Providers are designing silicon and services specifically for agents, not as an afterthought. If your architecture doesn't evolve to support these workloads, you'll fall behind.
- The cost of inaction: If spending on agentic AI doubles year-over-year, waiting for the technology to mature further means missing the opportunity to capture operational efficiencies today.
- The convergence of data and intelligence: The real value lies not just in the AI model, but in the ability to connect it to enterprise data safely and with proper governance. Snowflake and AWS are building that as a single integrated block.
In a context where speed and operational efficiency determine winners in markets, enterprises that consolidate their data and AI infrastructure will have a durable advantage.