Snowflake Bets Big on AWS AI Chips in New $6 Billion Cloud Agreement

Snowflake and Amazon Web Services have signed a new five-year agreement worth $6 billion, deepening their long-running cloud partnership as demand for AI infrastructure continues accelerating across enterprise computing.

The deal significantly expands Snowflake’s use of AWS infrastructure, particularly Amazon’s custom-built Graviton processors, as the data storage company scales AI-related workloads tied to its Cortex AI platform and broader enterprise data services.

Snowflake has operated on AWS since its founding, though its products are also available through Microsoft Azure and Google Cloud. According to AWS, Snowflake has generated roughly $7 billion in sales through AWS Marketplace since 2012, making the new agreement nearly equivalent to the total business the companies had done together previously.

The companies say customer spending is growing rapidly alongside enterprise adoption of AI tools. Snowflake said its customers are on pace to spend $2 billion on AWS infrastructure in 2025 alone, roughly double prior levels.

A major focus of the agreement involves expanded access to AWS’s Graviton chips, Amazon’s internally developed ARM-based processors. While GPUs remain central to training and running large AI models, CPUs continue handling much of the infrastructure surrounding AI applications, particularly workloads tied to automation and AI agents.

Snowflake’s Cortex AI platform has increasingly positioned the company at the center of enterprise AI deployments by allowing businesses to use natural language tools and automated analytics directly on top of stored corporate data. Features include conversational database queries and AI-generated summaries built around enterprise datasets already housed inside Snowflake systems.

The agreement also underscores how aggressively cloud providers are pushing their own AI hardware as an alternative to Nvidia’s dominance in AI computing. Amazon CEO Andy Jassy recently argued that Amazon’s internally developed chips deliver “better price-performance” than Nvidia hardware, even as AWS continues deploying Nvidia GPUs throughout its cloud business.

AWS has been expanding similar infrastructure agreements across the AI sector. Last month, the company signed a deal with Meta involving millions of Graviton chips to support Meta’s AI computing needs, following a separate multibillion-dollar cloud agreement Meta reached with Google Cloud earlier this year.

At the same time, competition around AI chips is intensifying across the industry. Google has spent years developing its own AI processors, while Microsoft recently introduced its Maia AI chip platform.

Nvidia, meanwhile, has continued defending its position at the center of the AI hardware market. CEO Jensen Huang recently described the company’s new AI-focused CPU, Vera, as part of a “brand new” $200 billion market and said Nvidia had already sold $20 billion worth of the product.

For AWS, the Snowflake agreement reflects how the AI boom is driving massive infrastructure spending across cloud computing. As companies race to build AI products and automation systems, cloud providers supplying the underlying compute resources are capturing a growing share of the spending surge.

This analysis is based on reporting from TechCrunch.

Image courtesy of Yahoo Finance.

This article was generated with AI assistance and reviewed for accuracy and quality.