Snowflake Partners with AWS in $6B Deal for AI CPUs, Boosting Amazon

Cloud data storage leader Snowflake has finalized a significant $6 billion, five-year deal with Amazon Web Services, as confirmed by both companies on Wednesday.

Since its inception in 2012, Snowflake has primarily operated on AWS, although it has since expanded to Microsoft Azure and Google Cloud. To highlight the magnitude of this agreement, it’s notable that Snowflake has generated a total of $7 billion in sales through AWS Marketplace over its entire history. Consequently, this new arrangement nearly matches its total earnings from this platform.

The deal’s feasibility is attributed to a notable increase in spending by Snowflake’s clients on AWS, which is expected to double to $2 billion in 2025 alone.

This growth is driven largely by advancements in AI. Snowflake significantly invested in its AI development tool, Cortex AI, over the past couple of years. The tool is invaluable as it operates on Snowflake’s extensive data repositories, offering features like natural language queries and concise reporting.

A crucial aspect of this agreement includes expanded access to AWS’s proprietary ARM-based CPU chip, Graviton.

As AI transitions to everyday usage and automation, the demand for CPU power intensifies. While GPUs are essential for training AI models, CPUs take charge of various tasks essential for running AI agents.

Last month, Amazon CEO Andy Jassy highlighted that AWS’s custom AI chips provide superior price-performance compared to Nvidia’s. However, AWS continues to utilize Nvidia’s chips in its cloud offerings. The soaring demand for AI computation is prompting cloud providers like AWS to accelerate their chip deployment. Additionally, many AI developers have tailored their applications for Nvidia’s technology.

Despite this, Amazon’s chips offer a cost-effective alternative. Staying true to its reputation as a cost-efficient provider, Amazon claims that these savings are passed on to its clientele.

Consequently, new large-scale contracts are emerging. Recently, AWS secured a deal to supply millions of Graviton chips to Meta to support its expanding AI needs — a notable achievement considering Meta previously entered a $10 billion agreement with Google Cloud.

These collaborations signify a competitive shift, signaling to Nvidia that cloud giants are encroaching on its territory. Google has been developing its AI chips, and Microsoft unveiled its Maia AI chip this January.

Nvidia CEO Jensen Huang recently expressed his readiness to safeguard and expand his company’s market share. With the launch of their AI-specific CPU named Vera, Huang emphasized a ‘brand new’ $200 billion market opportunity for Nvidia, following another record-breaking quarter, claiming $20 billion in sales.

While Nvidia is unlikely to concede market share effortlessly, the recent multi-billion-dollar contracts secured by AWS illustrate how AI is positively impacting its growth. As AI continues to integrate into everyday practices, cloud service providers are also set to benefit significantly.