Exciting News for Amazon: Snowflake Forms $6B Partnership with AWS for AI CPU Chips

Data storage leader Snowflake has finalized a new five-year deal valued at $6 billion with Amazon Web Services, as reported on Wednesday.

Historically, Snowflake has functioned on AWS, but it has also become available on Microsoft Azure and Google Cloud. To illustrate the magnitude of this deal, AWS noted that Snowflake has earned $7 billion through AWS Marketplace since its launch in 2012. Therefore, this new agreement is nearly equivalent to all the revenue it has accrued from that cloud service.

This expansion is driven by rising expenditures from Snowflake’s clients on AWS, which are expected to double to $2 billion by 2025, as projected by Snowflake.

A significant factor in this growth is AI. Over the past few years, Snowflake has introduced its AI development tool, Cortex AI. This tool utilizes the extensive data stored within Snowflake, enabling functionalities such as a straightforward text interface for database queries and summary reports.

It’s also worth noting Snowflake’s new contract that provides greater access to AWS’s proprietary ARM-based CPU chip, Graviton.

As AI evolves from training to widespread application and automation through agents, CPU utilization rises substantially. While GPUs handle training and reasoning tasks, CPUs are primarily responsible for most other AI-related duties, particularly those involving agents.

Amazon CEO Andy Jassy recently asserted that Amazon’s proprietary AI chips deliver “better price-performance” than Nvidia’s products, even though AWS continues to use Nvidia’s chips in its cloud services. With skyrocketing demand for AI processing, cloud providers like AWS are swiftly deploying chips. Furthermore, most leading AI model developers (and a plethora of other AI services) have optimized their applications to function best with Nvidia’s chips.

Nevertheless, Amazon’s own chips offer a more economical choice for the cloud giant. Consistent with its cost-sensitive strategy, Amazon claims that it passes these savings on to its customers.

Consequently, these chips are attracting new multi-billion-dollar collaborations. For example, AWS recently secured a deal to provide millions of Graviton chips to Meta to meet its growing AI computational needs. This marked a significant victory for AWS, especially given that Meta had previously entered into a $10 billion agreement with Google Cloud.

In addition, these contracts indicate to Nvidia that competitive CPUs from cloud giants are now challenging its market position. Google has been developing its own AI chips for several years, while Microsoft launched its Maia AI chip in January.

Unsurprisingly, Nvidia CEO Jensen Huang remarked last week that he is fully prepared to safeguard and expand his market share. He stated that the new AI-specific CPU introduced by his company, dubbed Vera, signifies a “brand new” $200 billion market for Nvidia, following yet another record-breaking quarter. He also mentioned that he has already sold $20 billion worth of this new technology.

While Nvidia is unlikely to easily cede market share to Amazon or any other cloud provider, AWS’s multi-billion-dollar cloud contracts highlight how AI is driving its growth. Regardless of which companies garner the most benefits from the increasing integration of AI in our personal and professional lives, cloud providers are undeniably capturing a portion of it.