Cloud data storage giant Snowflake has signed a new $6 billion, five-year agreement with Amazon Web Services, the companies announced Wednesday.
Snowflake has always run on AWS, but has recently become available on Microsoft Azure and Google Cloud. To give you a sense of how big this deal is for these companies, Snowflake has sold a total of $7 billion worth of services through AWS Marketplace since its founding in 2012, AWS said. So this new deal is close to all the money we’ve ever made from that cloud.
That’s possible because Snowflake’s customers have recently accelerated their spending on AWS, which will double to $2 billion in that calendar year alone by 2025, Snowflake said.
Naturally, AI is driving growth. Snowflake has been offering Cortex AI, an AI building tool, for several years. This is a sensible tool. Snowflake stores much of a company’s data. AI tools can provide features such as textual interfaces for database queries (just ask questions in regular language), summary reports, and more.
Most notably, Snowflake is signing this deal to increase access to AWS’ Graviton, its own ARM-based CPU chip.
As AI moves from training to daily use to agent automation, CPU usage skyrockets. While the GPU handles training and inference, the CPU handles most of the remaining tasks related to AI, especially agents.
Amazon CEO Andy Jassy boasted last month that Amazon’s own AI chips offer “better value for money” than Nvidia’s products, although AWS still uses Nvidia chips in its cloud. The demand for AI processing is so high that cloud providers like AWS are deploying chips as quickly as possible. On top of that, all the major AI model makers (and many other AI products) are designing their apps specifically for Nvidia chips.
Still, Amazon’s own chips are a more affordable option for the cloud giant. Amazon, which has always been a price-focused company, says it is passing those savings on to customers.
As a result, these chips are attracting new multibillion-dollar deals. Last month, for example, AWS signed a deal to provide Meta with millions of Graviton chips to support its growing AI computing needs. This was a big win for AWS, as Meta had signed a $10 billion deal with Google Cloud a few months earlier.
More than that, these deals serve as a signal to Nvidia that the cloud giant’s competitive CPUs are about to come to lunch. Google has also been manufacturing its own AI chips for years. Microsoft just launched its Maia AI chip in January.
Unsurprisingly, NVIDIA CEO Jensen Huang said last week that the company is ready to defend and even grow its turf. His company’s new AI-only CPU, called Vera, represents a “brand new” $200 billion market for Nvidia, he declared after it delivered another record quarter last week, saying it has already sold $20 billion worth.
Nvidia may not easily cede market share to Amazon and other cloud providers, but AWS’s multibillion-dollar cloud deal shows how AI is lifting the ship. Regardless of which companies stand to benefit most from the rise of AI in our work and home lives, cloud providers are capturing their share.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
