Snowflake CEO Sridhar Ramaswamy poses in front of the company’s sign outside the New York Stock Exchange on September 30, 2025.
new york stock exchange
Amazon On Wednesday, the company’s cloud division announced it had committed to spending $6 billion. snowflakeThis includes the use of the company’s custom silicon and chips for artificial intelligence.
Snowflake’s purchase of services and technology from Amazon Web Services will be over a five-year period, according to a press release about the deal. Snowflake plans to expand its use of Amazon’s Graviton general-purpose chips and cloud-based graphics processing units for AI.
This is the latest sign that AWS is gaining momentum in the AI space as customers turn to the market-leading cloud for more advanced technology. Claude creator Anthropic said in April that it aims to invest more than $100 billion in AWS over 10 years. Amazon also has an agreement with OpenAI.
Both contracts with AI modeling companies include equity investments, but Snowflake’s deal does not. Snowflake, which went public in 2020 with a market capitalization of just over $60 billion, has long relied on AWS.
Snowflake’s stock price soared as much as 35% in after-hours trading after the company announced strong results for its fiscal first quarter, which ended April 30. The company reported adjusted earnings of 39 cents per share on revenue of $1.39 billion, an increase of 33% from the same period last year. Analysts polled by LSEG had expected earnings of 32 cents a share and revenue of $1.32 billion.
Snowflake also expected product revenue of $1.415 billion to $1.42 billion and adjusted operating margin of 12.5% in the fiscal second quarter, with guidance also strong. Analysts surveyed by StreetAccount expected product revenue of $1.37 billion and a profit margin of 11.9%.
Snowflake also announced that it will acquire AI startup Natoma for an undisclosed sum.
At the time of Snowflake’s IPO, the company disclosed an amended agreement with an unnamed cloud provider to spend $1.2 billion over five years and pay $350 million in the final year. The provider was Amazon, a Snowflake spokesperson told CNBC. By 2023, the agreement has grown to $2.5 billion.
The new $6 billion deal means average annual spending of $1.2 billion.
AWS introduced its first Arm-based Graviton chip in 2018, and it remains the company’s most successful custom chip to date. Snowflake first talked about hiring Graviton in 2022.
Snowflakes also have a close relationship NvidiaIn November, Snowflake touted an update that simplifies the process of running AI workloads on Nvidia GPUs.
The deal means yet another big tech company opts for custom arm–based processors are preferred over processors based on traditional x86 architectures.
For decades, server chips intel 1970s and advanced micro device Decades later. Arm’s alternative power-efficient architecture has become mainstream. apple But it was Amazon that brought Arm chips to data centers with Graviton. cloud rivals google and microsoft Following Amazon, they released a custom Arm chip.
Fast forward to 2026, and the demand for central processing units like Graviton is surging again as mass adoption of AI moves from call-answering chatbots to task-oriented agent apps.
GPUs like Nvidia’s are great for training AI models because they have thousands of small cores focused on performing many operations simultaneously, while CPUs have fewer powerful cores that perform general-purpose tasks sequentially. Agentic AI requires a lot of general computing power to move large amounts of data and coordinate across multiple agents for AI workflows.
In April, meta He said it will utilize hundreds of thousands of Graviton chips.
“Graviton is our industry-leading CPU chip that enables Meta to run CPU-intensive workloads behind Agent AI with the performance and efficiency needed at its scale,” Amazon CEO Andy Jassy said during the company’s April earnings call.
WATCH: How Amazon chips power Anthropic without Nvidia
