Snowflake commits $6 billion to AWS for AI processing power
The agreement, valued at nearly the total revenue Snowflake has generated via the AWS Marketplace since 2012, signals intensifying competition between cloud providers’ proprietary chips and Nvidia’s dominant GPU market.
Cloud data storage provider Snowflake has entered into a five-year, $6 billion agreement with Amazon Web Services (AWS) to secure access to artificial intelligence processing chips. The contract, announced on Wednesday, focuses on AWS’s home-grown, ARM-based Graviton CPUs, which are increasingly deployed for AI inference and agent-based tasks.
The scale of the commitment is significant for both parties. According to AWS, the $6 billion deal represents nearly the total revenue Snowflake has generated via the AWS Marketplace since its founding in 2012. Snowflake has historically operated on AWS, although its services are also available on Microsoft Azure and Google Cloud.
Snowflake’s spending on AWS doubled in 2025 to $2 billion for that calendar year alone. This acceleration is driven by the adoption of Cortex AI, an AI building tool that allows enterprises to interact with their data through natural language queries and automated reporting. As AI workloads shift from training to daily usage and automation, the demand for CPU processing power has surged, making Graviton chips a critical component of Snowflake’s infrastructure.
The agreement highlights the broader strategic push by cloud providers to deploy custom silicon as a cost-effective alternative to Nvidia’s graphics processing units. Amazon CEO Andy Jassy has previously asserted that Graviton offers superior price-performance compared to Nvidia’s offerings. AWS is rapidly deploying these chips to meet high demand, with Amazon CEO Andy Jassy claiming the company passes these savings on to customers.
This deal follows recent moves by AWS to secure other major clients, including a separate agreement to provide millions of Graviton chips to Meta. Meta had previously signed a $10 billion deal with Google Cloud. Meanwhile, Microsoft launched its Maia AI chip in January, and Google has been developing its own AI chips for several years.
Nvidia remains a dominant force in the sector, with CEO Jensen Huang stating that his company’s new AI-specific CPU, Vera, represents a “brand new” $200 billion market. Huang noted that $20 billion worth of Vera chips had already been sold following a record-breaking quarter. Despite Nvidia’s entrenched position in AI model architecture, the influx of multi-billion-dollar contracts for cloud providers’ custom silicon signals a competitive shift in the market.
