Amazon and Snowflake Join Forces: $6B Deal for AI CPU Chips

TL;DR

• Snowflake has signed a five-year, $6 billion agreement with AWS to increase its use of Graviton CPUs and other cloud infrastructure for AI workloads.
• The deal deepens Snowflake’s reliance on AWS at a time when the industry is shifting toward AI deployment and inference, where CPU demand is rising alongside GPU demand.
• For Amazon, the contract is a major win for its in-house chip strategy and adds pressure on competitors, including Nvidia, even though AWS still also uses specialized AI accelerators and GPUs.

Snowflake and Amazon Web Services have struck a major new cloud deal that underscores how quickly AI infrastructure spending is evolving. The agreement is worth $6 billion over five years, and Snowflake will expand its use of AWS’s Graviton processors as well as additional AWS AI-related compute services.

What the deal includes

According to the announcement reported by multiple outlets, Snowflake’s commitment centers on AWS’s ARM-based Graviton CPUs, which Amazon developed in-house for data-center workloads. CNBC reported that the arrangement also includes broader AWS technologies and services, with Snowflake increasing use of both general-purpose processors and cloud-based GPUs for AI workloads.

The scale of the deal is notable. CNBC said the agreement implies an average spend of about $1.2 billion per year. Yahoo Finance added that Snowflake has generated about $7 billion in revenue through AWS Marketplace since its founding, making the new commitment nearly as large as the company’s total historical AWS Marketplace revenue.

Why AWS Graviton matters

Graviton is central to Amazon’s effort to build a stronger internal chip ecosystem inside AWS. WSJ described Graviton as a CPU family launched by AWS in 2018 and positioned it as part of the broader surge in demand for computing hardware driven by AI.

The strategic appeal is straightforward: CPUs remain critical for cloud infrastructure, and as AI systems move from training into deployment, orchestration, and agentic workflows, the need for efficient general-purpose compute rises. That makes Graviton especially relevant for companies like Snowflake that run large-scale data and analytics workloads tied to AI applications.

What it means for Snowflake

For Snowflake, the deal strengthens its cloud foundation on AWS while preserving its multi-cloud footprint. Yahoo Finance noted that Snowflake runs on AWS, Microsoft Azure, and Google Cloud, so the AWS commitment expands one major relationship without necessarily eliminating the others.

The agreement also signals that Snowflake is leaning harder into the infrastructure needed for AI-era products and services. CNBC reported that the company plans to increase its use of Amazon’s specialized silicon and chips for AI applications, alongside Graviton and GPUs. In practical terms, that suggests Snowflake is optimizing for both cost efficiency and AI performance as enterprise demand grows.

The competitive pressure on Nvidia

The deal does not directly replace Nvidia in the AI hardware market, but it does highlight the growing importance of custom CPU silicon inside hyperscale clouds. Nvidia remains dominant in AI accelerators, especially for training large models, and AWS still offers chips for that purpose as well.

Even so, Amazon’s expanding chip strategy creates a more layered competitive landscape. As more workloads shift toward inference and agentic AI, buyers may increasingly mix GPUs with custom CPUs and other specialized chips, rather than relying on a single vendor. That trend can create pricing and architecture pressure across the market, including on Nvidia, which has benefited enormously from the AI buildout.

Why this deal matters now

The timing matters because the AI market is moving beyond the initial rush to train giant models. WSJ noted that AWS also provides chips designed specifically for training and running AI models, while other reporting emphasized Snowflake’s growing need for compute tied to AI workloads.

That shift helps explain why a CPU-focused deal is drawing so much attention. It is not just a procurement story; it is a signal that the next phase of AI infrastructure may be defined as much by efficiency, deployment scale, and cloud economics as by raw model training power.

The bigger picture for Amazon

For Amazon, the Snowflake agreement is another validation of its long-running push into custom silicon. AWS has spent years building an infrastructure stack that blends proprietary CPUs, AI accelerators, and cloud services, and Snowflake is now one of its most prominent customers in that ecosystem.

TechCrunch reported that the deal is being viewed as another piece of good news for Amazon because it reinforces confidence in AWS’s long-term AI strategy. In a market where cloud and AI spending are increasingly intertwined, that confidence can be as valuable as the contract itself.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.