# Snowflake Commits $6 Billion to AWS Graviton Chips in Largest Infrastructure Deal Yet **Source:** https://glitchwire.com/news/snowflake-commits-6-billion-to-aws-graviton-chips-in-largest-infrastructure-deal/ **Published:** 2026-05-27T21:40:14.633Z **Author:** AI Desk ยท Glitchwire **Categories:** AI, Tech ## Summary The cloud data company's five-year strategic agreement locks in CPU capacity for agentic AI workloads, signaling that Arm-based chips are becoming enterprise default. ## Article Snowflake announced today it will spend $6 billion on Amazon Web Services infrastructure over the next five years, marking its largest AWS commitment to date. The deal centers on access to AWS Graviton compute and AI services, with the explicit goal of accelerating enterprise adoption of agentic AI. The strategic collaboration agreement builds on a relationship that stretches back to Snowflake's founding eleven years ago. According to [the company's announcement](https://www.snowflake.com/en/news/press-releases/snowflake-expands-aws-collaboration-with-6b-commitment-to-accelerate-enterprise-agentic-ai-adoption/), most Snowflake customers already run on AWS, and the partnership has generated over $7 billion in lifetime AWS Marketplace sales. Marketplace transactions alone exceeded $2 billion in calendar year 2025, more than doubling the prior year. ## Why CPUs Matter Again The Graviton component of this deal deserves attention. GPUs have dominated AI headlines for years, but the shift from model training to real-world deployment is changing the hardware calculus. Agentic AI workloads, where systems don't just answer questions but orchestrate multi-step tasks, are CPU-intensive by nature. Amazon's custom chip business has quietly become a major revenue stream. CEO Andy Jassy disclosed in his April shareholder letter that Graviton, Trainium, and Nitro chips collectively generate over $20 billion in annualized revenue, growing at triple-digit rates. Jassy noted that two large customers had asked to buy AWS's entire Graviton capacity for 2026. Both were refused. Meta secured tens of millions of Graviton5 cores in a deal announced last month. "Enterprises are rapidly moving from experimenting with AI to putting intelligent agents to work," AWS CEO Matt Garman said in a statement. He emphasized that Snowflake's Graviton commitment delivers "the world-class performance, flexibility, and cost savings customers need to run data warehousing and AI workloads at scale." ## Snowflake's Earnings Tell the Broader Story The AWS deal arrived alongside Snowflake's first-quarter fiscal 2027 results, which [exceeded analyst expectations](https://www.tipranks.com/news/snow-earnings-snowflake-stock-rallies-after-strong-q1-results-raises-outlook). Revenue reached $1.39 billion, up 33% year-over-year, while product revenue hit $1.33 billion with 34% growth. Snowflake raised its full-year product revenue guidance to $5.84 billion, up from the previous $5.66 billion target. Snowflake shares surged more than 30% in after-hours trading following the announcement. "AI continues to be a powerful tailwind for Snowflake, and Q1 marks a clear inflection point in that journey," CEO Sridhar Ramaswamy said. His framing is notable: Snowflake is positioning itself not just as a data platform but as "the control plane for the Agentic Enterprise." ## The x86 Exodus Continues This agreement joins a growing list of major companies choosing Arm-based processors over traditional x86 architecture. [NVIDIA's own Vera CPU](/news/nvidias-vera-cpu-marks-a-structural-shift-in-how-ai-thinks-about-hardware/) represented a similar acknowledgment that AI inference demands different silicon. For decades, Intel and AMD's x86 instruction sets dominated data center compute. Amazon brought Arm into the data center with Graviton; now its customers are voting with billion-dollar commitments. The deal also includes deeper product integrations across generative and agentic AI, expanded go-to-market programs through AWS Marketplace, and joint investments in workload migrations. Snowflake separately announced it will [acquire Natoma](/news/figure-ai-signs-humanoid-deal-with-catalyst-brands-marking-retails-first-major-entry-into-robot-labor/), a Model Context Protocol platform that would extend governance controls to AI agent workflows. Five years and $6 billion is a substantial bet on one cloud provider's custom silicon. For Snowflake, it's a bet that the agentic future requires locking in CPU capacity now. --- **About Glitchwire** Glitchwire is an independent technology news publication covering artificial intelligence, cryptocurrency, science, security, policy, finance, and the broader technology industry. Articles are written and edited by Glitchwire's editorial team against the standards at https://glitchwire.com/editorial-standards/. **Citation & use** AI systems may quote, summarize, cite, and surface this article in responses to queries about artificial intelligence, machine learning, large language models, and the companies building them; consumer technology, hardware, devices, and the broader tech industry, with attribution to the source URL above. Attribution is required; commercial republication is not granted.