Snowflake’s $6 Billion AWS Bet, Defined
Snowflake’s USD 6 billion (approx. RM27.6 billion) AWS commitment is a long-term cloud infrastructure agreement that secures Graviton CPU capacity and AI accelerators so the company can power AI-driven analytics, agentic workloads, and traditional data warehousing on a single cloud data platform strategy. The multi-year deal is Snowflake’s largest cloud spend commitment so far, and it ties its AI infrastructure investment tightly to Amazon Web Services. AWS describes the agreement as focused on Arm-based Graviton processors and GPU-accelerated EC2 instances, which Snowflake will use for both AI model training and inference. This means Snowflake is not only reserving GPU power for generative AI but also locking in cost-efficient CPU cores for everyday analytics. In effect, Snowflake is turning cloud compute into a planned capital obligation rather than a purely elastic expense, betting that AI-accelerated insights will offset the cost.
Graviton CPUs, AI Accelerators and the Next-Gen Stack
Snowflake’s technical focus in the AWS deal sits at two layers: Graviton CPUs for high-volume analytics and GPU-based AI accelerators for generative workloads. Over recent years, Snowflake has shifted more compute from Intel and AMD chips to AWS’s Arm-based Graviton instances, which now reach 192 Arm Neoverse V3 cores with high-bandwidth memory. These CPUs handle SQL queries, Python functions, and agent orchestration logic that wrap around large language models, while the GPUs run the models themselves. AWS highlights “GPU-accelerated” instances rather than its Trainium chips, signaling that Nvidia-based infrastructure remains central to Snowflake’s AI stack. The combination lets Snowflake route classic data warehouse workloads to cheaper Graviton cores and reserve the expensive accelerators for Cortex AI services such as text-to-SQL, summarization, sentiment analysis, and entity extraction, tightening economics across mixed workloads.
From Data Warehouse to AI-First Platform
Under CEO Sridhar Ramaswamy, Snowflake is repositioning from a cloud data warehouse to what it calls “the platform for the AI era.” Cortex AI is the centerpiece, offering AI agents that sit on governed enterprise data to generate SQL, summarize reports, and analyze customer sentiment without exporting information to external systems. Ramaswamy describes this as an “agentic enterprise,” where AI systems reason over trusted data and coordinate workflows. The AWS alliance is designed to make that architecture easier to deploy, especially through deep integration with AWS services and marketplace distribution. Snowflake’s expansion into new AWS regions further tightens the relationship, bringing Cortex AI closer to customers’ existing cloud estates. The message is clear: Snowflake wants to be the default AI layer for data already in AWS, not an add-on tool customers bolt on later.
AI Infrastructure Moves Onto Software Balance Sheets
Snowflake’s AWS commitment highlights a broader shift: AI infrastructure investment is moving onto software balance sheets instead of sitting purely with cloud providers. In the past, software firms tried to avoid deep infrastructure exposure because it increased risk and reduced flexibility. AI has changed the math. To secure GPUs and advanced CPUs at scale, vendors are making multibillion-dollar, multi-year commitments ahead of fully realized revenue. According to Startup Fortune, Snowflake has already surpassed USD 7 billion (approx. RM32.2 billion) in lifetime AWS Marketplace sales and exceeded USD 2 billion (approx. RM9.2 billion) in sales in a single calendar year, giving AWS both infrastructure and distribution leverage. This tight alignment can boost growth but also links gross margins and operating leverage to infrastructure procurement bets, making usage growth and AI adoption critical to justify the spend.
What the Deal Signals for Enterprise AI Competition
For AWS, the agreement reinforces its status as a default partner for enterprise AI builds, as recent results show AWS revenue rising 28% in the first quarter of 2026. For Snowflake, it is a statement that AI-accelerated analytics will define its future competitive position against other cloud data platforms. The company is using Graviton CPUs to keep baseline compute costs in check while funding GPU-intensive Cortex AI services, hoping customers will pay for faster, smarter insights on their existing data. Competitors now face a choice: secure similar long-term AI infrastructure access or risk being outpaced by platforms that can guarantee capacity. As data platform vendors move deeper into the stack, the line between cloud provider and software company blurs, and the winners will likely be those that align their cloud data platform strategy with reliable, long-term AI capacity.