What Snowflake’s $6B AWS Commitment Really Means
Snowflake’s USD 6 billion (approx. RM27.6 billion) AWS commitment is a long-term cloud computing spending plan in which a software company pre-buys large amounts of infrastructure capacity to secure compute for future artificial intelligence workloads and deepen its strategic ties with a hyperscale cloud provider. Snowflake’s new multi-year strategic collaboration with Amazon Web Services centers on Graviton CPU deployment and GPU-accelerated EC2 instances that will power both data warehousing and AI services. The deal is Snowflake’s largest-ever cloud spend commitment and signals how AI accelerators investment is moving onto software balance sheets. Rather than staying abstracted from infrastructure, Snowflake is locking in compute to support its Cortex AI suite, which includes text-to-SQL, summarization, sentiment analysis, and coding agents. This marks a shift from treating infrastructure as a pass-through cost toward treating it as a core strategic asset.
From Data Warehouse to AI Platform, Powered by Cloud Hardware
Snowflake’s strategy under CEO Sridhar Ramaswamy is to reposition the company from a cloud data warehouse into what it calls “the platform for the AI era.” Cortex AI brings AI workloads directly to governed data, enabling applications like text-to-SQL and entity extraction without moving data out of Snowflake. To make this vision real, Snowflake needs dependable access to AI accelerators and CPU capacity at scale. According to Amazon, Snowflake’s lifetime AWS Marketplace sales crossed USD 7 billion (approx. RM32.2 billion) and exceeded USD 2 billion (approx. RM9.2 billion) during the 2025 calendar year. That sales volume makes AWS both the infrastructure bedrock and a key distribution channel for Snowflake’s AI offerings. The tighter Snowflake integrates with AWS-native AI services and infrastructure, the more it can promise predictable performance and scale for enterprise AI projects built on its platform.
Graviton CPUs and AI Accelerators: Why the Chips Matter
At the heart of the Snowflake AWS infrastructure plan is a specific hardware mix: ARM-based Graviton CPUs for general compute and GPU-accelerated instances for AI training and inference. Snowflake has already shifted more compute from Intel and AMD to Amazon’s in-house Graviton chips, now in their fifth generation with up to 192 Arm Neoverse V3 cores. These CPUs handle SQL queries, Python functions, and orchestration logic around AI models, leaving GPUs and other AI accelerators to run the models themselves. This division of labor matters for software infrastructure costs. Cost-efficient Graviton instances can support traditional data warehousing workloads, freeing budget headroom for the more expensive AI accelerators investment required by generative AI. Snowflake is also avoiding deep lock-in to AWS-only AI chips like Trainium by prioritizing “GPU-accelerated” instances, which likely means Nvidia hardware that remains compatible with multi-cloud strategies.
AI Is Pulling Software Balance Sheets Into Infrastructure
Snowflake’s AWS deal reflects a broader shift: AI is pushing software vendors closer to the hardware layer than in the past. Historically, cloud-native software companies tried to avoid owning infrastructure risk; they sat above the hyperscalers and paid as they went. With AI, that model strains. To guarantee capacity for agentic AI workloads that plan, retrieve data, call tools, and execute tasks, vendors are making multi-billion-dollar commitments ahead of revenue. Snowflake’s USD 6 billion (approx. RM27.6 billion) obligation turns infrastructure into a balance-sheet item with implications for gross margin, operating leverage, and strategic flexibility. If AI adoption grows as expected, these long-term cloud computing spending bets can secure supply and defend premium pricing. If demand lags, the same commitments become a drag. Either way, software companies are looking more like infrastructure investors as AI reshapes their economics.
Tighter Cloud Partnerships as an AI Competitive Strategy
Snowflake’s move also shows how tighter partnerships with hyperscalers are becoming a competitive requirement in AI. The company is expanding its AWS footprint into more regions and deepening joint go-to-market efforts through AWS Marketplace, where it has already surpassed USD 7 billion (approx. RM32.2 billion) in lifetime sales. For AWS, Snowflake’s commitment reinforces its appeal as a preferred platform for enterprise AI builds, supported by scale, security features, and established procurement processes. For Snowflake, AWS is no longer only the infrastructure substrate but also a channel and co-selling engine for its AI products. That interdependence raises the stakes: switching costs increase, but so does the potential upside from shared growth. As other data platforms follow similar paths, AI-era competition will hinge not only on software features but on who has secured the most reliable, cost-effective pipeline of compute in the cloud.