In more good news for Amazon, Snowflake signs $6B deal with AWS for AI CPU chips

Background

Snowflake is a leading data cloud platform that heavily relies on powerful computing infrastructure for its services. Amazon Web Services (AWS) is not only a major cloud provider but also develops its own custom AI chips, such as the Trainium and Inferentia processors. This deal represents a strategic alignment between a major data platform and a cloud giant's proprietary silicon.

Key Points

• Scale and Commitment: The agreement is described as "enormous" and spans five years, indicating a profound long-term commitment and investment from Snowflake in Amazon's custom chip technology.
• Direct Challenge to Nvidia: The article explicitly states Nvidia is "being put on notice." This deal is not just a procurement choice but a strategic signal that large-scale AI infrastructure can be built and sustained on non-Nvidia hardware, particularly custom ASICs (Application-Specific Integrated Circuits) from cloud providers.
• Strategic Motivation: For Snowflake, securing a dedicated, long-term chip supply mitigates risk from the global shortage of Nvidia GPUs and potentially reduces costs. For Amazon, it validates its in-house chip strategy and deepens the lock-in of a high-profile customer onto its ecosystem.

Significance

• Erosion of Nvidia's Moat: While Nvidia's GPUs and CUDA software ecosystem remain dominant, deals like this demonstrate that hyperscalers are successfully building viable alternatives for specific, large-scale AI workloads. This could gradually erode Nvidia's pricing power and market share.
• Rise of the Custom Silicon Model: This agreement underscores the growing trend where cloud providers (AWS, Google, Microsoft) design their own optimized chips for internal and customer use. The battleground is shifting from pure chip performance to vertically integrated cloud ecosystems that bundle compute, storage, and AI services.
• Market Signal: It serves as a powerful endorsement of Amazon's custom AI chips, encouraging other large enterprises to consider alternatives to Nvidia. This could accelerate investment and development in competing AI chip architectures, fostering a more diverse and competitive market.