Archive — US — techcrunch
In more good news for Amazon, Snowflake signs $6B deal with AWS for AI CPU chips
Snowflake has signed a $6 billion agreement with AWS to enhance AI capabilities, using AWS's Graviton CPU chips. This deal is significant as it nears the total revenue Snowflake has earned through AWS since its inception, driven by the increasing demand for AI services. Amazon's own chips are positioned as a cost-effective alternative to Nvidia's, strengthening its cloud positioning amid rising AI trends.
Key points
- Snowflake's $6 billion deal with AWS spans five years, enhancing its AI services.
- The deal is approaching Snowflake's total $7 billion revenue from AWS Marketplace.
- Snowflake's customer spending on AWS is projected to double to $2 billion in 2025, driven by AI adoption.
- Snowflake uses AWS’s ARM-based Graviton chips to support AI operations.
- CPUs manage daily AI tasks while GPUs are used for training; demand for AI processing grows.
- AWS claims its Graviton chips offer better price-performance than Nvidia, though it still utilizes Nvidia's chips.
- Amazon's competitive pricing aims to attract new multi-billion-dollar contracts, posing a challenge for Nvidia.
Why it matters
- The rise of AI has significantly increased demand for cloud computing resources across the industry.
- Amazon and other cloud providers are developing their own chips to compete against Nvidia, which has dominated the market.
- Past collaborations and shifts in deals (e.g., Meta's previous $10 billion deal with Google Cloud) showcase competitive tensions in AI infrastructure.
- The AI landscape is rapidly evolving with new players like Microsoft introducing AI-specific chips.
The $6 billion deal between Snowflake and AWS illustrates the growing synergy between AI development and cloud capabilities, highlighting a competitive shift in the chip market where established players like Nvidia may need to adapt to maintain their dominance.