Amazon Joins $130.6B Q1 AI Spend; Plans $50B Graviton Chip Unit
Amazon joined peers in spending $130.6 billion on AI during Q1 as memory chip prices climbed 50%, pushing infrastructure capex higher. The company plans a standalone Graviton and Trainium chip offering to scale its $20 billion AWS run rate toward a $50 billion external market once capacity expands.
1. Q1 AI Investment Surge
Amazon joined Google, Microsoft and Meta in spending a combined $130.6 billion on AI in Q1, marking the highest quarterly investment on record. This ramp-up reflects heavy infrastructure capex to scale compute resources for AI workloads.
2. Memory Chip Costs Climb
Prices for memory chips have risen by 50% year-to-date, inflating Amazon's AI infrastructure expenses. Higher component costs are prompting the company to revise its capex forecasts upward.
3. Standalone Chip Unit Plans
Amazon is evaluating launching a standalone business selling its custom Graviton and Trainium chips to third parties. The chip operation currently runs at a $20 billion annual revenue pace internally within AWS.
4. Scaling Capacity for External Sales
The company aims to expand production capacity beyond AWS demand before commercializing chips externally, targeting potential external revenues of up to $50 billion. Executing this strategy will depend on securing sufficient manufacturing resources.