Amazon Eyes Chip Unit to Hit $50B Revenue as AI Costs Surge
Amazon contributed to the $130.6B Big Tech AI spend in Q1 and will further increase R&D, driving capital needs as memory chip prices climb 50%. Meanwhile, Amazon’s custom Graviton and Trainium chips could expand from $20B to $50B annual revenue if sold externally once production scales.
1. AI Investment and Cost Pressures
Amazon joined Google, Microsoft and Meta in spending $130.6 billion on AI in Q1 2026, surpassing costs of historic R&D programs. With memory chip prices up 50% this year, the company plans to boost infrastructure capex, raising concerns about margin compression and longer-term ROI on AI initiatives.
2. Chip Business Expansion
Leveraging its custom Graviton and Trainium processors, Amazon's internal AWS chip run rate has reached $20 billion annually. Senior executives anticipate scaling production capacity to serve third-party customers, potentially lifting external chip revenues to $50 billion if market demand mirrors Nvidia’s growth trajectory.