Overview
- AWS says Project Rainier is fully operational less than a year after announcement, running across multiple datacenters with nearly 500,000 Trainium2 chips.
- Anthropic is training and serving Claude on Rainier now, with plans to scale to more than 1 million Trainium2 chips by year-end.
- Amazon characterizes the deployment as the largest cluster of non‑Nvidia AI chips and highlights its end‑to‑end control of chips, servers, and datacenter design.
- To support AI capacity, the company reports adding over 3.8 gigawatts of power in the past 12 months, while local utilities expect significant grid strain as the campus expands.
- AWS posted about $33 billion in Q3 revenue, up roughly 20% year over year, which executives link to accelerating AI workloads.