Musk's Colossus 1 AI supercomputer's inefficient mixed-architecture design couldn't be used to train Grok, so Anthropic's using it for inference instead — Musk readies unified Blackwell-only Colossus 2 for frontier training and potential IPO

Anthropic has leased xAI’s Colossus 1 data center to add inference capacity for Claude and ease usage limits.

Colossus 1 reportedly has more than 220,000 GPUs and 300 MW of power, making it one of the largest AI compute sites.

The article says its mixed GPU setup likely made it a poor fit for training Grok, even if it was useful for serving workloads.

xAI is now said to be planning Colossus 2 as a unified Blackwell-based cluster for future frontier training and possibly public-market ambitions.