Anthropic is expanding to Colossus2. Will use GB200

Anthropic says it’s expanding with SpaceX and moving more of Claude, its chatbot, onto the new Colossus 2 system using Nvidia’s latest chips. On paper, it’s a big business update: more computing power, more capacity, more AI demand. But in the comments, nobody wanted to just clap politely. They immediately turned it into a full-blown “is xAI backing down?” saga.

The loudest theory was that Musk’s own AI company might be handing the keys to rivals instead of keeping every machine for itself. One commenter basically read the tea leaves and declared this yet another sign xAI could be slipping in the race to build the smartest AI. Others piled on with the obvious irony: Musk has blasted Claude as too politically correct, yet now his broader empire is helping give it a home. That contradiction lit up the thread, with people joking that principle suddenly gets very flexible when there’s rental money and business optics involved.

Then came the split-screen debate. Some argued this is just practical: maybe top-tier AI is becoming too expensive, and the future is cheaper copycat versions that are “good enough.” Others saw a darker angle, pointing to environmental criticism around the giant data center and its power setup, saying the real cost isn’t showing up on the invoice. So yes, Anthropic got more room for Claude — but the crowd’s real obsession was whether Musk is making a master move, a money move, or a panic move.