Anthropic expands compute on Colossus 2

The company is expanding its partnership with SpaceX. Throughout June, Anthropic will ramp up capacity powered by NVIDIA GB200 chips in the Colossus 2 cluster.
Inference for Claude models on this infrastructure will scale up aggressively in the coming days. Anthropic’s leadership notes that meeting surging AI demand requires massive physical resources, and SpaceX’s expertise in logistics and infrastructure construction is a key asset.
The strategic move will let the company increase model throughput significantly. The latest GB200 accelerators ensure a performance boost on complex queries.
Up next
Anthropic to pay SpaceX $15 billion annually for compute access
Anthropic will pay SpaceX $15 billion a year to rent compute capacity. This deal shows that in the AI race, the scarcest resource isn’t algorithms but pure compute power. Elon Musk is effectively turning his data centers into a goldmine by renting them out to a direct competitor.
Related stories
- Mythos put Claude on AMD in about 3 days, not 6-12 months
- FT: Anthropic strikes $3.4B gas power deal for data centers
- Anthropic’s amended S-1 lists $895B in compute commitments
- Banks shelve bonds for Anthropic-tied Texas data center
- Anthropic’s S-1 caps SpaceX compute spend at $120.5B
- Anthropic offers Blackstone a double-digit yield on chips
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

