Anthropic Locks In SpaceX's Colossus 1 Compute Capacity
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- SpaceX signed an agreement with Anthropic granting access to Colossus 1, which SpaceX calls one of the world's largest and fastest-deployed AI supercomputers, and said Anthropic is interested in partnering for orbital compute capacity.
- Anthropic inked a deal to use all of Colossus 1's compute capacity, per the Wall Street Journal, securing roughly 220,000 NVIDIA GPUs through the partnership (Coin Edition).
- Ross Nordeen, xAI cofounder, left to join Anthropic "to focus on compute" following the Colossus lease, per DatacenterDynamics.
- Axios describes the moment as an "intelligence explosion," anchored on reporting tied to Anthropic co-founder Jack Clark.
- Spyglass frames the deal as pulling SpaceX into the "neocloud orbit," recasting SpaceX as a compute provider alongside its rocket business (Semafor notes the tie-up paves way for space data centers).
Why it matters: Anthropic has locked in the full capacity of SpaceX's Colossus 1 supercomputer and pulled in an xAI cofounder, per the WSJ and DatacenterDynamics. With 220,000 NVIDIA GPUs and orbital compute talks in play (Semafor, Coin Edition), compute scarcity is redrawing the AI alliance map — SpaceX is no longer just a launch company but a neocloud player.
