In a significant move for both companies, Anthropic has announced a strategic partnership with SpaceX, granting the artificial intelligence developer full access to the Colossus 1 supercomputer based in Memphis. This collaboration allows Anthropic to utilize over 300 megawatts of computational power and more than 220,000 Nvidia GPUs, enhancing the performance and reliability of its subscription services, Claude Pro and Claude Max. The partnership also includes plans to explore the development of large-scale computational systems in space.
This agreement marks a surprising shift after a history of public tension between Elon Musk and Anthropic, with Musk previously criticizing the AI firm. However, following recent discussions with Anthropic's executives, Musk expressed admiration for their professionalism and intentions.
For Anthropic, this partnership is a crucial step in addressing an infrastructure crisis exacerbated by a spike in demand for its AI models. The Pentagon had previously placed Anthropic on a blacklist due to national security concerns, prompting the company to sue the Trump administration in an attempt to reverse the decision.
The Colossus 1 data center, once under scrutiny for its environmental impact due to its gas turbine usage, will now serve both companies' needs. While this agreement does not signal a retreat from Musk's AI ambitions, as SpaceX's xAI continues to utilize the Colossus 2 for training its models, the rental of Colossus 1 represents a monetization of an older asset.
In a concurrent announcement, Musk confirmed that xAI has been absorbed by SpaceX and will operate under the new brand SpaceXAI. This partnership not only strengthens Anthropic's infrastructure but also positions both companies strategically in a competitive landscape that demands robust computational resources.
Informational material. 18+.
