Anthropic announced on Wednesday that it has struck a deal to access the computing resources of Elon Musk’s SpaceX, signaling a resolution with its former critic and providing a mutual boost in the competitive field of artificial intelligence. The agreement allows Anthropic to utilize the full computing power of SpaceX’s Colossus 1 facility in Memphis, Tennessee, which is equipped with over 220,000 Nvidia processors. This partnership will grant the Claude chatbot creator an additional 300 megawatts of capacity within a month.
The deal not only gives SpaceX, which is preparing for an initial public offering, a prominent client to showcase its AI aspirations to investors but also assists Anthropic in addressing capacity constraints resulting from escalating demand for its products like the AI coding tool, Claude Code. The announcement coincided with Anthropic’s developer day event in San Francisco, where the company introduced a new feature for its Claude AI called “dreaming.” This feature is designed to aid the AI systems in learning by analyzing work during intervals, identifying patterns, and updating stored user preferences and context data.
With the increased capacity from the SpaceX collaboration and other recent agreements, Anthropic revealed that it is enhancing its services. This includes doubling the rate limits for Claude Code’s paid plans, eliminating peak-hour usage restrictions for Pro and Max accounts, and significantly raising the number of requests developers can make to its Claude Opus models.
Anthropic also expressed interest in collaborating with SpaceX to create multiple gigawatts of space-based orbital data centers, aligning with one of Musk’s primary objectives and a key driver behind SpaceX’s upcoming public offering, which is anticipated to be capital-intensive and technically intricate. Ryan Mallory, CEO of data center operator Flexential, emphasized the industry’s fervent pursuit of power and scale as evidenced by discussions about space-based compute capacity.
In a statement, Musk acknowledged a shift in his stance towards Anthropic, attributing his decision to lease computing power to the company’s commitment to ensuring its AI technology benefits humanity. Musk praised Anthropic’s work and emphasized the importance of developing AI for the greater good. He mentioned that SpaceX has transitioned its AI training activities to Colossus 2 and expressed willingness to provide computing resources to other AI firms that share a similar vision of using technology to advance humanity.
The comments from Musk represent a departure from his previous criticisms of Anthropic’s AI, where he had expressed concerns about bias. Musk’s willingness to support companies focused on ethical AI development marks a significant change in his perspective. Anthropic’s introduction of the “dreaming” feature, aimed at enhancing its tools for managing AI agents, is part of the company’s strategy to attract enterprise customers amidst growing competition in the AI sector.
During the San Francisco event, Boris Cherny, Anthropic’s head of Claude Code, demonstrated how developers can utilize routines to optimize the AI computer programmer’s actions. Cherny emphasized the productivity potential of AI technology, highlighting the need to expedite its implementation for maximum impact.
