Anthropic taps Elon Musk's SpaceX for more AI compute power

In a significant move to enhance its AI capabilities, Anthropic has partnered with SpaceX to utilize the Colossus One data center for computing power. This collaboration comes as the demand for Anthropic's AI coding service, Claude Code, continues to escalate. Ami Vora, the Chief Product Officer at Anthropic, made this announcement at the company's developer conference held in San Francisco. She revealed that Anthropic is set to double the existing five-hour rate limit for developers across its Pro, Max, Team, and Enterprise plans. Additionally, the company will increase rate limits for developers utilizing the Claude Opus model's API, allowing for greater input and output per minute. Faced with the challenge of strained computing resources, Anthropic turned to SpaceX, which acquired Elon Musk's xAI lab earlier this year. The Colossus One data center, located in Memphis, is one of the largest globally, equipped with tens of thousands of advanced Nvidia chips designed to handle intensive AI computations. However, the rapid expansion of this facility has raised concerns among local residents regarding pollution from gas turbines. Anthropic anticipates that this partnership will provide over 300 megawatts of computing capacity within a month, leveraging more than 220,000 Nvidia GPUs. This initiative aligns with Anthropic’s strategy of collaborating with major hyperscalers, as the company also seeks partnerships with tech giants like Google, Amazon, and Microsoft. The deal marks a noteworthy step in SpaceX's nascent compute sales business. Additionally, reports earlier this year indicated that Cursor, a coding startup, planned to train its latest model using xAI's GPUs. Through these computing agreements, SpaceX and xAI aim to generate revenue from their infrastructure while simultaneously advancing AI model development. Interestingly, Musk had previously criticized Anthropic's Claude model, labeling it as "misanthropic and evil." However, his perspective appears to have shifted after spending time with Anthropic leadership last week. He expressed on social media that he was impressed by their commitment to ethical AI development, stating, "Everyone I met was highly competent and cared a great deal about doing the right thing." Moreover, in a subtle hint within the press release regarding the partnership, Anthropic mentioned that it has shown interest in collaborating with SpaceX to create data centers in space, potentially paving the way for future innovations.