Anthropic will gain access to SpaceX’s multi-billion dollar supercomputer known as “Colossus,” according to an announcement, striking a deal that could also result in the two firms partnering up on the development of space-based AI data centers.
Key Facts
- Under the agreement, Anthropic will use the supercomputer’s computing power to “directly improve capacity for Claude Pro and Claude Max subscribers.”
- Like using other AI models, those who use Claude’s premium versions are subject to rate limits, which are mechanisms put in place to cap the number of requests made by a user.
- With access to SpaceX’s supercomputer, which is powered by more than 220,000 Nvidia graphics cards and high-end accelerators to handle high workloads, Anthropic can give its premium Claude users higher rate limits and faster response times.
- The improvements for Claude premium users were effective Wednesday, according to Anthropic, which listed three changes that included doubling of the five-hour rate limits for Claude Code’s premium users.
- Anthropic’s agreement with SpaceX strengthens its supply chain, as the company already has compute agreements with Amazon Web Services and Google worth hundreds of billions of dollars.
What to watch for
Anthropic also showed interest in teaming up with SpaceX to develop “multiple gigawatts of orbital AI compute capacity” as part of the agreement. Space, specifically Earth’s orbit, is being eyed by AI companies as the next home for the data centers needed to develop and power artificial intelligence—with SpaceX noting in its Wednesday announcement the computing power needed to train the next generation of AI is “outpacing what terrestrial power, land, and cooling can deliver on the timelines that matter.” The Elon Musk-led company added that space-based data centers can offer “near-limitless sustainable power with less impact on Earth.”
Key Background
Anthropic’s AI hardware is now further diversified with the SpaceX deal in place. Claude, which already utilizes hardware from Amazon, Google and Nvidia, will now tack on SpaceX’s supercomputer to its backers. The “Colossus” supercomputer was built in Memphis, Tennessee, in just 122 days with the help of Nvidia. The computer, which has doubled the number of graphics cards it uses since 2024, requires a vast amount of energy that led to concerns from locals about its strain on energy and the water supply, underscoring the larger issue powering increasingly large AI workloads. Musk’s aerospace firm SpaceX merged with xAI earlier this year, positioning the company to establish an edge over its competitors by putting AI data centers in space. Musk has claimed in two to three years, the most cost-effective method to strengthen compute power for AI will be to build orbital data centers, adding solar power could create “little operating or maintenance costs.”
This story was originally published on forbes.com and all figures are in USD.
Want to see more Forbes articles on your feed? Tap here to make Forbes Australia a preferred source on Google.
Look back on the week that was with hand-picked articles from Australia and around the world. Sign up to the Forbes Australia newsletter here or become a member here.
