Keep your friends close and your enemies closer.
Yesterday, the artificial intelligence startup developer of the Claude model, Anthropic, and Elon Musk’s aerospace company, SpaceX, announced a strategic agreement to use the computing capacity of the Colossus 1 data center in Tennessee.
Colossus 1 is one of several data center projects that xAI – an AI startup founded by Musk and absorbed by SpaceX – has undertaken in Tennessee and Mississippi as part of its strategy to expand computing capacity.
The agreement will provide Anthropic with over 300 megawatts of new computing power, equivalent to more than 220,000 Nvidia GPUs, strengthening the infrastructure needed to support the growth of the Claude models and at the same time expanding the company’s network of technological partnerships.
But that’s not all: the two companies also aim to develop several gigawatts of AI computing capacity in orbit in the future. A project dear to Elon Musk, convinced that space computing could overcome the space and cooling limitations of terrestrial data centers.
Moreover, this move comes as SpaceX is working to become a powerhouse in the AI field ahead of the planned IPO, one of the most significant financial debuts in history.
Not only that, the agreement is noteworthy for having united two rivals, notes Quartz. The American billionaire, owner of Tesla and SpaceX, has publicly attacked Anthropic on several occasions, recalls Cnbc.
Not to mention that the agreement with Musk, one of the main backers of President Donald Trump’s campaign, comes despite the recent public dispute between Anthropic and the Pentagon over limits on the military use of its technology, highlights the Financial Times.
All the details.
THE AGREEMENT BETWEEN ANTHROPIC AND SPACEX FOR COMPUTING POWER
Anthropic has announced a partnership with SpaceX to use the entire computing capacity of the Colossus 1 data center. According to the two companies, the new capacity will be available within the month and will directly improve service for Claude Pro and Claude Max subscribers.
The company producing Claude AI will use over 300 megawatts of computing power at SpaceX’s data center in Tennessee, as part of a broader infrastructure expansion strategy necessary to support the growing demand for its artificial intelligence models.
According to Anthropic, the agreement will allow it to restore some of the speed limits it had reduced in recent weeks.
THE LATEST PARTNERSHIPS MADE BY AMODEI’S COMPANY
Also Quartz recalls that the agreement with SpaceX adds to a series of strategic deals recently announced by Anthropic to strengthen its computing capacity. These include an agreement for up to 5 gigawatts with Amazon, which provides for nearly 1 GW of new capacity by the end of 2026, and a contract of similar size with Google and Broadcom operational from 2027.
The company led by Amodei has also partnered with Microsoft and Nvidia for $30 billion in Azure capacity and announced a $50 billion investment in U.S. AI infrastructure with Fluidstack.
SPACEXAI AND BEYOND, MUSK’S AMBITIONS FOR ORBITAL DATA CENTERS
Beyond the Memphis infrastructure, the two companies stated that Anthropic has expressed the intention to reach several gigawatts of computing capacity for artificial intelligence in orbit with SpaceX in the future.
“As part of this agreement, Anthropic has also expressed interest in collaborating on the development of orbital computing capacity for artificial intelligence of several gigawatts,” SpaceX said in a blog post.
Meanwhile, also yesterday Musk stated that xAI would be “dissolved” as a separate company and that its AI products would be renamed SpaceXAI.
Last month SpaceX announced an agreement with Cursor to provide computing capacity to the startup and collaborate on AI tool development, also obtaining the right to acquire the company for $60 billion within the year.
SpaceX explained that “The computing power required to train and operate the next generation of these systems is exceeding terrestrial energy, land, and cooling capacities, especially on the expected timelines.”
The aerospace company also plans to build a large chip production plant called “Terafab.” Last March Musk presented the project, a joint venture between his companies Tesla, SpaceX, and xAI, for the production of chips for artificial intelligence, robotics, and data centers in space. Terafab will therefore be a manufacturing facility located near Austin, Texas, aiming to produce one terawatt of computing power per year, according to Musk.
The facility is expected to produce two types of chips: one optimized for edge computing and inference, and the other a high-power chip designed for space, which could be used by SpaceX and xAI.
MUSK’S STRATEGY
Returning to the recent agreement with Anthropic, regarding SpaceX (SpaceXAI?), the main advantage it seems to derive from the deal is money, observes Gizmodo. Since the company is reportedly planning to go public by the end of the year, it probably needs to strengthen its financials to attract investor interest.
At the same time, the agreement between Musk and Anthropic gains significance also in light of recent tensions in the AI sector. Musk is indeed engaged in a lawsuit against OpenAI and Sam Altman, accused of turning the lab into a for-profit company contrary to the organization’s original mission.
ANTHROPIC IS NO LONGER “EVIL” ACCORDING TO TESLA AND SPACEX’S OWNER
Gizmodo also highlights how Anthropic and SpaceX make an unlikely pair, considering how the CEO of the latter has spoken about the former.
Besides insults, Musk also accused Anthropic of having stolen data on a “massive scale” to train its models. What made Musk change his mind?
Commenting on the agreement, the American billionaire wrote on X (his social platform formerly Twitter): “I spent a lot of time last week with senior members of the Anthropic team to understand what they do to ensure Claude is a benefit to humanity and I was impressed.” He then added: “Nobody triggered my evil detector.”
In the message posted on the social network, Musk also stated: “As long as they engage in critical self-analysis, Claude will probably be a good thing.” And again: “After that, I had no problem renting Colossus 1 to Anthropic, since SpaceXAI had already moved training to Colossus 2.”
Claude is woke and their logo looks like an anal sphincter 😂 https://t.co/5QNItd9HS2
— Elon Musk (@elonmusk) March 7, 2026
