A huge agreement is coming to light that will completely change the stability in the world of artificial intelligence. According to the official documents published by Elon Musk’s space company SpaceX during the public offering process, the well-known artificial intelligence developer Anthropic is paying an incredible amount to use SpaceX’s “Colossus” information centers in Memphis.

In order to reach the high process capacity offered by SpaceX, Anthropic agrees to pay the company $1.25 billion per month, i.e. a total of $15 billion annually. This multi-billion dollar agreement, which will last until May 2029, clearly reveals how desperately technology giants need hardware power in the artificial intelligence race and how large budgets they allocate for this purpose.

SpaceX’s IPO Document Leaked All Details

SpaceX, which recently gained strength by merging with xAI, shares striking financial data in the S-1 documents submitted to the US Securities and Exchange Commission (SEC) as a module of the public offering process.

According to this detailed report, which exceeds the expectations of analysts in the industry, Anthropic’s annual payment of $15 billion is almost double SpaceX’s total revenue of $18.7 billion through 2025. In the first months of the agreement, a special discounted pricing is applied to Anthropic as the capacity increase process of the information center continues.

However, when the systems reach full capacity, Anthropic continues to train its artificial intelligence models called Claude much faster by taking full advantage of this huge processing power, which contains approximately 220 thousand GPUs.

A Strategic Partnership in the Midst of Competition

The Claude model developed by Anthropic is actually in direct competition with xAI’s Grok model in terms of market share and user base. On the other hand, the fact that Anthropic allocates such a high budget to Elon Musk’s company, one of its biggest Competitors, creates great surprise in the technology world.

The cancellation clause added between the parties, which allows the parties to terminate the agreement within 90 days, provides companies with the flexibility they need in the rapidly changing artificial intelligence market.

While local governments’ objections to the construction of huge data centers for environmental reasons are increasing, it is clearly seen how large companies such as Anthropic are having difficulty finding server capacity and therefore are negotiating an agreement even with their Competitors.

The Era of Artificial Intelligence Data Centers in Space Begins

In his latest statement on the social media platform

The company aims not only to establish giant information centers on Earth, but also to establish giant orbital information centers in space in the future. These orbital information centers, which will work with solar power and cool in the natural vacuum environment of space, are expected to offer a radical solution to the power consumption and overheating problems experienced by existing land-based facilities.

Financial Numbers Are Astounding

SpaceX’s capital expenditures in the field of artificial intelligence are reaching quite remarkable levels. The company spends $12.7 billion on artificial intelligence infrastructure in 2025 alone, and this number accounts for 61 percent of total capital expenditures.

On the Anthropic front, things are going pretty well financially. The company is expected to reach operating profit and quarterly sales revenue to exceed $10.9 billion. Considering the growth momentum of both companies and their positions in the market, it can be easily said that the need for artificial intelligence process power will increase exponentially in the coming years and this market will turn into a trillion-dollar ecosystem.