SpaceX rents its AI computing power to Anthropic for $1.25 billion per month

SpaceX Monetizes Its AI Computing Power: A Historic Deal with Anthropic

SpaceX, Elon Musk's company known for its space ambitions, is revealing an unexpected facet of its operations: the monetization of its own computing infrastructure dedicated to artificial intelligence. In a filing with regulatory authorities, the company details a major cloud services agreement with Anthropic PBC, an AI research and development firm. This agreement, effective from May 2026, involves making significant computing capabilities available, particularly on supercomputers named COLOSSUS and COLOSSUS II. These infrastructures are also used internally by SpaceX for training its own AI models, such as Grok 5, demonstrating their dual use.

This agreement marks a turning point in SpaceX's strategy, capitalizing on its massive investments in computing power. The company has entered into Cloud Services Agreements with Anthropic, a pioneer in the field of AI, recognized for its focus on safety and ethics. The valuation of these computing resources reaches new heights, with a contract stipulating a monthly payment of $1.25 billion by Anthropic. This colossal sum underscores the growing demand for high-performance computing infrastructure, essential for the development and training of the most advanced AI models on the market.

A $1.25 Billion Per Month Contract: AI, SpaceX's New Revenue Engine

The core of this agreement lies in the valuation of the raw computing power that SpaceX has developed. The Cloud Services Agreements stipulate that Anthropic commits to paying SpaceX $1.25 billion per month until May 2029. It is specified that this computing capacity will be gradually made available in May and June 2026, with reduced rates during this initial phase. Flexibility is also integrated into the contract, as both parties retain the option to terminate the agreement with 90 days' notice. This contractual structure reflects an ambitious and potentially evolving business relationship.

The stakes for Anthropic are high: access to massive and high-performance computing infrastructure to accelerate its own research and AI model development. The company, often compared to OpenAI, is investing heavily in creating safer and more ethical AI. Access to SpaceX's supercomputers, particularly COLOSSUS II which hosts the training of Grok 5, represents a major strategic opportunity for Anthropic to remain competitive in an increasingly intense AI race. The scale of the financial investment reflects the criticality of these resources for leading players in artificial intelligence.

COLOSSUS and COLOSSUS II: The Supercomputers at the Heart of SpaceX's AI Strategy

The COLOSSUS and COLOSSUS II systems are not mere server farms. They represent SpaceX's commitment to the field of high-performance computing for AI. The company uses these resources not only for its own projects, such as training the Grok 5 model, but also to offer services to third parties. This ability to handle very large-scale AI workloads is a strategic asset. It allows SpaceX to diversify its revenue streams, moving away from sole reliance on space contracts and satellite launches.

The construction and maintenance of such computing infrastructures require considerable investment, both in hardware and energy. By renting out a portion of this capacity, SpaceX is transforming a potentially high cost center into a substantial revenue source. This business model could become a standard for companies with surplus or specialized computing resources. The agreement with Anthropic could thus be just the prelude to other similar partnerships, positioning SpaceX as a key infrastructure provider for the AI ecosystem.

A Booming AI Infrastructure Market

The agreement between SpaceX and Anthropic is part of a context of exponential demand for AI computing power. Companies worldwide, from startups to tech giants, are struggling to secure access to the GPUs and infrastructure needed to train ever larger and more complex artificial intelligence models. Players like NVIDIA currently dominate this market, but new offerings are emerging. SpaceX's ability to offer an alternative, leveraging its own infrastructure, creates a new competitive dynamic.

For the AI sector, this agreement has several implications. Firstly, it validates the strategy of specialization and monetization of computing infrastructure. Secondly, it highlights the diversification of players like SpaceX into areas adjacent to their historical core business. Finally, it underscores the strategic importance of mastering physical infrastructure – from data centers to communication networks – for future AI advancements. Companies that can combine software expertise with control over their hardware infrastructure will have a clear competitive advantage.

Our Take: SpaceX's Bold AI Strategy

The announcement of this agreement reveals a bold strategy on the part of SpaceX. By transforming its massive investments in computing power into a high-value commercial service, the company is positioning itself as a key player in the AI ecosystem. The challenge for SpaceX is no longer just to conquer space, but also to provide the infrastructure that will power the next generation of artificial intelligence. The ability to secure such a contract, of unprecedented financial scale for computing services, demonstrates the confidence of major players like Anthropic in SpaceX's capabilities.

It will be interesting to follow the evolution of these partnerships. SpaceX's ability to maintain the performance and availability of its COLOSSUS and COLOSSUS II infrastructures, while managing its own computing needs for projects like Grok 5, will be crucial. Furthermore, the 90-day termination clause leaves some flexibility for both parties. This agreement could well be the first in a series, confirming the trend towards specialization and resource pooling in artificial intelligence, a field where raw power has become a determining factor of success.