Anthropic has signed a massive long-term infrastructure agreement with SpaceX as competition in the AI industry continues to intensify. According to details revealed in recent securities filings, Anthropic is expected to pay nearly $45 billion to Elon Musk’s SpaceX over the next three years for computing capacity tied to its Claude AI systems. The agreement gives Anthropic access to large-scale infrastructure needed to train and operate advanced AI models. Payments are expected to reach around $1.25 billion per month through May 2029, though some early capacity ramp-up periods in 2026 will reportedly operate at lower rates. SpaceX disclosed the financial details as part of paperwork connected to its upcoming initial public offering.
The scale of the deal immediately drew attention because of how large it is compared to SpaceX’s current business size. The company reportedly generates around $18 billion annually, meaning the Anthropic agreement alone could eventually represent revenue far beyond its present yearly totals. Earlier this month, Anthropic confirmed a separate agreement for access to more than 300 megawatts of compute power from SpaceX’s Memphis-based AI infrastructure project known as Colossus 1. At the time, financial terms were not shared publicly.
Anthropic later expanded the partnership to include additional capacity at another SpaceX data center. The update was mentioned by Anthropic co-founder and chief compute officer Tom Brown. The partnership highlights how quickly AI companies are racing to secure computing resources. Training modern AI models now requires enormous amounts of hardware, electricity, cooling systems, and networking infrastructure. Demand has risen so rapidly that compute access is becoming one of the industry’s biggest bottlenecks.
Anthropic has faced growing pressure from rising Claude usage across coding, business automation, and enterprise AI tools. The company’s reported revenue run rate climbed sharply over the past year as more businesses adopted Claude for daily workflows. That growth created major infrastructure needs. Colossus 1 reportedly contains more than 220,000 NVIDIA GPUs, making it one of the largest AI computing clusters currently operating.
The agreement also reflects SpaceX’s growing involvement in artificial intelligence infrastructure after the company merged with Musk’s xAI earlier this year. Since then, Musk’s broader strategy appears increasingly focused on combining AI systems with large-scale compute operations and energy infrastructure.
The relationship between Anthropic and SpaceX is especially notable because the companies are connected to competing forces inside the AI industry. Anthropic continues competing directly with OpenAI, Google DeepMind, and xAI itself in the race to develop more advanced language models and AI agents. At the same time, Musk’s relationship with parts of the AI industry remains complicated. He recently lost a major legal battle connected to OpenAI and Sam Altman after a jury ruled that he waited too long to bring parts of the case forward. The ruling added to a growing list of legal disputes involving Musk and AI-related companies. Despite competitive tensions, the infrastructure deal shows how critical computing power has become in modern AI development. For companies building frontier AI systems, access to GPUs and large-scale data centers is now just as important as the models themselves.
