
The recent deal was not directly between Anthropic and Elon Musk personally, but between Anthropic and Musk’s AI-computing infrastructure business, now operating through SpaceX/xAI (often referred to as SpaceXAI in recent reporting).
The agreement gives Anthropic access to the full capacity of the Colossus 1 AI supercomputer facility in Memphis, Tennessee. The facility reportedly contains more than 220,000 NVIDIA GPUs and roughly 300 megawatts of AI computing power. Anthropic intends to use that capacity to expand and accelerate its Claude AI services.
Key reported details include:
- Anthropic gains access to the entire Colossus 1 compute cluster.
- The deal was announced in early May 2026 amid surging demand for Anthropic’s Claude models and coding tools.
- Later financial disclosures suggested the arrangement could be worth as much as $1.25 billion per month and potentially tens of billions of dollars over several years if maintained.
- Elon Musk subsequently clarified that SpaceX only committed to a 180-day (six-month) lease initially, with both parties having cancellation rights, rather than an unconditional long-term commitment.
What makes the deal noteworthy is the political and corporate context. Musk had previously criticized Anthropic publicly, yet the shortage of AI computing resources has become so severe that competitors are increasingly doing business with one another. Several commentators described the arrangement as a striking example of strategic necessity overriding rivalry in the AI race.
There is also an intriguing secondary aspect: reports indicate Anthropic expressed interest in exploring future space-based or orbital data-center infrastructure with SpaceX, though that appears to be exploratory rather than a firm project commitment.
In short, Anthropic’s deal is essentially a massive AI-compute leasing agreement: Anthropic gets access to one of the world’s largest AI supercomputers, while Musk’s infrastructure business earns substantial revenue from otherwise available computing capacity
