Anthropic has signed a strategic agreement with SpaceX to enhance its computing capabilities, a move aimed at meeting the increasing demand for the AI model Claude and expanding the platform's ability to serve users.

Under the agreement, Anthropic will rely on more than 300 megawatts of computing capacity within SpaceX's massive data center in Memphis, known as Colossus 1, giving it access to over 220,000 Nvidia graphics processing units.

Addressing Claude's Limitations

Through this partnership, Anthropic seeks to address one of the most prominent issues faced by Claude users in recent months, following repeated complaints from paid subscribers about usage restrictions, which forced them to pause after a limited number of tasks or pay additional fees to continue working.

The company confirmed that the new capacity will allow it to significantly raise usage limits, especially for Claude Code subscribers, its AI-powered programming platform, with doubled usage limits for paid subscribers.

Anthropic also announced the removal of restrictions imposed during peak hours, along with a substantial increase in usage limits for Claude Opus models via API interfaces.

Ensuring AI Does Not Harm Humanity

This deal brings together two prominent companies in the AI race, as Anthropic is now one of the strongest competitors to OpenAI, while Elon Musk, CEO of SpaceX, continues his legal battle against OpenAI and its board over the company's shift to profitability, contrary to its original founding purpose.

Commenting on the agreement, Tom Brown, co-founder of Anthropic, said the company will need to "move enormous amounts of atoms" to keep up with the accelerating demand for AI, considering SpaceX as the best at executing this kind of rapid expansion.

For his part, Elon Musk said he spent time with senior officials at Anthropic to understand how Claude works, adding that he left with a positive impression of the team, describing it as highly efficient and focused on building systems beneficial to humanity.

Musk pointed out that SpaceX will retain the right to reclaim computing resources if it turns out that the AI system is used in activities that could harm humanity, indicating that the partnership is not separate from safety and long-term oversight considerations.