Orbital Ambitions: Anthropic Taps Musk’s “Colossus” to Double Claude’s Power and Eye the Stars

According to the latest intelligence, Anthropic has secured a landmark partnership agreement with Elon Musk’s SpaceX, granting it access to the colossal computational reservoirs housed within the Colossus 1 data center. Bolstered by this infusion of processing power, Anthropic has announced an immediate doubling of the usage quotas for Claude Code across its premium tiers, alongside a significant relaxation of various operational constraints.

Under the terms of this accord, Anthropic has been authorized to utilize the “entire computational capacity” of the SpaceX Colossus 1 facility this month, an infusion estimated to yield over 300 megawatts (MW) of novel power. More arresting, however, is the shared vision for the future; Anthropic has signaled a profound interest in further collaboration to pioneer “gigawatt-scale orbital AI computing.” When juxtaposed with SpaceX’s ambitious proposal to the FCC—which envisions the deployment of one million satellites to construct an “orbital data center”—this move ventures into the realm of speculative science fiction.

Elon Musk clarified that while the resources of Colossus 1 have been dedicated to Anthropic, the training operations for xAI—recently merged with SpaceX—have been seamlessly transitioned to the Colossus 2 facility, thereby ensuring no disruption to internal resource allocation.

Leveraging this vast computational backbone, Anthropic has swiftly extended the benefits to its user base. For subscribers of the Pro, Max, Team, and Enterprise plans, the five-hour rate limits for Claude Code will be doubled instantaneously. Furthermore, Pro and Max users will no longer be subjected to the frustrations of “peak hour” throttling. The API invocation limits for the flagship Claude Opus model have also been augmented to a “considerable degree,” offering a decisive advantage to development teams reliant on the model for intensive coding and complex reasoning.

The alliance has elicited significant industry astonishment, given the historical friction between the two parties. Not long ago, Musk characterized Anthropic as “misanthropic and evil” and maintained a scathing critique of CEO Dario Amodei. However, Musk’s stance on the X platform has undergone a dramatic softening; he recently expressed that he was “impressed” by the Anthropic team and conceded that Claude “should perform well.”

Nevertheless, maintaining his characteristic brand of eccentric oversight, Musk appended a singular caveat in a subsequent post: “We reserve the right to reclaim this compute should their AI engage in activities inimical to humanity.”