Anthropic nears first quarterly profit, agrees to pay SpaceX $1.25 billion monthly for computing power

Anthropic, the San Francisco-based AI startup, is on track to achieve its first quarterly operating profit in the June quarter, according to a person familiar with the matter. The company’s revenue is expected to reach at least $10.9 billion, more than double the $4.8 billion recorded in the March quarter, with an anticipated operating profit of $559 million. This surge reflects growing demand for its Claude AI, used by developers for programming tasks, and its Mythos model, which enterprises deploy to identify code vulnerabilities. The financial projections highlight Anthropic’s ability to offset the high costs of AI development and deployment, a challenge faced by many in the industry. To meet its computing needs, Anthropic has secured a deal with SpaceX, agreeing to pay $1.25 billion per month through May 2029 for access to SpaceX’s AI training data centers, Colossus and Colossus II. The agreement allows either party to terminate with 90 days’ notice, with fees adjusted during the ramp-up phase in May and June. SpaceX’s IPO filing revealed the deal as part of its broader strategy to offer AI compute services at scale, despite its own AI segment reporting a $2.5 billion operational loss in the March quarter on $818 million in revenue. Elon Musk, SpaceX’s CEO, announced on X that the company is in discussions with other firms about expanding this service, signaling a potential shift toward profitability in its AI operations. The partnership underscores the competitive landscape in AI infrastructure, where companies like Anthropic rely on external computing power to sustain growth. While Anthropic’s financial outlook improves, SpaceX’s AI segment remains unprofitable, creating a dynamic where both entities depend on scaling their operations to justify the massive investments required for AI development.