In a significant move within the competitive artificial intelligence sector, Anthropic has terminated access to its Claude AI models for developers at Elon Musk's AI startup, xAI. The decision was communicated internally at xAI this week, highlighting escalating tensions between leading AI labs.
Internal Email Reveals Sudden Block
The restriction came to light through an internal email sent by xAI cofounder Tony Wu to his team. On Wednesday, Wu informed the team that Anthropic's models had stopped responding on Cursor, a popular AI-powered coding tool widely used by developers. He attributed this to a new policy being enforced by Anthropic against its major competitors.
Wu acknowledged the immediate impact on the team's workflow, stating that the block would affect productivity. However, he framed it as a motivating factor, pushing xAI to accelerate the development of its own proprietary coding products and AI models.
Violation of Competitive Terms of Service
The action by Anthropic is firmly rooted in its commercial terms of service. These terms explicitly prohibit customers from using its services, including the Claude models, to "build a competing product or service" or to "train competing AI models." While Cursor is a legitimate development platform, xAI's specific use of Claude through it to advance its own AI lab's work was deemed a violation.
This is not an isolated incident for Anthropic. The company has a history of enforcing these terms against rivals. In August of last year, Anthropic revoked API access for OpenAI under similar circumstances. More recently, in June 2025, it limited access for the Windsurf coding environment following reports of a potential sale to OpenAI.
Broader Crackdown on Third-Party Access
The block on xAI coincides with a wider enforcement effort by Anthropic. The company is simultaneously tightening technical safeguards against third-party tools that spoof its official Claude Code client. Thariq Shihipar, an Anthropic employee working on Claude Code, confirmed the company had strengthened protections against such spoofing, which allowed access at consumer subscription rates instead of the higher, metered API pricing.
The timing of these actions is particularly noteworthy. Elon Musk recently announced that a "major upgrade to Grok Code" is expected next month, promising superior performance on complex coding tasks. In response to the ban, Nikita Bier, head of X, suggested retaliating by banning Anthropic from the X platform.
This development underscores the fiercely competitive and protective nature of the current AI landscape, where companies are aggressively guarding their technological advantages while racing to innovate.
