Anthropic blocks xAI from using Claude, deepening rivalry in the AI race

Anthropic has stopped Elon Musk–owned xAI from using its Claude artificial intelligence models, marking a fresh escalation in the growing rivalry between leading AI firms. The decision was communicated internally at xAI this week, after its engineers found that Claude was no longer accessible through Cursor, a widely used AI-powered coding platform.

In an internal email to staff, xAI cofounder Tony Wu said the shutdown appeared to stem from a new policy being enforced by Anthropic across its major competitors. He acknowledged that the loss of Claude would affect the company’s development work but framed it as a push towards building more in-house tools. “We will take a hit on productivity, but it really forces us to develop our own coding products and models,” Wu wrote, underlining how central AI systems have become to xAI’s workflow.

Also read: Grok NSFW backlash reignites debate over how far AI chatbots should go

The move is believed to be linked to Anthropic’s commercial terms, which prohibit customers from using Claude to develop competing products or to train rival AI systems. While Cursor itself is a legitimate software platform, xAI’s use of Claude through it to assist in building its own AI technology appears to have triggered enforcement of those rules.

The decision prompted a sharp reaction from X executives. Nikita Bier, who leads product operations at X, wrote online that it was “time to ban Anthropic from X,” reflecting the growing tension between the two sides.

Anthropic has previously taken similar steps against rivals. Last year it withdrew API access from OpenAI, and earlier this year it limited the AI coding service Windsurf after reports emerged of a potential acquisition by OpenAI. More recently, the company has also tightened technical controls to prevent third-party tools from imitating its own Claude Code client in order to access its models at discounted rates.

Also read: Apple’s AI-powered Siri edges closer to launch after year-long delay

Engineers at Anthropic have said new safeguards were introduced after some users were found bypassing pricing and security systems, triggering account bans and stricter enforcement of usage policies.

The clash comes amid wider changes in Musk’s AI and social media empire. Over the weekend, Musk announced that X will soon open-source its latest recommendation algorithm, including the code used to determine which organic and advertising posts are shown to users. He said the company will release updates to the algorithm every four weeks, alongside detailed technical notes.

As competition intensifies across the AI sector, the breakdown between Anthropic and xAI highlights how access to powerful models is becoming both a technical and strategic battleground.