Anthropic has cut off xAI’s access to Claude AI models via the Cursor tool, enforcing a policy that blocks competitors from using Claude to build rival products.

An email from xAI cofounder Tony Wu confirms the ban on xAI’s access to Claude through Cursor, signaling a direct access restriction.

Anthropic’s terms prohibit using Claude to train competing AI or create rival offerings, and the restriction appears linked to xAI’s use of Claude through Cursor for development work.

Market context suggests access controls by AI firms are shaping investments and competitive dynamics in the AI race.

Coinciding with the main story, Elon Musk announced X will open-source its recommendation algorithm and release code and developer notes every four weeks.

Kylie Robison is cited as the reporter highlighting Anthropic’s policy change.

xAI warned staff to expect a short-term dip in coding speed, framing the move as a push to accelerate internal tool development and reliance on its own models.

StartupNews.fyi notes potential conflicts of interest and mentions site upgrades and a contact email for glitches.

The development underscores intensified competition between Anthropic and xAI as both race to improve speed and quality of large AI models.

Anthropic’s stance signals it will not support direct rivals even through third-party tools, signaling tighter access controls in the AI ecosystem.

The situation fits into the broader shift from open access to more restricted, strategic control over tools and data in the AI race.

Anthropic has strengthened technical measures to prevent spoofing or abuse of Claude Code by third-party tools and to clamp down on access at cheaper rates.