Anthropic is quietly limiting access to its most powerful AI models for certain tasks while launching competing products, sparking backlash from customers who fear they'll be squeezed out.

What happened: Anthropic released Claude Fable 5 this week but throttled its performance when customers used it for tasks related to building their own AI software or hardware. After public backlash, the company said it would at least notify customers when a weaker model is being served. The company also asked partners like Figma and Canva to collaborate on AI features weeks before launching Claude Design—a competing product that eventually pulled Figma out of the partnership.

Why it matters: Anthropic's annualized revenue grew fivefold in just five months to nearly $50 billion（約8兆円）, now exceeding OpenAI, which faces the same tension between being a platform and a competitor. Customers like Cursor and Amp are now rivals to Claude Code and Claude Design, yet depend on Anthropic's models. As Martin Casado, an Andreessen Horowitz partner and Cursor investor, wrote: "It's only a matter of time before only the model creators have access to the most powerful models. Everyone else gets access to smaller, distilled versions." This dynamic echoes earlier cases where Microsoft and Google used platform dominance to favor their own apps and squeeze out partners—both were later found to be illegal monopolies.

What to watch: Anthropic's throttling officially aims to stop foreign adversaries or rival AI labs from improving their own tech, but developers worry even basic features for AI-powered apps could get restricted. Figma CEO Dylan Field said Anthropic was "not consistently candid in their communications," suggesting the conflict between model provider and customer-turned-competitor may deepen as Anthropic and OpenAI grow faster than the next 32 large AI startups combined.