Anthropic Co-founder and Chief Science Officer Jared Kaplan recently explained why the company terminated Windsurf’s direct access to its AI models, Claude 3.5 Sonnet and Claude 3.7 Sonnet. This decision came amidst rumors that OpenAI, Anthropic’s biggest competitor, might be acquiring Windsurf for $3 billion. During an interview at TechCrunch's TC Sessions: AI 2025 on Thursday, Kaplan stated that the move was aimed at fostering sustainable long-term relationships with future partners. "It would be odd for us to be selling Claude to OpenAI," Kaplan remarked, emphasizing the company's commitment to its own ecosystem. Anthropic’s decision to cut ties with Windsurf was also driven by its current computing limitations. The company wants to reserve its computational resources for what it considers "lasting partnerships." Windsurf, which has now been forced to seek third-party computing providers, expressed disappointment over the decision. They noted that it could lead to short-term instability for users trying to access Claude through their platform. Neither Windsurf nor OpenAI have officially commented on the acquisition rumors, and an OpenAI spokesperson did not respond to TechCrunch’s inquiries. Kaplan, however, reassured the audience that Anthropic is working to expand the availability of its models for users and developers in the near future. The company has begun leveraging a new and scalable computing cluster from Amazon, which will significantly boost its capacity. This new resource will help Anthropic support a broader range of clients and projects. In the meantime, Kaplan highlighted that Anthropic is collaborating with other companies focusing on AI development, such as Cursor. He emphasized that Anthropic sees Cursor as a long-term collaborator rather than a rival, despite the latter’s efforts to build its own AI models. Kaplan further indicated that Anthropic is shifting its focus toward developing its own specialized coding products, like Claude Code, instead of primarily concentrating on AI chatbot experiences. He believes that the chatbot paradigm, while currently popular, is static and limiting. In contrast, AI agents, which can interact more dynamically and contextually, hold greater potential for assisting users in the future. This strategic shift reflects a broader trend in the AI industry, where companies are moving beyond simple conversational interfaces to create more sophisticated and versatile tools. Kaplan's comments suggest that Anthropic is positioning itself to stay ahead in the evolving landscape of AI applications, particularly in coding and software development.