Replit, the coding platform that has become a darling of startup founders and developers, has made a significant shift in its AI partnerships. The company, co-founded by Amjad Masad, has switched its default AI model from Anthropic to OpenAI's ChatGPT. This move, driven by cost, performance and partnership considerations, will likely see a reduction in Replit’s spending with Anthropic, a company that had previously been one of its top customers.

While Masad clarified that customers can still access Anthropic models, the default model shift could have significant financial implications for Anthropic. Prior to this move, Replit had been heavily reliant on Anthropic’s Claude models for its platform, with a partnership that had significantly boosted its revenue. Recent blog posts and social media activity, however, suggest that Replit was still promoting Anthropic products as recently as a few weeks ago.

During a TechCrunch Disrupt appearance, Masad also revealed insights into Replit's growth strategy, noting a significant shift towards enterprise contracts. He stated that enterprise revenue now makes up more than 15% of Replit’s revenue, with over 50% of its revenue coming from B2B sales. He also highlighted the importance of good AI safety measures, likening them to the necessary environment for human productivity.

This shift in partnerships comes at a time when the AI landscape is rapidly evolving. The decision by Replit could signal a broader trend, as companies navigate the complex and often competitive ecosystem of AI providers. Whether this will lead to a more unified approach to AI or simply a reshuffling of partnerships remains to be seen.