Replit, the popular AI coding platform utilized by developers and startup founders, announced a shift in its strategic alliances. Amjad Masad, the company’s CEO and co-founder, declared on Thursday, October 16, 2026, while speaking at TechCrunch Disrupt that Replit was ceasing its collaboration with Anthropic.
Reasons for the Strategic Change
During his address, Masad attributed the decision to a combination of factors, including cost considerations, performance metrics, and overall partnership alignment. He noted that Replit had recently established a partnership with OpenAI in August.
Masad stated that the decision was based on comprehensive evaluations, adding that the company utilized an internal benchmark system to assess performance. He elaborated that, according to their internal metrics, “Sol and Astra perform better on building vibe-coded applications from scratch.” He summarized the rationale by stating, “It was a decision based on both cost, partnership, and performance. And for all these reasons, we just felt like they were a better partner for now.”
Background and Conflicting Reports
Previously, Replit had integrated the AI lab’s Claude models into its cloud-based software development platform, a move that sources indicated substantially boosted the company’s revenue. This platform allows users to develop and deploy applications directly within their web browser.
However, Masad’s public announcement was disputed by individuals familiar with the relationship between the two companies. One source, who requested anonymity, indicated that the partnership between Anthropic and Replit remained active.
Furthermore, evidence gathered from Replit’s own public channels, including its website, blog, and social media, suggested continued use of Anthropic products in the weeks preceding the announcement. For instance, a blog entry published in late September cited Anthropic’s Claude Fable 5 and Claude Fable 5.1 as models operating within the Replit Agent’s production environment. Additionally, an employee of Replit posted on X around the same timeframe, promoting the capabilities of Anthropic’s flagship model, Opus 5.5.
Company Growth and AI Safety Focus
Beyond the partnership news, Masad provided insights into Replit’s financial trajectory and the necessity of robust AI safety protocols. He highlighted the increasing contribution of enterprise clients to the company’s revenue.
According to Masad, the company’s business model is showing strong growth in the B2B sector. He specified that while he was not disclosing Annual Recurring Revenue (ARR) figures, enterprise contracts accounted for 2% of the revenue the previous year, rising significantly to 15% of revenue this year. Moreover, he stated that over 50% of the company’s revenue derived from B2B contracts.
Regarding the future of AI development, Masad stressed that sophisticated AI agents require a secure “habitat” to operate effectively, drawing a comparison to the necessary working environment for humans. He outlined that this required habitat must incorporate strong orchestration, robust error recovery, comprehensive safety and security measures, sandboxing capabilities, dependable networking, and proper identity and authorization protocols to ensure safe deployment of AI technology.
