Replit Shifts AI Partnership
Replit, the AI coding platform popular among developers and startup founders, has shifted away from using Anthropic as its primary partner, according to CEO and co-founder Amjad Masad.
Masad stated that the decision involved multiple factors, including cost and performance. Although Replit announced a partnership with OpenAI in August, the extent of the shift away from Anthropic had not been previously clarified.
Masad highlighted internal benchmarks and application building performance as key drivers for selecting OpenAI as a preferred partner for the time being.
Model Availability and Platform Default
Despite the change, Anthropic models remain accessible on Replit's cloud-based software development platform, which allows users to build and deploy applications directly from their browsers.
OpenAI's ChatGPT now serves as the default option on the platform. Because most users retain default settings rather than manually selecting a model, this change is expected to decrease Replit's spending with Anthropic.
Previously, Replit relied primarily on Anthropic’s Claude models, a relationship that contributed to significant revenue growth and positioned Replit as one of Anthropic's top customers.
Masad emphasized that Replit maintains a model-agnostic approach, allowing users to choose Anthropic Claude or other supported models, while designating OpenAI as the preferred partner for powering Replit Agent.
Conflicting Reports on Partnership Status
Some industry insiders familiar with the partnership disputed aspects of the announcement, maintaining that the relationship with Anthropic remained active.
Recent company blog posts and employee social media statements from late September had actively promoted Anthropic models running within Replit Agent environments.
Business Growth and Enterprise Revenue
Alongside updates regarding AI partnerships, Masad shared details regarding the company's financial growth, highlighting an expansion in enterprise contracts.
Enterprise contracts grew from approximately 2% of revenue last year to an estimated 15% this year, with B2B transactions accounting for more than half of total revenue.
The company projects that enterprise and government customers will generate over half of Replit's total revenue by 2027.
Infrastructure and AI Safety
Masad also emphasized the importance of robust safety measures and proper operational environments for deploying artificial intelligence agents securely.
Proper orchestration, error recovery, sandboxing, and security protocols are essential elements required to safely deploy AI agents into production environments.
This article was originally published on October 16 at 2:22 pm PT. It has been updated to include more context around Replit’s partnerships with Anthropic and OpenAI.
