2 Conflicting Reports: Replit CEO Cites Cost for Anthropic Split, Insiders Say Deal Lives

Replit, a cloud-based software development platform, is facing conflicting reports regarding its relationship with AI provider Anthropic. CEO and co-founder Amjad Masad stated on stage at TechCrunch Disrupt 2026 that the company is no longer working with Anthropic. He cited cost, partnership dynamics, and performance metrics as drivers for the decision, noting that OpenAI’s ChatGPT now performs better on internal benchmarks for building applications from scratch.

By the numbers

• 15% of Replit revenue now comes from enterprise contracts, up from 2% last year.
• More than 50% of total revenue is derived from B2B business.
• Replit was previously a top-five customer of Anthropic before this shift.

What the figures mean

The financial data indicates a significant pivot toward enterprise clients, with enterprise revenue jumping from two percent to fifteen percent of the total. Over half of the company’s income now comes from business-to-business contracts. This growth coincides with the strategic shift in AI model providers. Masad emphasized that OpenAI is the preferred partner for powering Replit Agent due to economics and performance. However, he clarified that Anthropic models, including Claude, remain accessible to customers who choose them, even though OpenAI is the new default. This suggests the operational split may be narrower than the initial headline implied.

What should you do?

Organizations using Replit should verify their current model selections immediately. Do not assume the platform has automatically migrated all workloads to OpenAI. Check whether your specific projects or agents are pinned to Anthropic’s Claude models or if they are set to use the default provider. If your compliance or security policies mandate specific model vendors, you must manually confirm that your configuration reflects those requirements. Review any automated scripts that might rely on default settings, as a change in the default could alter behavior or output quality unexpectedly.

Operational impact

When a major supplier changes its default configuration or partnership status, internal teams usually need to audit their dependencies. Check whether critical applications rely on specific model behaviors that may have shifted. Identify which stakeholders, such as engineering leads or compliance officers, need to be informed about potential changes in output quality or vendor liability. Review service level agreements to see if they specify particular model versions or providers. Ensure that logging and monitoring systems can detect performance anomalies that might arise from the new default settings.

What to do and how to stay safe: Replit

• Audit current Replit project settings to confirm which AI model is actively selected.
• Verify that any automated deployment pipelines are not inadvertently switching models.
• Monitor application performance and output quality for deviations after the default change.
• Consult internal security policies to ensure the new default provider meets compliance standards.

Step-by-step guide: Shadow AI Explained: 8 Questions Answered for Secure Operations

General security guidance from the BitShield News newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Did Replit completely stop using Anthropic?

Replit CEO Amjad Masad stated they are no longer working with Anthropic, but later clarified that Anthropic models remain available to users who select them.

Why did Replit switch its default provider?

Masad cited cost, partnership strength, and better performance metrics for OpenAI’s models in building applications from scratch.

Is the Anthropic partnership still active according to insiders?

Yes, one insider familiar with the partnership stated that the deal is still in place, contradicting the CEO’s initial announcement.