Enterprise AI Giants Restrict Anthropic's Fable Model Over Data Retention Disputes

Major technology companies including Nvidia, Palantir Technologies, and Booz Allen Hamilton are restricting deployments of Anthropic's Fable model for sensitive work, citing concerns about data retention policies that conflict with enterprise security requirements. The conflict centers on Anthropic's June 2026 introduction of a 30-day data retention policy for Fable 5, which the company said was necessary to detect sophisticated attacks unfolding across multiple user sessions. However, major enterprise customers view the policy as incompatible with handling proprietary business data, even though Anthropic pledged not to use retained logs for model training.