Anthropic Introduces ‘Enterprise Frontier Safeguards’ for Enhanced Data Control
Responding to persistent demands from corporate clients, Anthropic has launched a comprehensive replacement for its contentious data retention policy. On September 1, 2026, the AI company unveiled Enterprise Frontier Safeguards—a revolutionary system that provides businesses with increased control over the storage, review, and management of their sensitive data when utilizing Anthropic’s most advanced AI models.
The original policy, which was rolled out in June alongside the release of Claude Fable 5 and Mythos 5, mandated a 30-day retention of all model traffic. This policy was met with severe backlash from enterprise customers in regulated sectors such as finance, healthcare, and law.
With the new system, enterprise customers now have the option to store data in their own cloud infrastructure—including Amazon S3, Azure Blob Storage, or Google Cloud Storage—under their own encryption keys and access controls. Importantly, automated safety monitoring will persist, but without the necessity for human review by Anthropic employees.
The fresh system will be accessible across Claude Code, Claude Enterprise, Amazon Bedrock, Google’s Agent Platform, and Microsoft Foundry. The rollout will occur in phases, with wider availability aimed for fall 2026. Anthropic has confirmed that there will be no separate charges for Enterprise Frontier Safeguards.
Kate Jensen, Head of Americas at Anthropic, highlighted that the company dedicated hundreds of hours collaborating with over 100 customers to develop this solution.
Source: CNBC