Skip to content
AI IntelligenceSep 16, 2026AI Intelligence
Article

Nvidia and Palantir reportedly scale back Anthropic model use over data logs

Despite assurances from OpenAI and Anthropic that customer data is not used for training, AI companies continue to face data trust issues. Palantir, Nvidia, and Booz Allen Hamilton reportedly scaled back their use of Anthropic's Fable model after learning it stores usage logs for 30 days.

Frontier EditorialSource: The Decoder
01

Source Brief

Nvidia and Palantir reportedly scale back Anthropic model use over data logs: Despite assurances from OpenAI and Anthropic that customer data is not used for training, AI companies continue to face data trust issues. Palantir, Nvidia, and Booz Allen Hamilton reportedly scaled back their use of Anthropic's Fable model after learning it stores usage logs for 30 days.