Nvidia, Palantir and Booz Allen restrict Claude over data fears

Three large Anthropic customers are putting limits on how its models touch their data.

What the report says

The Information reported on September 14, 2026, according to Reuters, that Nvidia, Palantir and Booz Allen are restricting their use of Anthropic’s models over concerns about their proprietary data.

The restrictions differ by company. Nvidia limits Claude to less sensitive work. Palantir wants a zero-data-retention guarantee before it offers Anthropic’s models through its software. Booz Allen barred staff from running Anthropic’s commercial model on cybersecurity projects that touch proprietary data.

Anthropic’s side

The restrictions follow a 30-day data retention policy Anthropic introduced in June with Fable 5. On September 1, Anthropic announced Enterprise Frontier Safeguards, which let enterprise customers keep monitoring data in their own cloud accounts under their own encryption keys, with no human review by Anthropic employees required. The system is rolling out in phases starting later this fall, and the report does not say whether it satisfies the three companies.

What it shows

The report names specific work: Nvidia’s sensitive internal tasks, the software through which Palantir offers AI models, and Booz Allen’s cybersecurity projects that touch proprietary data. According to Reuters, the concern is that model providers could retain or extract value from sensitive corporate information.

For enterprise buyers, the question is shifting from which model performs best to what happens to the data sent to it.