Palantir, Nvidia, Booz Allen Restrict Anthropic and OpenAI Models Over Data Concerns
Palantir Technologies (PLTR), Nvidia (NVDA), and Booz Allen Hamilton (BAH) are restricting or threatening to drop AI models from Anthropic and OpenAI due to concerns over data retention policies. The companies fear that sensitive corporate information could be retained or used by the model providers, potentially violating privacy standards. Palantir is demanding guarantees of zero data retention before offering access to Anthropic's models through its platform. Nvidia has limited Anthropic’s use within internal tasks and prefers its own Nemotron models for proprietary work. Booz Allen Hamilton has banned staff from using Anthropic’s commercial AI on cybersecurity projects involving sensitive data. The pushback intensified after Anthropic introduced a 30-day data retention policy in June, which it claimed was necessary to detect sophisticated attacks across multiple sessions. While the company stated that retained data would not be used for model training, enterprise customers—especially those in regulated industries—raised concerns. In response, Anthropic launched Enterprise Frontier Safeguards, allowing clients to store activity logs on their own cloud infrastructure with encryption keys under their control. Nvidia and Palantir recently announced a joint platform using Nvidia’s Nemotron models within Palantir software, enabling organizations to run AI without exposing data externally. Meanwhile, Microsoft is capitalizing on enterprise customer concerns by promoting its segmented cloud solutions.
Tickers: $BAH $MSFT $NVDA $PLTR
Topics: AI models Anthropic data privacy enterprise AI usage OpenAI