Investing.com -- Escalating concerns over corporate data privacy are driving major enterprise and defense clients to restrict their use of frontier artificial intelligence models, according to reporting from The Information. The pushback highlights a growing friction point for AI developers such as Anthropic (NASDAQ: ANTP) and OpenAI (NASDAQ: OAI), as prominent clients, including NVIDIA Corporation (NASDAQ: NVDA), Palantir Technologies Inc (NASDAQ: PLTR), and Booz Allen Hamilton Holding (NYSE: BAH), seek stronger data protection guarantees or pivot toward isolated cloud infrastructure to safeguard their intellectual property.

At the center of the dispute is Anthropic’s policy shift in June regarding its flagship Fable model, which introduced a rolling 30-day data retention requirement to prevent malicious misuse of its systems. While Anthropic has since offered an option allowing select customers to store usage logs on their own servers, the lack of irrevocable zero-data-retention (ZDR) guarantees has prompted key corporate clients to scale back deployment. Nvidia has limited its use of Fable to less sensitive projects like open-source software, relying instead on its proprietary Nemotron models for core operations such as supply chain monitoring.

The widening corporate anxiety is creating strategic opportunities for cloud infrastructure providers like Microsoft Corporation (NASDAQ: MSFT), which is aggressively pitching private, air-gapped server environments that keep customer data entirely isolated from external AI developers. Other large enterprises are taking similar precautions; utility operators and pharmaceutical leader Novo Nordisk maintain strict boundaries against feeding proprietary operational data into commercial models, while aerospace contractor Northrop Grumman relies primarily on open-source models deployed on isolated internal networks.

Despite these headwinds, both Anthropic and OpenAI reiterate that standard enterprise contracts do not allow customer inputs or outputs to be used for general model training by default. However, ambiguous disclosures surrounding technical metadata and usage analytics continue to leave enterprise customers seeking absolute clarity. As frontier AI labs increasingly roll out commercial software that directly overlaps with their clients' core operations, the battle over data retention signals that control over proprietary intellectual property will remain a central catalyst in enterprise AI adoption.

Against that backdrop, market anxiety triggered a broad sell-off across AI infrastructure and semiconductor equities as investors weighed potential friction in enterprise adoption. Major chipmakers felt immediate pressure, with Nvidia declining 3.3% and Advanced Micro Devices falling 6.2%, while broader hardware and networking suppliers suffered steeper drawdowns, including Coherent (-11%), GE Vernova (-9.2%), Lumentum (-9%), Vertiv (-8.6%), Applied Optoelectronics (-8%), Eaton (-7.6%), Marvell (-7%), SanDisk (-6%), Micron (-6%), Broadcom (-4.7%), and Digital Realty (-3.4%).