Palantir and Nvidia reportedly restrict use of certain AI models amid data-sourcing concerns

Palantir Technologies and Nvidia have moved to limit how their staff and products use some third-party artificial intelligence models, citing concerns about the provenance of the data used to train them, according to a report by The Information picked up by Reuters.

The reported restrictions reflect a broader challenge facing companies deploying large language models (LLMs) — AI systems trained on vast datasets that can include publicly scraped material whose licensing status remains contested. Vendors working with government and enterprise clients, where scrutiny of data handling is heightened, have particular incentives to be cautious about which models are embedded in their workflows.

Palantir Technologies Inc. (NASDAQ: PLTR) builds software platforms used by government agencies and commercial customers, including its Gotham platform deployed across the intelligence community in the United States, the United Kingdom and internationally. For a company whose contracts depend on strict data-governance standards, the sourcing of training data behind AI components is a material consideration, even as the firm promotes its own AI tooling for large customers.

Nvidia, best known as the dominant supplier of the graphics processing units (GPUs) that power AI training and inference, has been expanding into software and services layers of the AI stack, making its internal policies on model use notable for the wider ecosystem.

The report arrives amid intensifying legal and regulatory attention on training data. Rights holders have filed multiple copyright lawsuits against AI developers, and governments on both sides of the Atlantic are drafting rules around transparency in model training. Corporate policies that restrict which external models employees or products may use are becoming a common risk-management practice in the sector.

Palantir shares were trading at $170.50 in recent activity, marginally below the prior close of $170.62 (down 0.07%), valuing the company at roughly $439.7 billion. The stock sits in the software–infrastructure category of the technology sector, where AI adoption narratives have been a significant driver of valuations over the past two years.

Neither company’s reported policy changes were accompanied by disclosed financial impact. Restrictions on model use typically affect internal development practices rather than immediate revenue, though they can shape product roadmaps in competitive AI markets.

Source: original release

What to watch

• Whether Palantir or Nvidia publicly confirm or detail the reported internal AI model policies.
• Outcomes of ongoing copyright litigation involving AI training data, which could reshape vendor model choices.
• Palantir’s next quarterly earnings report for commentary on AI product demand and governance practices.
• Regulatory developments on AI training-data transparency in the US and UK, where Palantir maintains significant public-sector business.