Palantir Takes Aim at Usage-Based Pricing Models in Enterprise AI

Palantir Technologies (NASDAQ: PLTR) is publicly pushing back against the prevailing “pay-per-token” pricing model that has come to define much of the commercial artificial intelligence market, according to recent reporting by Benzinga. The company’s stance signals a different philosophy about how AI software should be priced and delivered to enterprise and government customers.

Under the token-based model, organizations are charged for each unit of text processed by a large language model, a structure that can make costs unpredictable as AI usage scales. Palantir’s critique suggests the company sees this approach as a friction point for customers who want more predictable economics when deploying AI across their operations — particularly in regulated and mission-critical environments where budgeting certainty matters.

The Denver-headquartered company has built its business on long-term software deployments for government agencies and large enterprises, including platforms used by the intelligence community. That installed base gives it a vantage point different from AI infrastructure providers whose revenue depends directly on metered model usage. Palantir describes itself as focused on integrating AI into existing enterprise workflows rather than selling raw model access.

Market Context

Shares of Palantir were trading at $170.87, down 2.04% from the previous close of $174.43, giving the software and infrastructure company a market capitalization of roughly $439.7 billion. The stock’s valuation reflects the market’s elevated expectations for AI-linked software providers, which makes debates over pricing models more than academic — how vendors charge for AI can shape margins, contract structures, and customer retention across the sector.

The broader industry remains split. Hyperscale cloud providers and model developers have largely embraced usage-based billing, while a number of application-layer software firms have favored subscription or outcome-based arrangements. Palantir’s position places it firmly in the latter camp, arguing that predictability serves customers better as AI becomes embedded in everyday operations.

What to watch

• Upcoming quarterly earnings, including commercial customer count and U.S. government revenue figures
• Any new contract announcements or renewals that reveal how Palantir structures AI pricing in practice
• Competitor responses, as other enterprise software vendors weigh usage-based versus subscription AI billing
• Guidance updates that indicate how pricing philosophy affects revenue visibility

Source: original release