Palantir CEO calls AI labs 'Marxist' after record $1.9B quarter

Karp's shareholder letter crystallizes a tension that has been building in enterprise AI adoption. Palantir and its CEO are not alone in raising this concern—Microsoft CEO Satya Nadella has articulated similar points. The underlying issue is real: companies like OpenAI and Anthropic have received substantial investment and partnership commitments from enterprises, yet simultaneously launched or expanded business lines in domains where their customers operate (healthcare, legal services, design, drug discovery). From the enterprise customer's perspective, this creates a paradox: they fund the AI capability, provide data and use-case expertise through their work with the lab, and then watch as the lab pivots that capability into a competing product.

Karp's use of Marxist framing—deliberately provocative in a shareholder letter—is designed to highlight what he sees as a structural imbalance in value capture. His point, stripped of rhetorical flourish, is that traditional vendors let customers own and control their intellectual property and data exhaust, whereas AI labs do not. However, the article also acknowledges a countervailing reality: the AI market is growing so rapidly that there may be room for multiple business models to coexist. Palantir's record quarter suggests that enterprises do have alternatives and are voting with their wallets for vendors that promise data sovereignty—but OpenAI and Anthropic's continued growth indicates the market has not yet reached consensus on which model customers prefer.