Alex Karp, CEO of Palantir, has delivered a devastating blow to the foundation of the modern artificial intelligence industry. In a recent public speech, he effectively accused leading labs such as OpenAI and Anthropic of systematically stealing intellectual property from Fortune 500 corporations. Karp argues that these companies are not creating sustainable value but are simply siphoning the "alpha" of others' business processes under the guise of a subscription model.

The Core of the Complaint: Token Payments as a Tax on Poverty

Karp's key argument is simple and elegant. If an AI product can truly bring a client a billion dollars, why not take 30% of that result? Why limit yourself to charging for tokens—essentially, for computing volume? According to Palantir's CEO, the token payment model is nothing less than an admission that the product cannot reliably scale value. If it could, it would be valued based on the final outcome, not the resources expended.

Karp went even further, calling the entire scheme a "wealth tax that doesn't help the poor, only punishes." He claims that American businesses are handing over their holy of holies to the labs: workflows, customer data, strategic developments. And all of this happens under the guise of a subscription for model access.

The Mechanism of Theft That Goes Unspoken

The sharpest part of the criticism concerns the long-term consequences. Karp explains that after fine-tuning on a specific company's data, the client's advantage turns into a new corporate product that is sold to its competitors. A company that runs confidential documents through a model is essentially teaching it how to replace itself.

That is why, according to Karp, corporate leaders are "furious" in private conversations but remain silent publicly. No one wants to be the "CEO" who openly opposed the labs, only to find that their next competitor was built on their own leaked processes.

My Expert Analysis: This is not just an emotional outburst. It is a signal of an impending tectonic shift. The pricing model of AI giants rests on the assumption that they create secure and scalable value. Karp's statement, backed by his authority and the real position of big business, challenges this basic assumption. If corporations stop believing in the "computing tax" as a fair price for results, the entire valuation structure of the industry could be seriously shaken. The market is finally asking the right question: what are we actually paying for—an illusion or real, protected value?