Palantir CEO Alex Karp criticizes OpenAI and Anthropic over AI dependency

Palantir CEO Alex Karp has criticized frontier AI companies, including OpenAI and Anthropic, accusing them of attempting to “drug addict” humanity into a future controlled by these providers. In an interview with CNBC, Karp argued that these companies encourage businesses to become increasingly dependent on their models, potentially leading to a future where enterprises lose control over their data, intellectual property, and profitability.

Karp specifically targeted the business models of these firms, such as token-based pricing, which he suggests leads to “tokenmaxxing”—the maximization of AI token usage without necessarily generating measurable business value. He expressed concern that such models could result in job losses and allow adversaries to gain a competitive advantage.

In response to these criticisms, OpenAI and Anthropic have stated that customer data is protected and not used to train their models without explicit opt-in. Karp has advocated for the use of open-weight AI models, which allow businesses to customize and run models independently. Palantir recently joined other major tech firms in signing an open letter emphasizing the importance of open-weight models for national security.