Palantir CEO Alex Karp Criticises AI Labs After Record-Breaking Quarter

Palantir Chief Executive Officer Alex Karp used the company’s latest quarterly shareholder letter and earnings call to sharply criticise leading artificial intelligence labs, arguing that enterprises should be cautious about handing over their data and AI infrastructure to model developers.

Karp made the remarks as Palantir reported record financial results for its second quarter, benefiting from growing enterprise demand for artificial intelligence software. The company posted $1.9 billion in revenue, up 93% from the same period a year earlier, and $1.1 billion in profit. In his shareholder letter, Karp noted that the quarterly profit exceeded Palantir’s total revenue during the corresponding quarter of the previous year.

Despite the strong results, Karp devoted part of his shareholder letter to criticising the business practices of some frontier AI companies. He argued that many developers of large language models ultimately seek to control the “means of production” used by their enterprise customers, describing those ambitions as having “Marxist overtones and undertones.”

Karp questions AI labs’ business models

During the earnings call, Karp expanded on those comments, suggesting some AI providers expect businesses to contribute valuable intellectual property and operational knowledge that can later be incorporated into competing AI products. Using provocative language, he questioned whether companies should support an ecosystem that could eventually strengthen future competitors rather than their own organisations.

Palantir positions itself differently, offering model-agnostic software that allows governments and enterprises to use multiple AI models while retaining control over their data, prompts, orchestration workflows and operational context. The company argues that customers should own what Karp described as their AI “exhaust,” rather than allowing that information to become part of another company’s platform.

Karp also criticised organisations that continue paying AI providers while potentially helping improve models that could later compete against them. He argued enterprises should carefully evaluate how their data and expertise are used when deploying generative AI technologies.

Broader debate over enterprise AI

Karp’s comments reflect a broader discussion emerging across the technology industry about how enterprises should deploy artificial intelligence. Microsoft Chief Executive Satya Nadella has recently encouraged organisations to adopt multiple AI models and avoid depending on a single provider, arguing that flexibility reduces both costs and vendor lock-in.

Supporters of that approach point to the growing number of companies that have partnered with or purchased services from AI developers such as OpenAI and Anthropic, while those same AI companies have expanded into additional commercial markets, including design software, healthcare operations, legal services and drug discovery.

Although Karp portrayed Palantir’s strategy as an alternative to those business models, the company’s latest financial performance also illustrates the rapid expansion of enterprise AI spending across the industry. With demand continuing to grow, the market remains large enough for multiple companies to compete while pursuing different approaches to enterprise artificial intelligence.