Palantir CEO Alex Karp used a record-breaking quarter to deliver one of the more unusual arguments in recent corporate history: that the companies building the world’s most powerful artificial intelligence models are, in their own way, Marxists.
The charge arrived in Palantir’s Q2 2025 shareholder letter, published on 4 August 2025, as the Denver-based data analytics firm reported results that few in its sector could match. According to the Palantir Q2 2025 SEC press release, total revenue for the three months ended 30 June 2025 reached $1.004 billion, a 48% increase year on year, with a 27% operating margin. Net income attributable to common stockholders came in at approximately $327 million.
A note on figures: the original earnings announcement circulating in some coverage cited $1.9 billion in total revenue and 93% overall growth. The SEC filing puts total revenue at $1.004 billion, up 48%. The 93% growth figure does appear in Palantir’s own results, but refers specifically to U.S. commercial revenue, which rose from $159 million to $306 million year on year, per the official Q2 2025 shareholder letter.
Palantir CEO Alex Karp on the Model-Agnostic Alternative
Karp, who holds a doctorate in social theory, has long been comfortable reaching for philosophical frameworks to explain business strategy. In the shareholder letter, he trained that habit on frontier AI labs, writing: ‘There are Marxist overtones and undertones to our business. Others, including many of those building large language models, intend, knowingly or otherwise, to capture the means of production of their purported partners.’
On the subsequent earnings call with Wall Street analysts, he expanded the argument, asking whether companies are going ‘to buy into a future’ where their work helps their ‘adversaries win, and everybody who does win is a small, tiny group of people living in a tiny place that somehow believe because they eat vegetables and they don’t support war fighters that they deserve to have the total means of production of this country? And the rest of us should just sit back and absorb the cost of that revolution, which we’re paying for.’
His critique has a specific mechanism in mind. When enterprises sign up for large language model services, Karp argued, they are effectively subsidising a competitor: ‘You are paying for the right for them to migrate your IP, your know-how, your expertise to their model, so that they can build a competitive business that doesn’t require your business or people.’ The motive, he added, is ideological as much as commercial: ‘They are superior to you. They deserve to colonize your enterprise.’
Palantir’s answer to that dynamic is a model-agnostic platform: software that lets governments and enterprises run AI workflows while retaining control of their data and what Karp calls their AI ‘exhaust’, meaning the prompts, orchestration, and context generated during model use.
The Numbers Behind the Rhetoric
Karp’s philosophical riffs land differently when the underlying business is performing. According to Yahoo Finance’s summary of the Q2 2025 earnings call, Palantir’s Rule of 40 score, which combines revenue growth rate and profit margin, reached 94 in the quarter, up 11 points sequentially from 83 in Q1 2025. The company also cleared $1 billion in quarterly revenue for the first time.
That commercial momentum gives Karp’s polemic a particular edge. The AI labs he is criticising have not crowded Palantir out. Their proliferation has, if anything, expanded demand for the kind of enterprise AI governance layer that Palantir sells.
The underlying concern Karp is raising has gained traction beyond Palantir’s shareholder letters. Microsoft chief executive Satya Nadella has made related points about the competitive dynamics between AI platform providers and their enterprise customers. Observers have pointed to the list of companies that paid for or partnered with Anthropic and OpenAI, only to watch those labs enter adjacent markets spanning design, healthcare, legal, and drug discovery.
Palantir CEO Alex Karp is not the first to identify the tension, but he may be the most flamboyant in naming it. Whether enterprises act on the warning or continue buying tokens from the labs he is targeting will determine whether his analogy ages as prophecy or as theatre. The U.S. commercial revenue line, confirmed in Palantir’s own August letter, already growing at 93% year on year, suggests there is a market for the alternative he is selling. The question is how much of the broader enterprise AI spend ultimately flows through a neutral layer, and how much stays inside the labs’ own ecosystems.
