After killer quarter, Palantir CEO Alex Karp calls AI industry ‘Marxist’
Palantir CEO Alex Karp used the company's second-quarter shareholder letter and earnings call to accuse major AI labs of behaving in a "Marxist" manner, arguing that firms building large language models are effectively seizing the "means of production" from the enterprises and governments they partner with. The remarks, made despite Palantir's own record results, highlight growing tension in the AI industry over how much control frontier labs retain over clients' data and expertise even as they sell them AI tools.
Karp, who holds a PhD in social theory, contrasted Palantir's model-agnostic approach, which he says lets organisations keep control of their data and AI outputs, with rivals he accused of using partnerships to absorb clients' intellectual property and expertise to build competing products. The comments came as Palantir reported $1.9 billion in second-quarter revenue, up 93% year-on-year, and $1.1 billion in profit, exceeding its total revenue from the same period a year earlier. Karp's critique echoes concerns raised elsewhere in the industry, including by Microsoft CEO Satya Nadella, about companies that pay AI labs like Anthropic and OpenAI while those same labs expand into competing sectors such as design, healthcare, legal and drug discovery.
- Palantir CEO Karp calls AI labs "Marxist" for controlling client data and IP
- Palantir posted $1.9bn revenue (up 93%) and $1.1bn profit in Q2
- Karp contrasts Palantir's model with rivals like OpenAI and Anthropic
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Those sympathetic to Karp's critique argue that several frontier AI labs have built unusually centralised, mission-driven cultures, in which safety commitments, ideological frameworks and long-term societal goals shape product decisions as much as customer demand does. They contend this can look less like ordinary commercial behaviour and more like a small group of technologists deciding, on the public's behalf, how a transformative technology should be controlled and distributed. For enterprises and governments needing dependable, mission-critical systems, they argue, a vendor whose priorities are openly aligned with practical deployment and accountability is preferable to one shaped primarily by abstract ideological commitments. Palantir's commercial success, on this view, reflects genuine demand for that alternative approach, even if Karp's chosen language is deliberately provocative.
The case against
Sceptics of the remark argue that labelling the AI industry 'Marxist' is a loose and inflammatory generalisation that obscures a sector overwhelmingly funded by venture capital and organised along conventional profit-seeking lines. They note the comment arrived immediately after Palantir itself posted a $1 billion profit, which invites the reading that it functions as competitive positioning against rivals rather than a considered structural critique. They would also point out that caution among AI labs around safety and controlled deployment reflects a genuine, widely shared concern about real risks from powerful AI systems, not an ideological attempt to seize control of the technology. On this view, Karp's framing risks reducing a legitimate debate about AI safety and governance to a political slogan.