Palantir Technologies Inc. CEO Alex Karp has launched a sharp critique against leading AI labs like OpenAI and Anthropic, asserting their business models are fundamentally flawed. He warns corporate clients are paying escalating token costs for limited value while exposing their core data and intellectual property to risk.
In a CNBC interview on Wednesday, Karp used strong language, describing the business models of top AI labs as "just insane" and accusing Anthropic and OpenAI of misleading enterprise partners. He claims they exaggerate AI risks while selling their most powerful models to global businesses and governments.
He characterized this as enterprises being forced to pay a "wealth tax," effectively surrendering their competitive advantage to a third party. When CNBC host Becky Quick remarked, "You sound angry," Karp responded, "This is the voice of American business, speaking through me."
These comments reflect growing dissatisfaction among corporate clients with U.S. AI labs. High costs, questionable return on investment, and increasing regulatory pressure are prompting some major American corporations to consider cheaper Chinese alternatives, creating multiple challenges for entities like OpenAI and Anthropic.
The Core of the Criticism
Karp specifically targeted the token-based billing model used by Anthropic and OpenAI. "I'm not picking on them, but something is fundamentally broken," he stated. "The general consensus in the business community of this country is that I'm just wasting time and burning tokens."
He argued that while AI labs hype risks, they aggressively market powerful models, leading enterprise clients to pay a "wealth tax" to these platforms. This transfers their "alpha"—their core competitive edge—to external parties. On national security, his language was even stronger: "Are we really going to outsource the battlefields of this nation to the consensus view in Silicon Valley? That's just insane."
A Broader "AI Sovereignty" Manifesto
Karp's public remarks are not an isolated incident. The day before the interview, Palantir published a nine-point "AI Sovereignty" manifesto on X, systematically outlining its critical stance on the current AI ecosystem.
The manifesto's core logic revolves around "sovereignty." It argues that institutions relinquish their future agency by ceding control over their data, model weights, and competitive edge to parties likely to profit from it. The manifesto explicitly states that "tokenmaxxing" incentives promote disposable scripts over robust software systems and create an "addiction to false progress," noting, "There is a reason those selling tokens refuse to charge for value."
It further emphasizes that controlling model weights means controlling an institution's destiny. "Weights are the distillation of the knowledge an institution has worked so hard to accumulate. To let someone else control your weights is to allow them to migrate your business alpha to themselves."
Part of a Wider Enterprise Backlash
Karp's criticism is not an outlier but a concentrated expression of broader corporate discontent with U.S. AI labs. Soaring usage costs, difficult-to-quantify returns, and an increasingly complex regulatory environment are collectively eroding trust in mainstream AI platforms.
Some large U.S. companies are already looking toward more cost-competitive Chinese AI alternatives, a trend posing direct commercial pressure on firms like OpenAI and Anthropic. Palantir's position is clear: enterprises should treat data as a core asset and view AI sovereignty as a prerequisite for institutional survival, rather than entrusting their fate to the "consensus judgment" of external platforms.
Comments