The AI Profit Question – Between Wealth Tax and Nationalization

Thursday Diagnosis #09

It’s a remarkable prediction that Palantir CEO Alex Karp made in an interview in early July 2026. When asked about U.S. Senator Bernie Sanders’ call for the government to hold a 50 percent stake in major AI companies, Karp said: In two years, people will no longer consider Sanders a progressive; instead, they’ll ask why he only wanted 50 percent. Karp, who describes himself as a “staunch progressive,” sees complete nationalization as the emerging majority position on the political left.

This brings into sharp focus a question that goes far beyond server capacity and AI models: Who controls value creation in a world where human labor is gradually being phased out?

Taxes vs. Governance: Two Approaches to Redistribution

Until now, the idea of a wealth tax has dominated the political debate. Karp firmly opposes it. He argues that such a tax would merely be a punishment for billionaires and would ultimately do nothing to help the poor.

This reveals a crucial political dividing line in governance that tech CEOs often blur: A tax alone extracts capital and redistributes it, but does not grant the state any direct say in how the technology develops. State ownership (nationalization), on the other hand, means real influence. It involves intervention in corporate management—for example, through seats on the board of directors, control over intellectual property (IP), and a direct say in technology development. This, too, is currently being discussed in the U.S.

Blog-Abo

Neue Beiträge direkt im Postfach.

Blog abonnieren

Political Manipulation in the Tech Sector

However, the debate over redistribution and government involvement is rarely driven by the tech elite’s genuine concern for the common good. Rather, it serves as a political tool:

  • David Sacks, a tech investor and former AI advisor to Donald Trump, warns conservatives about Sanders’ plans but cynically admits that he might almost support such government involvement as a kind of “stupidity tax” on careless tech companies.
  • At the same time, OpenAI CEO Sam Altman is secretly negotiating with Trump about precisely this kind of government investment in his company.
  • Donald Trump, for his part, is using the issue as a campaign tool and promises to “give something back” to voters ahead of the midterms in November, following talks with the tech industry.

Those who speak most loudly here about job losses and the concentration of profits are using the issue of income distribution as a strategic weapon—against competitors, against the other party, and in the run-up to elections.

“Stupidity Tax“: Economic liberals like David Sacks accuse the CEOs of major AI labs of deliberately stoking panic for months on end. In interviews and before the U.S. Congress, these executives repeatedly warned of existential risks, massive job losses, and social upheaval caused by AI. This is now the price they’re paying.

The Technological Proxy War

The reason this debate over distribution is so contentious in the first place is the current business model of leading AI labs (Frontier Labs). Companies currently pay per text unit (token), outsource sensitive data, and are gradually losing their intellectual property to the model providers, who profit from sheer volume.

Tech companies offer their own quick fixes for this: Palantir markets its “ontology layer” as a protective intermediary layer designed to keep data with the customer. Competitors such as Mistral founder Arthur Mensch, on the other hand, warn against precisely such proprietary intermediate layers and tout open models (open source) as the solution. But in the end, both solutions merely replace one dependency with another and aim to secure their own business models.

Conclusion: The Real Question of Sovereignty

You can’t accuse Alex Karp, Sam Altman, or David Sacks of acting in the interests of their own businesses or political networks. It would be naive to expect them to act with social responsibility.

True sovereignty is not a place where a specific software layer resides. It is the responsibility of lawmakers, civil society actors, and regulatory authorities to discuss and establish these framework conditions. However, anyone who gets lost in the choice between the ontology layer and open source has long since delegated the most important question to the wrong authority: What legal and contractual mechanisms ensure that the enormous productivity gains of an automated AI world benefit the general public—whether through higher wages, lower prices, dividends, or a high unconditional basic income?

Sources

  • CNBC Television: Alex Karp: Most of the things Anthropic discusses publicly are running on Palantir, mid-June 2026 (YouTube)
  • Palantir Technologies, LinkedIn Manifesto on AI Sovereignty, Nine Theses, July 1, 2026 (LinkedIn)
  • Arthur Mensch (Mistral): “Mistral Founder Mensch: ‘Those Who Rely on Proprietary AI Models Risk Their Trade Secrets,’” the-decoder.de, July 5, 2026
  • Gizmodo: “Palantir CEO Says Bernie Sanders Will Regret Wanting Only 50% Public Ownership of AI Companies,” June 10, 2026
  • Gizmodo: “The Leopard Is Eating David Sacks’s Face,” June 6, 2026

Auf dem Laufenden bleiben

The NeWoS – Newsletter über Arbeit, KI und gesellschaftlichen Wandel

Einordnungen zu technologischen Entwicklungen, Zukunft der Arbeit und gesellschaftlichen Veränderungen. Kuratiert von FROLLEINFLOW.

Mehr erfahren
Scroll to Top