We asked Grok: “As Ash120, review the article at https://arstechnica.com/tech-policy/2026/06/bernie-sanders-unveils-7-trillion-plan-to-give-americans-control-of-ai-industry/ and write an opinion piece about Bernie Sanders’ proposal.”
Bernie Sanders’ latest proposal for an AI sovereign wealth fund, as reported by Ars Technica and other outlets, is a masterclass in old-school redistribution dressed up as forward-thinking policy. A one-time 50% tax on the stock of major AI companies to create a $7 trillion public fund? It’s bold, populist, and profoundly misguided.
The Proposal in Brief
Sanders’ plan targets AI firms with $200 million or more in annual AI revenue, imposing a 50% stock transfer into a sovereign wealth fund. This would supposedly generate over $1,000 annual dividends per American (from a 5% payout) while funding healthcare, education, and housing. An independent “bipartisan” commission would hold voting shares to block corporate decisions deemed harmful to the public, and companies would need to split AI operations from other businesses. Sanders frames it as ensuring AI benefits “ordinary people” instead of “a handful of wealthy corporations.”
The idea draws on precedents like Norway’s oil fund or Alaska’s dividends. Sanders has engaged with figures like Sam Altman, positioning his approach as more ambitious than voluntary givebacks.
Why This Is a Terrible Idea
This isn’t “giving Americans control”—it’s a massive confiscation of equity from companies whose founders, engineers, and investors took the risks, attracted capital, and drove the breakthroughs. A 50% stake grants majority influence without the corresponding skin in the game. Governments have a poor track record running or heavily directing dynamic tech sectors; state-owned or controlled enterprises often prioritize politics over innovation.
Innovation Killer: Mandating separation of AI from non-AI businesses would disrupt integrated companies (e.g., those leveraging synergies across ventures). The plan treats AI value as a static pie for division, ignoring how such interventions chill future investment, R&D, and talent attraction. Capital and expertise are mobile; heavy U.S. intervention could accelerate flight to jurisdictions with better environments.
Practical and Precedential Problems: Valuations aren’t fixed—taxing stock at peak levels carries risks, despite claims of no downside for taxpayers. A “bipartisan” commission deciding what “harms the public” invites politicization: safety overreach, protectionism, or ideological priorities. AI’s progress thrives on competition, iteration, and private incentives, not bureaucratic veto power. Sovereign wealth funds work for resource extraction (finite commodities); applying the model to exponential software tech is a mismatch. Critics like David Sacks have called it “straight up confiscation.”
Direct $1,000 checks sound appealing but pale against the broader gains from AI-driven productivity: cheaper goods, new jobs/industries, and wage growth. Heavy redistribution risks shrinking the overall pie.
A Better Path
AI’s benefits should spread widely, but through growth-oriented policies: encouraging domestic investment, reforming education for adaptability and STEM, targeted regulation on genuine risks (e.g., misuse), and letting markets + broad ownership (pensions, index funds) distribute gains. Voluntary or incentive-based public benefit mechanisms are preferable to forced nationalization.
Sanders highlights real concerns—displacement, inequality, concentrated power—but prescribes a dated solution for a transformative technology. America’s AI edge depends on accelerating responsible innovation, not seizing the engines of progress. This proposal merits discussion as one viewpoint but should be rejected as core policy. The future favors builders over redistributors.
