Federal Gambling on AI: The High Cost of Buying a Boardroom Seat
The United States government is quietly considering a move that looks less like sound economic policy and more like purchasing front-row tickets to the Titanic’s final act. Reports suggest federal regulators are weighing whether to take direct financial stakes in artificial intelligence giants like OpenAI and Anthropic. This isn’t just a regulatory pivot; it’s a massive, unprecedented bet on an unproven, unprofitable sector.
The New Paradigm of State Ownership
Since January 2025, the U.S. government hasn’t just been tinkering with antitrust enforcement—they’ve completed at least 16 separate deals worth a combined $20.9 billion in direct ownership stakes, according to the Council on Foreign Relations. This marks a fundamental shift from the traditionally hands-off neoliberal framework that defined American tech policy for decades.
Take Intel, once a symbol of American technical prowess and now a national security backstop. The Department of Commerce has taken a 10% stake, and the Development Finance Corporation has invested in minerals, energy, and infrastructure. Meanwhile, the Department of Defense has undertaken at least seven similar deals.
In quantum computing, theCommerce Department recently invested $2 billion for minority controlling stakes in nine firms—including D-Wave, Quantinuum, IonQ, and Rigetti Computing. None of these companies are profitable yet, but the fear that quantum computing might someday accomplish tasks impossible with classical computers keeps the funds flowing.
Financial Quicksand and Strategic Err
The core justification for this intervention is national security, but the numbers tell a deeper, more troubling story. OpenAI alone carries over $1.4 trillion in financial commitments spread across the next eight years—and remains fundamentally unprofitable. That’s not just a red flag for investors; it’s a structural warning sign.
Sam Altman, OpenAI’s CEO, reportedly pushed for federal investment last year. But when CFO Sarah Friar reportedly warned about the strings attached to federal loan guarantees, Altman publicly walked back his support. Still, rumors persist of active negotiations for a 5% equity stake under the current administration—a desperate attempt to bridge the gap between astronomical R&D costs and sustainable, revenue-based valuation.
The Populist Challenge: Sanders’ Sovereign Wealth Fund
While the executive appetite for Silicon Valley equity grows, the legislative response is equally radical. Senator Bernie Sanders recently unveiled the American AI Sovereign Wealth Fund Act, a plan that would fundamentally reshape how the U.S. markets treat frontier tech.
The proposal calls for a one-time 50% stock tax on any AI company generating over $200 million in revenue, pooling the capital into a $7 trillion sovereign wealth fund. Crucially, it also outlines a bipartisan Independent Commission for Democratic AI with veto power over industry mergers and model training configurations.
This is an explicit challenge to the Silicon Valley model of concentrated, private control—and a promise to return a portion of AI-generated revenue directly to citizens. It frames AI not as proprietary hardware or software, but as a public utility built on publicly sourced training data.
Bubble Reality: The Cost of Intelligence
The rush to secure board seats and ownership stakes ignores the underlying fragility of the business models themselves. We are seeing market saturation risks that few seem willing to acknowledge.
Frontier model pricing is exorbitant. Claude Mythos, for instance, costs $25 per million input tokens and a staggering $125 per million output tokens—a 5x increase from its predecessors. This is already forcing corporate consumers to adapt: Uber, for example, capped usage at $1,500 per month per employee.
When your primary product becomes so expensive it forces customers to throttle their own reliance, the bubble isn’t just about valuation—it’s about structural operational viability. Harvard and Yale economists noted in a 2021 paper that government funding can bolster innovation only when private markets fail to invest due to uncertainty. But direct ownership in hyper-speculative consumer AI models may not meet that threshold.
Copyright Quagmire: The Liability Black Hole
There are at least 115 active copyright lawsuits currently mired in the courts against AI companies. The full liability landscape isn’t just unknown—it’s potentially catastrophic. OpenAI and Anthropic built their models on content scraped from the open web, often without compensation or permission.
Buying into this ecosystem with taxpayer money isn’t just speculative—it’s reckless. Before committing billions in public capital, the U.S. government should allow the legal system, consumers, and the market to establish a genuine, sustainable value. We’re rushing to buy the Titanic when we should be checking the hull for holes.
What’s the Alternative?
The smarter move is patience, not desperation. Let the bubble play out. Wait for courts to issue rulings on copyright infringement. Let pricing stabilize across multiple providers. Watch whether smaller, open-weight models like Meta’s Llama series or Mistral’s offerings gain enough traction to force Big AI to lower prices.
The government already has leverage: it’s a massive buyer of enterprise AI services itself, and its procurement power can influence market direction without putting taxpayer capital at risk.
The danger isn’t in failing to act—it’s in acting too early, on the strength of promises that don’t yet match reality.
Conclusion: Letting the Dust Settle
With over 115 active copyright lawsuits and pricing that’s already pricing out corporate customers, the full liability landscape for these AI companies isn’t just unknown—it’s cataclysmic.
Buying into this, with taxpayer money, is the definition of reckless. Before committing billions in public capital, the U.S. government should allow the legal system, consumers, and the market to establish a genuine, sustainable value.
We’re rushing to buy the Titanic when we should be checking the hull for holes. The smarter move is patience, not desperation. Let the bubble play out, rather than subsidizing its inevitable pop.
