Senator Sanders Wants to Own AI Companies — and Hand America’s Adversaries the Keys
Senator Bernie Sanders published an op-ed in the New York Times last week demanding that the federal government seize a 50% equity stake in America’s largest AI companies. His proposed American AI Sovereign Wealth Fund Act would accomplish this through a one-time tax paid not in cash but in stock — handing Washington board seats, voting rights, and a direct role in steering some of the most consequential technology companies on earth. Sanders frames this as democratic accountability. It is nothing of the sort. It is government envy dressed up as economic policy — and it would create a national security catastrophe in the process.
Consider the security implications: A sovereign wealth fund holding 50% equity in OpenAI, Anthropic, and xAI simultaneously becomes the single most valuable cyberattack target in the history of nation-state warfare. China, Russia, and Iran wouldn’t need to compromise each AI company individually — they’d have one target: The fund’s governance infrastructure, its board communication channels, its privileged visibility into proprietary model architectures, training methodologies, and safety research. Every cleared official with board-level access becomes an insider threat vector. Every administration transition becomes a potential intelligence exposure event. Senator Sanders wants to democratize AI; what he’d actually construct is the world’s most devastating single point of failure for the entire U.S. AI industry — gift-wrapped for America’s adversaries.
The core economic argument rests on an equally flawed premise — that because AI models are trained on publicly created knowledge, the public deserves ownership of the companies that are built from it. By that logic, every highway contractor owes the government a stake because trucks use public roads. Every pharmaceutical company should surrender equity because its researchers attended state universities. The logical endpoint isn’t economic justice — it’s the elimination of private enterprise, achieved one righteous-sounding claim at a time. And note what else that ownership claim would hand the government: Privileged access to the proprietary training data, model weights, and vulnerability research that America’s adversaries most desperately want. Sanders isn’t proposing accountability. He’s proposing the largest involuntary intelligence disclosure in U.S. history.
Israel offers the sharpest possible antidote to this thinking — and the most instructive security parallel. The country built the most celebrated startup ecosystem outside Silicon Valley not by demanding equity but by enabling private capital with discipline and conviction. The Israel Innovation Authority provides conditional grants and matched funding, recouping its investment through royalties on revenues — catalyzing private capital without displacing it. That approach produced a tech sector generating 54% of Israel’s exports, $15.6 billion in private investment in 2025, and — critically — the cybersecurity companies now protecting critical infrastructure globally. Wiz, Armis, and Cyera didn’t emerge from a state-ownership model. They emerged because the government stayed in its enabling lane, keeping sensitive threat intelligence, customer vulnerability data, and proprietary security research out of government hands and therefore out of the crosshairs that government co-ownership would have painted on them.
Israel’s story also delivers the warning that Sanders refuses to hear. When the Israeli government pursued its judicial reform agenda in 2023, the Israel Innovation Authority sounded the alarm loudly and publicly. Government instability drives capital away with stunning speed — by early 2023, between 50 and 80% of newly incorporated Israeli startups were forming outside the country. Not because tax rates changed. Because entrepreneurs understood that a government willing to overreach in one domain will eventually overreach in theirs. Imagine what explicitly handing government 50% ownership — and board-level visibility into every security architecture decision those companies make — would trigger.
The indirect returns from a thriving AI sector don’t require equity stakes to materialize. Tax revenue, high-wage employment, global talent attraction, and spillover innovation compound through the broader economy for decades. More importantly, the security architecture of competitive, privately-operated AI systems is fundamentally stronger than any centralized state-controlled alternative — which is precisely why the U.S. funded the internet as infrastructure investments rather than as ownership stakes. Distributed, competitive, privately-governed systems are harder to compromise, harder to manipulate, and harder for adversaries to exploit than any structure with a single governmental owner sitting at the center.
Why This Matters
Government ownership of AI companies wouldn’t democratize AI. It would lobotomize it — and expose it. Every consequential decision on safety standards, deployment priorities, and competitive strategy would become hostage to electoral cycles and bureaucratic incentives. Every oversight mechanism that currently provides independent security accountability — CISA’s authority to compel disclosure, NSA’s ability to assess AI-enabled threat vectors, the FTC’s enforcement independence — would collapse the moment the federal government holds board seats in the companies being overseen. The companies best positioned to build the future would incorporate elsewhere, raise capital elsewhere, and build elsewhere. Senator Sanders is right that AI’s benefits should reach broadly across society. He is catastrophically wrong about how to get there — and dangerously indifferent to what his proposal would hand America’s adversaries on the way.
