US floats sovereign wealth fund to stake citizens in AI companies directly

The Trump administration is exploring a sovereign wealth fund structure that would give individual Americans direct financial stakes in major artificial intelligence companies, according to public statements from Vice President JD Vance. The proposal, still in conceptual stages, represents a departure from traditional federal technology policy and introduces sovereign capital allocation mechanisms typically associated with resource-rich Gulf states to the US domestic equity landscape.

Vance did not specify funding sources, target companies, or implementation timeline. The statement came without accompanying legislative text or Treasury guidance. Elon Musk, who holds advisory proximity to the administration, separately advocated for direct cash payments to citizens funded by AI company equity stakes, though his comments did not reference official policy channels. No White House working group or interagency task force has been publicly named to structure the vehicle.

The proposal arrives as US AI companies trade at elevated multiples despite mounting regulatory scrutiny in the EU and nascent domestic debates over compute resource allocation. A federal sovereign fund holding minority stakes in five to ten large-cap AI firms would immediately become a top-twenty shareholder in most targets, creating governance complications and potential conflicts with existing antitrust frameworks. The Norway Government Pension Fund Global, the $1.6 trillion model most frequently cited in sovereign wealth discussions, operates under strict passive investment mandates and avoids domestic equity concentration—constraints a US vehicle might not adopt. If the administration pursues an active ownership model, the fund would face immediate questions about board representation, voting rights on AI safety governance, and whether stakes dilute or concentrate existing shareholder power.

The concept raises structural questions allocators are already gaming. Distribution mechanisms matter: per-capita dividend payments create different incentive structures than accounts citizens access at retirement or sell in secondary markets. Norway distributes zero oil fund proceeds directly to citizens; Alaska's Permanent Fund pays annual dividends averaging $1,600 but holds no equities in Alaskan companies. A US AI fund holding domestic equities and paying direct distributions would be without clean precedent. Fund managers would need to price regulatory risk—an administration holding 3-8% stakes in Anthropic, OpenAI's future public vehicle, or scaled competitors gains leverage over export controls, compute allocation, and model release timelines. That leverage is either worth a premium or a discount depending on which second-order effects dominate.

Operators should watch for three signals: first, whether Treasury or a new statutory body is named to lead structuring work, expected within 90 days if this advances beyond trial balloon. Second, whether initial funding comes from new debt issuance, reallocation of existing federal investment vehicles, or a novel revenue mechanism tied to AI compute taxes or data licensing fees. Third, whether the administration defines "major AI companies" to include only public equities or attempts to negotiate direct stakes in private firms like OpenAI, Anthropic, and xAI, which would require consent from existing cap tables and trigger valuation disputes.

The proposal's fiscal mechanics remain entirely unspecified, and no comparable democracy has attempted direct equity redistribution at this scale in a high-growth technology sector while the sector is still concentrating.