Trump Says He Is Considering a Government Equity Stake in Top AI Companies

On June 5, 2026, President Donald Trump told reporters that his administration is exploring direct equity stakes in the country's most valuable artificial intelligence companies. "There are concepts where pieces could be given to the American public, where the American public essentially becomes a partner with the companies," Trump said, adding that he planned to discuss the idea with AI company executives as soon as the following week [1][2].

The companies reportedly under consideration — OpenAI, Anthropic, and xAI — sit at the center of a technology sector whose combined valuation now rivals the GDP of major nations. OpenAI alone closed a $122 billion funding round in March 2026 at an $852 billion valuation, with an IPO targeting a valuation above $1 trillion planned for September [3]. The proposal would mark one of the most significant expansions of federal ownership in private enterprise in American history.

The Proposal: A "Public Wealth Fund" Seeded by AI Equity

The concept did not originate in the White House. OpenAI CEO Sam Altman first pitched the idea directly to Trump in early 2025 and has discussed it with senior administration officials periodically since then [4]. In April 2026, OpenAI published a policy proposal outlining a "Public Wealth Fund" that would be seeded by voluntary equity donations from AI companies. The fund would "invest in diversified, long-term assets" and distribute returns as dividend payments to American households [4][5].

The operative word is "voluntary." Unlike the 2008 financial crisis bailouts, where the government acquired equity in exchange for emergency capital, this proposal envisions companies ceding shares without receiving government funds in return. No formal agreement is in place, and no company has confirmed participation [2].

But the framing raises immediate questions: how voluntary is a donation to a government that simultaneously controls your regulatory environment, awards your defense contracts, and shapes your export licensing?

The Numbers at Stake

The dollar figures involved are staggering. Even a modest stake — say, 1% — in the companies under discussion would represent tens of billions of dollars in value.

Source: CNBC, Bloomberg, company filings

Data as of Jun 5, 2026CSV

OpenAI's valuation has grown from $300 billion in March 2025 to $852 billion in March 2026, a nearly threefold increase in a single year [3]. The company's February 2026 fundraise alone brought in $110 billion from Amazon ($50 billion), SoftBank ($30 billion), and Nvidia ($30 billion) [3]. For a private company, valuation is determined by the price investors pay in funding rounds — but there is no public market price discovery. Who would set the terms of a government equity grant? The company? An independent appraiser? The question is unanswered.

For public companies like Nvidia (market capitalization: approximately $3.2 trillion) and Microsoft (~$3.1 trillion), market prices provide a reference point — but government acquisition of shares in publicly traded companies would itself move markets and raise distinct securities law questions [6].

An Already-Growing Portfolio

What has received less attention is that the Trump administration already holds ownership stakes in roughly 20 private companies, according to an analysis by the Cato Institute [7]. These include:

• A 10% share in Intel, the semiconductor manufacturer
• A "golden share" in U.S. Steel Corporation, giving the government say over certain decisions
• Stakes in MP Materials (rare earth mining), IBM, and GlobalFoundries (quantum computing and semiconductors) [7]

This portfolio represents a significant and largely unreported expansion of federal ownership in the private sector. The AI proposal would extend this pattern into the most valuable and strategically sensitive technology companies in the world.

Historical Precedent: Lessons from 2008

The most relevant American precedent for government equity stakes in private companies is the Troubled Asset Relief Program (TARP) and related interventions during the 2008-2009 financial crisis.

Source: U.S. Treasury, Congressional Research Service

Data as of Dec 1, 2025CSV

AIG: The government provided $85 billion in emergency loans in exchange for a nearly 80% equity stake in the insurance giant, later committing an additional $40 billion through TARP. Government ownership peaked at 92% in January 2011 before the Treasury sold all common equity by December 2012, ultimately generating a positive return [8].

General Motors: The government acquired a 60.8% stake in exchange for $49.5 billion in bailout funds. It received 912 million GM shares and began selling when GM went public again in November 2010. The government sold its last shares in December 2013, losing approximately $10 billion on the investment [9].

TARP banks: The program invested $245 billion in financial institutions and ultimately recovered approximately $275 billion, including dividends and interest [8].

A critical legal footnote: in June 2015, the U.S. Court of Federal Claims ruled in Starr International v. United States that the government had overstepped its authority by taking an 80% stake in AIG, even during a genuine financial emergency [10]. The court found the government had no statutory basis for demanding equity as a condition of the rescue. If the legal basis was contested during a crisis, the foundation for equity stakes absent any crisis is even less settled.

The Legal Vacuum

The constitutional and statutory mechanisms for acquiring AI equity stakes remain unclear. Several distinct legal problems present themselves.

First, there is the question of Congressional appropriation. The Constitution vests spending authority in Congress. If the government is acquiring equity "voluntarily" — i.e., companies are donating shares — no appropriation may be needed. But if the donations are implicitly conditioned on favorable regulatory treatment, government contracts, or export licenses, courts could characterize them as unconstitutional conditions [10][11].

Second, conflict of interest problems are acute. The federal government is simultaneously the largest customer of AI technology (through the Department of Defense and intelligence agencies), the primary regulator of AI (through executive orders and potential legislation), and — under this proposal — a shareholder seeking financial returns. As the Lawfare Institute noted, officials would be "pulled between promoting competition, policing safety, and protecting taxpayer returns" [11].

Third, state constitutional restrictions add complexity. Many state constitutions prohibit state entities outside pension funds from holding corporate equity. California, for instance, restricts such ownership — relevant because many AI companies are headquartered there [11].

Legal scholars have identified minimum safeguards that would be needed: clear statutory mandates from Congress, strict separation of regulatory and investment functions, and public reporting requirements to manage conflicts [11].

The AI Industry's Extreme Concentration

Any government equity program would operate in a market defined by extraordinary concentration. The AI sector is dominated by a handful of firms across every layer of the technology stack.

Source: IDC, company reports

Data as of Mar 1, 2026CSV

Nvidia controls an estimated 81% of the AI data center chip market, with AMD holding roughly 10% [12]. In cloud infrastructure, which provides the computing power for AI training and deployment, three companies — AWS (~30%), Microsoft Azure (~20%), and Google Cloud (~13%) — control the majority of the market [12]. ChatGPT holds 64-68% of global web traffic among AI chatbots [12]. The global AI market reached $514.5 billion in 2026, growing 19% from the prior year, with 14 of the 15 companies in top AI rankings being U.S.-listed [12].

This concentration means a government equity program, even if nominally open to all, would inevitably be selective. The administration has named three companies — OpenAI, Anthropic, and xAI — but has not explained the selection criteria. Would companies need to meet a valuation threshold? A market share test? A national security designation? The absence of stated criteria invites concerns about favoritism, particularly given that xAI was founded by Elon Musk, who has served as a senior advisor to the Trump administration [2][5].

The China Comparison — And Its Limits

The closest international parallel is China's "golden share" program, through which the government acquires small stakes — typically 1% — in technology companies that carry disproportionate voting rights or veto power over business decisions [13][14].

Since 2017, the Cyberspace Administration of China has taken golden shares in subsidiaries of Alibaba, ByteDance (TikTok's parent), Sina Weibo, and Kuaishou, among others [13]. These stakes give the government the ability to influence content moderation, data-sharing, and strategic decisions. Analysts have described the program as "embedding the Chinese Communist Party within the nerve-centers of China's most important internet-content companies" to achieve "pervasive surveillance, censorship and policing capabilities from the inside out" [14].

The Trump administration has framed its proposal differently — as a wealth-sharing mechanism rather than a control mechanism. But the structural dynamics are similar: a small equity position, backed by the implicit power of the regulatory state, that gives the government a seat at the table.

Other models exist. France holds stakes in Airbus and other strategic firms through its investment arm. Singapore's sovereign wealth fund, GIC, and Norway's Government Pension Fund Global both invest in technology companies globally, but as passive financial investors without governance rights. The distinction between passive investment and active governance is critical — and the Trump proposal has not clarified which model it envisions [15].

The Taxpayer Investment Argument

Advocates for government equity stakes make a straightforward case: American taxpayers, through decades of federally funded research, built the foundations on which today's AI industry stands.

Federal AI research funding runs through multiple agencies. The National Science Foundation allocated $494 million for AI research, DARPA contributed $314 million, the National Institutes of Health $309 million, the Department of Defense $233 million, and the Department of Energy $187 million [16]. The FY2025 budget proposed $3 billion for AI across the federal government, including $2 billion for NSF alone — the largest increase for any single agency [16]. NSF has launched X-Labs, a $1.5 billion ten-year program for breakthrough research [16].

The technologies underlying modern AI — neural networks, natural language processing, computer vision — trace their origins to government-funded research programs stretching back decades. The argument is that if public investment created the conditions for private profit, the public should share in the returns. Senator Bernie Sanders has made this case explicitly, calling for 50% government equity stakes in AI companies [5].

The counterargument is equally direct: government equity ownership creates perverse incentives that distort markets and slow innovation. The Cato Institute has warned that the administration's growing portfolio of company stakes represents an unprecedented expansion of federal power over private enterprise [7]. If government officials have a financial interest in the companies they regulate, the incentive to regulate in the public interest — rather than to protect share value — is compromised.

Who Gets Hurt

The effects of government equity stakes would ripple across multiple stakeholder groups.

Existing shareholders would face dilution. If OpenAI donates, say, 5% of its equity to the government, existing investors — including Microsoft, Amazon, SoftBank, and employees with equity compensation — would see their ownership percentages reduced proportionally. For employees at AI startups, where equity compensation is often the primary form of wealth-building, dilution represents a direct pay cut [6].

Startup competitors face a different problem. If the government holds equity in OpenAI, Anthropic, and xAI, those companies gain an implicit guarantee of favorable treatment. Smaller competitors without government backing would face higher costs of capital, reduced access to government contracts, and the perception — justified or not — that the market is tilted against them [7].

Foreign AI firms would confront a new form of protectionism. Government-backed companies would enjoy advantages in competing for U.S. government contracts, potentially distorting global competition in AI [15].

Economists have identified specific market distortions: government-backed firms could access cheaper capital (because investors perceive reduced risk of adverse regulation), capture a larger share of government procurement, and enjoy informational advantages from their relationship with regulators [7][11].

Governance Rights: The Unanswered Questions

Perhaps the most consequential unanswered question is what governance rights would accompany any equity stake. The range of possibilities is vast:

• Board seats would give the government direct influence over corporate strategy, hiring, and investment decisions
• Veto power over data-sharing agreements, foreign partnerships, or export decisions could effectively make the government a co-decision-maker on every major business question
• Access to model weights — the trained parameters that define an AI system's capabilities — would give the government insight into proprietary technology that is currently among the most closely guarded intellectual property in the world
• Pre-release review authority, as suggested by Trump's June 2, 2026, executive order requiring AI companies to provide models for government assessment before full release, could become a permanent feature of the regulatory landscape [17]

The Trump administration has not specified which rights it would seek. But the precedent of China's golden shares — where 1% stakes carry outsized governance power — suggests that even a small equity position can translate into significant control.

The Broader Pattern

The AI equity proposal does not exist in isolation. It is part of a broader pattern of Trump administration technology policy that includes:

• A December 2025 executive order identifying "excessive state regulation" as an obstacle to U.S. AI dominance and directing an AI Litigation Task Force to challenge state AI laws [18]
• A March 2026 national policy framework centering on preemption of state AI regulation and a "light-touch" federal approach [19]
• A June 2, 2026, executive order requiring AI companies to provide models to the government for capability assessment before release [17]
• Conditioning $42 billion in broadband infrastructure funding on states' repeal of AI regulations [18]

Taken together, these actions suggest a strategy of consolidating federal authority over AI while simultaneously acquiring ownership positions in the companies subject to that authority. Whether this represents smart industrial policy or a dangerous concentration of governmental and commercial power depends on how — and whether — the inherent conflicts of interest are managed.

Source: FRED / S&P Dow Jones Indices

Data as of Jun 4, 2026CSV

The S&P 500, currently at 7,584, has risen 27% year-over-year, driven substantially by AI-related stocks [20]. The government's financial interest in the continued growth of AI companies would align with the market's trajectory — but would also create powerful incentives to subordinate safety, competition, and public interest to share price.

What Comes Next

No legislation has been introduced, no executive order signed, and no companies have formally agreed to participate. The proposal remains at the discussion stage. But the fact that it is being discussed at all — by a president who already holds stakes in roughly 20 companies, in an industry valued in the trillions, at a moment when AI's strategic significance rivals that of nuclear technology — means the conversation has moved from theoretical to operational.

The questions that remain are not abstract: How much equity? At what price? With what governance rights? Subject to what oversight? And who decides? Until those questions are answered, the proposal represents less a policy than a provocation — one that has surfaced a genuine tension between the public's stake in AI's development and the risks of government entanglement in the companies building it.