OpenAI Proposes Selling 5% Stake to U.S. Government, Valued at Over $40 Billion at $852 Billion Valuation

According to the Financial Times, OpenAI has initiated preliminary talks with the U.S. government, proposing to sell a 5% stake in the company. At the current valuation of $852 billion, this stake is worth approximately $42.6 billion. OpenAI CEO Sam Altman has held direct discussions with senior officials including Commerce Secretary Howard Lutnick and Treasury Secretary Scott Bessent. The talks are at a 'conceptual' stage, and any deal would likely require congressional legislation.

Background and Motivation

Altman's proposal is not a spur-of-the-moment idea. Since early 2025, OpenAI has been publicly discussing the concept of 'government equity':

• In April, it released a policy white paper proposing the establishment of a 'public wealth fund';
• In May, its foundation allocated $250 million to study the 'economic future of the AI era';
• In June, Anthropic filed for an IPO, followed by OpenAI, and Anthropic's CEO also proposed taxing AI companies and distributing the proceeds to the public.

OpenAI currently faces multiple regulatory pressures: Senator Bernie Sanders has introduced a bill requiring a 50% stock tax on large AI companies; the GPT-5.6 series has been delayed in phases due to White House requests; and with midterm elections approaching, AI's impact on jobs has become a bipartisan issue. Against this backdrop, selling equity is seen as a strategy to ease regulatory pressure and build a community of shared interests.

Details and Broader Ambitions

Altman's full vision consists of three layers:

1. Each of the top U.S. AI companies (OpenAI, Anthropic, Google, Meta) would contribute 5% equity;
2. These stakes would be pooled into a sovereign wealth fund, similar to the Alaska Permanent Fund;
3. The fund's returns would be distributed as dividends to all Americans.

However, there is currently no evidence that other companies are interested. Altman has suggested that all major AI companies join the same arrangement, but execution would be extremely challenging.

Reactions and Policy Divide

The proposal has drawn criticism from both the right and left:

• Right (e.g., free-market think tank R Street) argues that government equity is the first step toward White House control of AI, potentially leading to operational interference;
• Left (e.g., Senator Sanders) contends that the public should hold 50% equity, calling the 5% stake too small and noting that it would be passive (non-voting, non-interfering), offering limited public benefit;
• Safety advocates worry that government as a shareholder might relax AI safety regulations due to dividend interests.

Sanders' 50% proposal and Altman's 5% plan reflect two distinct approaches: the former emphasizes public control, while the latter focuses on profit-sharing.

Potential Impact

If the plan materializes, the government would gain an institutionalized window into AI company decisions, including model release schedules, safety assessments, data usage, and export controls. The development path of AGI might no longer be purely driven by technology and business; regulatory and public interest considerations would become hard constraints.

Previously, after the U.S. government took a 10% stake in Intel, the White House's attitude toward the company shifted significantly in favor, providing a precedent for government equity improving relations. OpenAI and Anthropic are preparing for IPOs, and offering equity now could lock in stable external relationships before going public.

Currently, the proposal remains at an early conceptual stage, far from a formal agreement. However, it marks a new dimension in the relationship between the AI industry and the government—the debate is shifting from 'whether to regulate AI' to 'whether the government should become an AI shareholder.'