OpenAI proposes five percent state stake and Alaska-style fund, according to FT

The Alaska Permanent Fund is, according to the report, a long-standing model: it is fed by oil revenues and invests the US state's returns in stocks, whose dividends are paid out to residents. Another portion of the revenues flows into Alaska's state budget. According to the report, Altman sees a state stake as a way to improve relations with the US government, as happened with Intel after the US government took a ten percent stake in the chipmaker.

Background on the Intel deal: After Donald Trump initially sharply criticized Intel CEO Lip-Bu Tan and called for his resignation, he supported the chipmaker after the US government received a 9.9 percent equity stake in exchange for the release of billions in subsidies. According to the FT report, Altman cites this example as a blueprint for a similar arrangement with AI corporations.

Intel as blueprint

Through a five percent stake by the government of US President Donald Trump, the company wants to remove political hurdles, the Financial Times reported on Thursday, citing two people familiar with the matter. According to the report, the proposal would not be limited to OpenAI itself: according to information from the Financial Times (FT), such a stake would also include competitors like OpenAI rival Anthropic, which would significantly increase the reach of the model.

In addition to OpenAI, Anthropic has also expressed similar proposals in economic policy discussions. The AI company had brought a so-called "digital dividend" into play — a payout to US citizens to be financed by an AI tax. This payout to US citizens was to be financed by an AI tax. With both companies' planned IPOs in mind, such a dividend payment to the population would be a novel instrument of profit distribution.

Anthropic and the idea of a digital dividend model

Both OpenAI and Anthropic are preparing for IPOs, which would bring the companies large profits. OpenAI is currently valued at 852 billion US dollars, Anthropic at 965 billion dollars. Should the state hold stakes in both companies, it would participate directly in possible price gains — a model that critics interpret as the beginning of a creeping nationalization of key technologies.

However, according to the FT, the talks are still at a very early stage. „Diese sollen sich die Gespräche zwischen Altman und der US-Regierung noch in einer frühen, konzeptionellen Phase befinden. For the implementation of such plans, however, a corresponding law would be necessary, to which the US Congress must give its consent. Without such a legislatively anchored foundation, the proposal remains for now a discussion offering to the political level.

It also remains open whether other industry heavyweights would join in. It is unclear whether corporations such as Google or Meta would also be willing to give up shares in their companies. While the negotiations are focused on OpenAI and Anthropic according to the FT, other providers could balk, as a state stake could be interpreted as interference in corporate sovereignty.

Political fault lines: Sanders demands nearly 50 percent

Within the political debate in Washington, the visions differ significantly. Sanders advocates for a significantly higher share of nearly 50 percent in the AI companies. Altman himself, by contrast, is in favor of the roughly five percent mentioned in the FT report. US President Donald Trump had already brought a state stake in AI corporations into play the previous month, so that US citizens can share in the expected profits of the industry.

The rapid advances in artificial intelligence (AI) are causing concern for the US government. This concern is also reflected in concrete regulatory steps: OpenAI postponed the general release of the latest program version at the authorities' instruction. In addition, the US government has temporarily banned Anthropic from providing its most advanced AI models to users abroad due to possible national security risks.

Regulatory pressure on the AI industry

Critics, on the other hand, interpret Altman's proposal as an attempt to counter the growing political pressure arising from concerns about the construction of data centers as well as the possible effects of AI on jobs and cybersecurity. In this reading, the offered state stake would be less a generous offer than a tactical move to avert regulations and to safeguard the corporation in the long term.

The model of a sovereign wealth fund fed by stakes in strategically important companies is so far unusual in the US. Individual sectors — such as semiconductor production — have in the past been supported through subsidies and minority stakes. A fund, however, that is fed directly from the equity stakes of private AI corporations and distributed to the population would be a structurally new step.

It is also unclear how such a fund could be set up administratively. An obvious construction would be one in which the contributed shares are held in trust by a state agency and the returns — for example in the form of regular dividends or one-time bonus payments — are paid out to taxable US citizens. Comparable instruments exist internationally, such as in the form of Norwegian or Saudi sovereign wealth funds; however, there is so far no direct model for a broad dividend distributed to private individuals from AI corporate profits.

Open questions and outlook

The discussion comes at a phase in which US politics is increasingly relying on industrial policy interventions to keep strategic technologies in the country. With the proposal of a state stake in OpenAI and Anthropic, this debate is now shifting from subsidies and export controls to direct state participation in the profits of private AI companies.

Observers view the move as an indication that the relationship between the US government and the leading AI corporations is in a phase of reordering. The form that a possible stake ultimately takes would, according to this, depend significantly on the outcome of the conceptual talks, the stance of Congress, and the willingness of other industry giants to join such a model.

Until then, it remains open whether the proposal will become a concrete legislative draft or whether it will, as with comparable discussion offerings of the past, remain in the political discourse. It is awaited with tension how the US government and Congress will react to the move — and whether other AI corporations will follow the example of OpenAI and Anthropic.