OpenAI is now a company, a startup, or what? The corporate structure of Sam Altman’s company is rather intricate, and it reflects the problems in the AI sector

Since its founding in 2015, OpenAI has promoted a rather cautious and transparent approach to the development of artificial intelligence – as the very name suggests. Originally, the company was structured as a non-profit organization and set itself the mission of developing “friendly” artificial intelligence, designed to generate benefits for all of humanity. This initial approach, however, came into conflict with what happened after the launch of ChatGPT. The tool’s success attracted enormous attention to the sector and kicked off an investment race that put OpenAI itself in an uncomfortable position, quickly turning it into the company that everyone in the tech industry was looking at.

In this scenario, in 2019 the company changed its corporate structure, abandoning the exclusively non-profit model to adopt a hybrid one that included a stronger for-profit component compared to its original mission. This was a significant change that effectively enabled further expansion and access to larger capital. A crucial step at this stage was the start of a collaboration with Microsoft, which in recent years has invested around $13 billion in OpenAI, consolidating a true commercial alliance. Last year, the Financial Times reported that the company was considering further changes to its structure to simplify and enhance its ability to raise investment. OpenAI initially denied the rumor, but shortly afterward the site The Information revealed that the company had attempted to raise additional funds, with the goal of reaching an overall valuation of around $100 billion.

It had long been known that OpenAI was pursuing contacts with investors and funds to secure new financial resources to be used in the long term. Faced with the growing number of rumors, the company then publicly stated that it was considering a reorganization to permanently move beyond its non-profit nature, thus freeing itself from some obligations under U.S. law. However, this possibility had sparked much criticism. Last May, the company led by Sam Altman announced that it had temporarily suspended the project of evolving into an entirely profit-oriented entity.

Today, OpenAI operates as a “benefit” corporation, that is, an organization that combines economic goals with a commitment to generate positive effects on society and the environment. The company has therefore maintained its non-profit division, while the other part of the organization has become a Public Benefit Corporation (PBC), which under U.S. law corresponds to a “public utility company”: this model allows the presence of investors and owners, and the distribution of dividends – that is, the share of revenue that the company chooses to redistribute to shareholders. Nevertheless, OpenAI’s non-profit section will remain the body responsible for appointing the members of the PBC’s board of directors and will continue to represent the main decision-making center of the entire structure.

According to OpenAI’s board chairman, Bret Taylor, the decision to review the company’s internal structure was necessary to ensure its survival in an increasingly competitive market. To develop new language models and at the same time support the operating costs of those already in use, OpenAI requires massive financial resources. Furthermore, as the number of players in the market grows and profit margins shrink, the required investments keep increasing, also because advances in language models entail ever higher costs – both in the initial development phase and in user deployment. OpenAI’s organizational changes should therefore be understood in light of this constant need for capital.

The search for investment, however, is not limited to Sam Altman’s company. In 2023, for example, Arthur Mensch, a former DeepMind researcher, together with Guillaume Lample and Timothée Lacroix, both from Meta’s artificial intelligence department, founded a new entity in Paris that today is considered the leading European company in the sector; in this case, the choice to establish themselves in France was not accidental: it was motivated by incentives and public investments introduced by Macron’s government in 2018, when the French president decided to focus on artificial intelligence through a series of targeted policies. Even today, Mistral emphasizes its European identity as a distinctive element to attract capital, but the overall picture of the sector shows how funding is never sufficient and how profit margins are still not enough to make AI a fully sustainable investment from an economic perspective.