What the Reporting Actually Says

According to Fortune, citing work by blogger and media critic Ed Zitron alongside reporting from the Financial Times, OpenAI's 2025 financial results look, in plain terms, rough. Neither outlet has published a full set of audited figures—OpenAI does not make those public—but the picture assembled from leaked data, investor briefings, and regulatory filings suggests the company's cost structure is outpacing its revenue growth by a meaningful margin.

That framing matters. OpenAI has raised capital at valuations that, at their peak, placed the company above $300 billion. When a company commands that kind of market confidence, its financials are implicitly expected to justify the trajectory. Reporting that they do not is not a minor footnote.

The Structural Problem With Reading OpenAI's Books

OpenAI occupies an unusual legal and financial position. It began as a nonprofit research laboratory and is currently mid-transition to a for-profit public benefit corporation (PBC)—a structure that carries fiduciary obligations to both shareholders and a stated public mission. That conversion, which is still subject to regulatory review in several U.S. states, means the company's balance sheet does not map cleanly onto standard corporate financial analysis.

Capitalization tables, debt obligations, and equity structures that would be routine disclosures for a public company are, in OpenAI's case, visible only to direct investors and, selectively, to regulators. For everyone else, the picture is assembled from secondhand reporting.

Why the Balance Sheet Matters Beyond OpenAI

OpenAI is not a bank, but its financial health has systemic characteristics that rhyme with one. Microsoft has committed tens of billions in compute and capital. SoftBank has made OpenAI a centerpiece of its Vision Fund strategy. Dozens of enterprise software companies have built products on OpenAI's API infrastructure, creating operational dependencies that would be painful to unwind.

When a company of that connectivity shows signs of financial strain, the question is not just whether OpenAI survives—it is what a disorderly deterioration would do to the companies and contracts downstream.

What Investors Should Be Asking

The core analytical tension is straightforward: OpenAI's revenue is growing, but so are its costs—primarily the compute infrastructure required to train and serve large language models (LLMs), the AI systems that underpin its products. Gross margin, the percentage of revenue remaining after direct costs, is the number that determines whether the business model is viable at scale. Without audited disclosures, that figure is unknown to the public.

Institutional investors who have participated in recent funding rounds have seen internal projections. Whether those projections are holding is a different question—and one the current reporting suggests may have a discouraging answer.

The Disclosure Gap

For a company that is, by most measures, among the most consequential technology businesses in the world, OpenAI's financial opacity is remarkable. Public benefit corporations are not required to file with the SEC unless they issue public securities. OpenAI has not done so. Until it does—or until a regulatory process compels disclosure—the balance sheet will remain, as Fortune's headline puts it, the most mysterious in business.

That is not a sustainable condition for a company seeking to be treated as critical infrastructure.