OpenAI and Anthropic File for Public Listings: What You Need to Know

OpenAI submitted a confidential draft prospectus to the U.S. Securities and Exchange Commission on June 8, 2026, according to a disclosure on the company's website. The filing is a first formal step toward a public offering, though the company stated it has not yet decided on timing for next steps — a standard hedge that gives it flexibility while meeting procedural requirements.

The move came one week after Anthropic filed its own confidential prospectus with the SEC, announced on June 1, 2026. In eight days, the two leading independent AI companies both entered the public-offering pipeline. This kind of clustering in the AI sector hasn't happened at this scale before.

What a Confidential Filing Actually Is

When a company files a "confidential" draft prospectus under SEC rules — a process enabled by regulations designed to help emerging companies — it can submit draft documents for the SEC's review without releasing them to the public right away. The company works with SEC staff to refine the documents, and only makes the filing public at least 15 days before it starts talking to investors about the offering. The system was designed to keep sensitive financial information from leaking during the back-and-forth between the company and regulators.

For OpenAI, this means the investment world doesn't yet see the revenue figures, costs, profit margins, or detailed risks that will shape how investors price the company when it goes public. But the confidential filing does confirm something important: OpenAI has formally engaged with the SEC process, which requires audited financial statements and a near-final draft prospectus. The basic organizational groundwork must already be done.

It's worth noting that OpenAI filed a related SEC registration back on March 20, 2025 — suggesting the company's work with the SEC goes back more than a year, though that earlier filing was likely different in scope from a full initial public offering.

Anthropic Faces Similar Challenges

Anthropic's filing, a week earlier, provides a useful comparison point. Both companies have unusual corporate structures that make them harder to fit into the standard story investors expect. Anthropic operates as a public-benefit corporation — a legal form that ties the company to a public-interest mission alongside shareholder returns. OpenAI has been converting from a capped-profit structure (where investor returns are limited) to a more conventional public-benefit corporation. Neither company maps neatly onto how most software or AI infrastructure companies work.

For OpenAI, the structural overhaul it underwent to convert into a clearer legal form was essential before any IPO could credibly happen. A public offering requires straightforward capital structure, clear shareholder rights, and a standard equity instrument that institutional investors can hold and trade. The confidential filing suggests, without explicitly confirming, that OpenAI has worked out those prerequisites enough for SEC review to begin.

Anthropic faces a parallel situation. Its public-benefit corporation status creates a legal tension: it has obligations not just to shareholders, but to a stated public mission. Prospectus writers will need to disclose that tension explicitly in documents meant for investors.

Why Now?

The timing is driven by real business conditions. Investment in AI infrastructure — the computers, data centers, and hardware needed to build and run large AI models — has reached levels rarely seen in enterprise technology. Major tech companies are spending heavily on GPU clusters and computing capacity. More AI models are moving from early testing into real production use across different industries. OpenAI and Anthropic have both benefited from this demand through API revenue (when companies use their models), enterprise contracts, and consumer subscriptions.

Going public would give each company something private investors cannot: listed stock that can be used to buy other companies, to give employees a way to cash in shares as they leave, and to raise money on the balance sheet. In a business where computing costs are a major expense that eats into profits, having public-company flexibility matters. The window for AI companies to attract premium valuations from public investors probably won't stay open forever, and both boards appear to believe now is the time.

The broader pattern here bears noting. Back in the mid-1990s, the browser wars produced a similar rush of IPO filings. Netscape's 1995 offering set a price for the internet sector, and competitors raced to get public in the following eighteen months. Not all of them survived the crash that followed. Today's leading AI labs have vastly more revenue and stable customer bases than those early internet companies did, but the dynamic is structurally familiar: a landmark filing pulls competitors into the queue.

What Remains Confidential

Because both filings are still confidential, the most important data points remain hidden from public view. Investors and analysts don't yet know OpenAI's or Anthropic's actual revenue, what percentage of every dollar they earn becomes profit, how much it costs to run their AI models, which customers represent the biggest chunk of revenue, or the details of their ongoing deals with Microsoft and Google. People are working with old private fundraising valuations (OpenAI was valued above $300 billion in its last private funding round), public comments from the companies, and educated guesses based on their pricing.

OpenAI's explicit statement that it hasn't decided on timing is a meaningful caveat. Confidential filings can expire if not pursued; changes in market conditions, regulatory developments, or unresolved disclosure questions can all delay the move from confidential to public. Anthropic faces the same uncertainties. Neither has announced a roadshow date, a group of underwriters, or a preferred exchange.

What the two filings together establish is that the top-tier AI lab sector — the companies that build and run cutting-edge foundation models — is preparing to move from private to public capital. When that happens, these companies will face quarterly earnings reports, analyst questions, and legal accountability to shareholders in ways they haven't before.

For people working in enterprise AI, cloud infrastructure, or AI development, the eventual public filings will be significant reading. They will be the first comprehensive, audited, legally binding accounts of what it actually costs to build and run frontier AI at scale. That information, when it becomes public, will reshape how the industry estimates the economics of AI, where competitive advantages actually lie, and how much profit companies can realistically expect from deploying foundation models over time.

The filings have been made. The process is underway.