According to several reports, OpenAI is reportedly targeting an initial public offering as early as September 2026, with the filing of its public prospectus expected in the coming weeks. It would be one of the largest tech IPOs ever. And above all, it would be the first time the public sees the real accounts behind ChatGPT.
For months, we have been documenting questions about the AI economy, from financing structures around data centers to Microsoft's results. Here is the event that will provide the most answers.
Why this document changes everything
A private company communicates what it wants. A company going public must file a detailed prospectus, under legal liability, containing information that no voluntary communication provides.
Concretely, four things previously unknown should come to light.
The real cost of inference. We explained in our article on the two costs of AI that the operating bill now exceeds training costs at established players. No one knows the real margins on each query.
The revenue breakdown. What share comes from consumer subscriptions, enterprises, the API? This information determines the strength of the business model far more than total revenue.
Off-balance-sheet commitments. Data center leases, compute purchase agreements, guarantees. This is where the risks we mentioned regarding cross-financing are lodged.
Risk factors. A prospectus contains a section where the company must honestly list what could go wrong. It is often the most instructive part of such a document.
Once listed, OpenAI will have to publish quarterly results. The entire sector will then have, for the first time, a verified reference point on the real economics of a leading AI lab. Today, debates about the bubble rest on estimates and statements. Tomorrow, they will rest on audited accounts. That is a change in the nature of the conversation.
The bet, and what it assumes
Going public at this stage is a strong strategic choice, and it cuts both ways.
The favourable reading: the company needs substantial capital to fund its infrastructure, and public markets offer a depth that private fundraising no longer provides. A listing also validates maturity.
The cautious reading: going public imposes a quarterly discipline that sits poorly with investments whose returns are measured in years. A listed company that misses its targets two quarters in a row faces pressure that a private company does not know.
Added to this is a governance question we had raised: OpenAI had proposed ceding a stake in its capital to the US government, as we mentioned in our comparison of strategies. How this arrangement squares with a public listing is one of the points the prospectus will need to clarify.
What to take away
This event deserves attention not for the stock performance, which only interests investors, but for the information it will make public.
For three years, we have discussed the AI economy from fragments: statements, estimates, figures released without audit. A prospectus, then quarterly results, will turn this speculative debate into documented analysis. Whatever the outcome, finally knowing what a generated response really costs will be the most useful information of the year.