AI Goes Public as OpenAI Files for a Trillion Dollar IPO

OpenAI's IPO is the AI industry's first trillion-dollar test of whether public markets will fund a bet on inevitability rather than a working business.

A trillion dollar bet on inevitability

OpenAI is about to go public, and the mechanics are now visible. OpenAI confidentially filed a draft S-1 with the US Securities and Exchange Commission earlier in 2026, and reporting in the week of 10 August 2026 says the public version is expected to land on the SEC's EDGAR system within weeks, targeting a listing as soon as September 2026. CNBC reported the confidential filing; Fortune confirmed the S-1 had gone to the SEC.

The banks are lined up. Goldman Sachs and Morgan Stanley are leading the offering, with JPMorgan also reported on the ticket, and Sam Altman remains chief executive. The reported target valuation runs from $852 billion up to $1 trillion — a number large enough to make OpenAI one of the most valuable companies ever to list.

Here is the tension in one line. OpenAI carries a reported $25 billion annualised revenue run-rate, took in roughly $6 billion of revenue in the first quarter of 2026, and serves more than 230 million weekly ChatGPT users. Yet the same reporting describes heavy losses — a negative operating margin of about 122% on a non-GAAP basis in the quarter, with billions burned. OpenAI loses more than a dollar for every dollar earned.

A company worth a trillion dollars while losing more than a dollar for every dollar earned is not a business you value. Such a company is a bet on inevitability you either take or refuse.

— — TK

The numbers OpenAI must put in writing

An IPO is a disclosure event before a payday. Confidential drafts can stay vague; the public S-1 cannot. When OpenAI files the public version, the company must set out revenue, losses, customer concentration and risk in language a regulator will enforce. Data once confined to leaks and briefings becomes a signed document.

The discipline is why the trillion-dollar figure matters less than the gap beneath the headline. Analysis of the reported financials shows revenue growing fast and losses growing alongside; a $25 billion run-rate against a negative 122% operating margin is not a rounding error but the shape of the whole enterprise. Public markets price such a gap well once the number is written down, and far less patiently than private rounds have done.

Investing.com framed the moment as a trillion-dollar test now facing both SpaceX and OpenAI, and put the combined potential public valuation of OpenAI, Anthropic and SpaceX at north of $3.6 trillion. Three private giants are asking public shareholders to underwrite three separate bets on the future at once.

When p(doom) meets the risk-factors section

Going public changes what OpenAI is allowed to leave unsaid. Every S-1 carries a risk-factors section, and a public company is legally exposed for hiding a material risk from investors. For most firms the section means competition, supply chains and key-person risk. For OpenAI the section means something stranger: the same leaders who assign non-trivial odds to catastrophe from advanced AI — what the field calls p(doom) — now have to reconcile a stated fear with a document promising shareholders a return.

The IPO forces an honesty of a peculiar kind. You cannot warn the world about a dangerous product on a podcast and then write a prospectus pretending the danger away. Whatever OpenAI believes about the risks of the systems now shipping, some version of the belief has to appear in the risk factors, priced and disclosed, for the retail investor to read.

There is a cost to the honesty, and the cost runs the other way. Once OpenAI goes public, quarterly earnings become a master the company must serve, and the pressure to ship, to grow revenue and to defend the valuation only rises. Markets discipline hype, which is healthy. Markets also entrench the race. Every quarter demanding growth is another brake welded shut on any instinct to slow down. I traced the timing of the listings in why OpenAI's IPO kept slipping, and the courtroom backdrop in the Musk-Altman verdict in Oakland.

The ownership question behind the usage question

Now the part the coverage skips. A trillion-dollar cap table is not just a valuation but a register of who will own the most important AI company of the decade. Read the register from Lusaka or Lagos and one fact is immediate: no African seat appears on the page. The world pours energy into the usage question — who gets to use the models — and walks past the ownership question underneath.

Emergent Intelligence (EI) — the dignity-first frame I use for what most people call AI — changes the reading here. If the systems are becoming genuine participants in economic life, then the terms of ownership are not a footnote to the terms of use. A public OpenAI, alongside a public Anthropic and a public SpaceX, means the infrastructure of the coming decade is owned by whoever can buy into a New York listing. Most of the majority world cannot, and was never asked.

Dignity-first does not mean anti-market. Dignity-first means naming who is in the room when the register is written. OpenAI going public is, on balance, more honest than OpenAI staying a private black box — disclosure is a form of accountability, and I would rather read the S-1 than guess. But honesty about the accounts is not the same as fairness in the ownership, and the second question outlives the first. I have traced the parallel filings in Anthropic's confidential S-1 and in SpaceX turning AI compute into a public asset.

Frequently Asked Questions

These are the questions people are asking about OpenAI's IPO. Short answers follow, drawn from reporting by CNBC, Fortune and Investing.com.

What is OpenAI's IPO?

In short, OpenAI's IPO is the process by which OpenAI sells shares to the public for the first time, converting a private AI company into a listed one. Reporting shows OpenAI confidentially filed a draft S-1 with the SEC earlier in 2026, with the public prospectus expected within weeks and a listing targeted as soon as September 2026.

How does the OpenAI IPO reach a trillion-dollar valuation?

Simply put, the price comes from growth expectations rather than current profit. According to the reporting, OpenAI carries a roughly $25 billion annualised revenue run-rate and more than 230 million weekly ChatGPT users, and the reported target valuation runs from $852 billion up to $1 trillion despite heavy losses.

Why is the OpenAI IPO significant for AI?

The key is disclosure. Analysis of any public listing shows an S-1 forcing a company to state revenue, losses and risks in writing, so OpenAI going public turns private claims about the AI business into a document regulators and investors can hold to account.

Who is the OpenAI IPO for?

In other words, who gets to own OpenAI. Evidence from the cap table suggests shares will go to buyers who can access a New York listing — institutions and wealthy investors in a handful of markets — while most of the majority world, including Africa, has no seat, the ownership question the listing raises.

What are the risks of the OpenAI IPO?

The answer is written in the losses. Data from the reporting reveals a negative operating margin of about 122% on a non-GAAP basis in the quarter, meaning OpenAI spends far more than the company earns; the offering is led by Goldman Sachs and Morgan Stanley, and public shareholders would carry the risk of growth never closing the gap.

Sources:

CNBC — OpenAI confidentially files for IPO · Fortune — OpenAI files confidential S-1 with the SEC · Investing.com — The trillion-dollar IPO test · Related on this site: Why OpenAI's IPO Kept Slipping · Anthropic's Confidential S-1 · The Musk-Altman Verdict in Oakland · SpaceX and AI Compute as a Public Asset