The most anticipated stock market debut of the decade just got more complicated. OpenAI confidentially filed for an initial public offering in May 2026 — and within weeks, reports emerged that the company is now leaning toward pushing the listing into 2027. Here is what has actually been confirmed, what is still reporting and rumor, and what an OpenAI IPO would mean for the AI industry and the people who build on its models.
Key facts
- Confidential S-1 filed: OpenAI submitted a confidential IPO registration to the SEC in late May 2026, with Goldman Sachs and Morgan Stanley reportedly leading the deal.
- Original target: a listing as early as Q4 2026 — potentially one of the largest IPOs in history.
- Now reportedly delayed: in late June, multiple reports said OpenAI is leaning toward waiting until 2027, citing a slumping public tech market and SpaceX’s rocky post-IPO debut.
- The financials: annualized revenue has passed $20 billion (about $2 billion per month), but the company remains deeply loss-making — with reported internal projections of roughly $14 billion in losses for 2026 and profitability not expected until around 2030.
- Nothing is final: a confidential filing starts a process; the timing, valuation and even the decision itself can still change.
What actually happened
In early June, reporting confirmed that OpenAI had confidentially filed an S-1 registration statement with the US Securities and Exchange Commission — the formal first step toward going public. A confidential filing lets a company begin the regulatory review process privately, refining its disclosures before anything becomes public. Goldman Sachs and Morgan Stanley are reported to be leading the offering.
The move followed months of groundwork. Reports as early as January 2026 indicated OpenAI was preparing for a potential listing in the fourth quarter of 2026, in what bankers expected could become one of the largest public debuts ever, with speculation around a valuation in the hundreds of billions of dollars — figures that remain unconfirmed by the company.
Then came the delay reports
In late June, the story shifted. Reports from major outlets said OpenAI is considering delaying the IPO to 2027. Two reasons dominate the reporting: a broader slump in public tech markets, and the underwhelming post-IPO performance of SpaceX — whose record-setting debut was followed by a sharp slide, cooling enthusiasm for mega-listings. Prediction markets like Kalshi and Polymarket, where traders bet on the announcement timing, promptly repriced toward a later date.
It is worth being precise about what this means: the S-1 does not expire, and “leaning toward 2027” is not a formal postponement. OpenAI has not publicly committed to a date. The honest summary is that the paperwork is in, and the timing is now a market-conditions question.
The numbers behind the story
OpenAI’s financial profile explains both the excitement and the caution. On the growth side, CFO Sarah Friar confirmed in January that annualized revenue had passed $20 billion, up from roughly $6 billion in 2024 — the company says it is generating about $2 billion in revenue per month and growing faster at this scale than Alphabet or Meta did at comparable stages. Few companies in history have grown this fast.
On the other side sits the cost of the AI race. OpenAI remains heavily loss-making — reported internal projections suggest losses of around $14 billion in 2026 alone, driven by enormous compute and infrastructure commitments, with profitability not expected until around 2030. An IPO at this scale would effectively ask public investors to fund the most expensive infrastructure build-out in tech history.
Why OpenAI needs the money
The context is the capital war at the top of AI. Training and serving frontier models requires data centers, chips and energy on a scale that private funding rounds — even OpenAI’s record-setting ones — strain to cover. Its main rivals have structural advantages: Google funds Gemini from a trillion-dollar advertising business, and Meta funds Llama from its social empire. Anthropic has deep-pocketed backers in Google and Amazon. Public markets are one of the few pools of capital large enough to keep pace. That is the strategic logic of an IPO — and also why the timing matters so much: a weak debut would ripple through the entire AI funding ecosystem.
What it means for people who use AI models
If you build on AI APIs or follow the model race, the IPO story is more than financial theater:
- Pricing pressure cuts both ways. Public-market scrutiny of losses could push OpenAI toward better margins — which can mean higher API prices or more aggressive upselling. At the same time, ultra-cheap open-weight rivals like DeepSeek V4 keep a lid on what anyone can charge. Our open-vs-closed cost study found open models are roughly 16× cheaper on average — a gap public investors will be watching too.
- Transparency arrives. An IPO forces detailed public disclosure of revenue, costs and risks. For the first time, everyone will see the real economics of frontier AI — data that today exists only as leaks and estimates.
- The competitive map hardens. A successful listing would hand OpenAI a war chest for compute and talent; a delayed or weak one would embolden rivals. Either way, the model landscape — tracked in our AI models database — is shaped by who can afford the next training run.
For teams budgeting AI spend, none of this changes today’s decision math: current API prices are what they are, and you can compare every major model’s real monthly cost with our AI API cost calculator.
The structural backstory
An OpenAI IPO would cap one of the strangest corporate journeys in tech. The company began in 2015 as a nonprofit research lab, added a “capped-profit” arm in 2019 to raise capital, and has spent recent years restructuring toward a more conventional for-profit model precisely so that outside investment — and ultimately public listing — became possible. Its deep partnership with Microsoft, which has invested well over $13 billion and holds extensive commercial rights to OpenAI’s models, adds another layer: public filings would finally spell out how revenue, compute commitments and model rights are actually divided between the two. That transparency alone makes the eventual S-1 one of the most anticipated documents in the industry — for competitors and customers as much as for investors.
Frequently asked questions
Has OpenAI officially gone public? No. It has confidentially filed the paperwork (an S-1) that starts the process. No shares are trading, and no date is set.
When is the OpenAI IPO expected? Originally as early as late 2026; recent reports say the company is leaning toward 2027. Nothing is confirmed.
What would OpenAI be worth? Speculation has centered on a valuation in the hundreds of billions of dollars, but no official range has been disclosed.
Can I buy OpenAI stock now? Not on public markets. Until an IPO completes, OpenAI equity is held by employees and private investors.
Why might the IPO be delayed? Reports point to slumping tech markets and SpaceX’s disappointing post-IPO performance, which cooled appetite for mega-debuts.
The bottom line
OpenAI’s IPO is no longer a question of if — the filing exists — but of when and at what price. The company brings historic revenue growth and equally historic losses to the table, and the timing now hinges on market conditions it cannot control. Whenever it lands, it will be a defining moment for the AI economy: the first time the market gets to put a hard, public number on the frontier-AI business model. We will update this story as facts are confirmed.
Reporting compiled July 2, 2026 from CNBC, Forbes, Reuters and company statements. Unconfirmed details are labeled as reported; figures may change as OpenAI’s filings become public.