# OpenAI's trillion-dollar IPO, explained > OpenAI filed confidentially for what could be the largest tech IPO ever — at a valuation near $1 trillion, while losing more than a dollar for every dollar it earns. Here's what's actually going on. - **Pillar:** News - **Author:** Aditya Marin Gasga (Founding Editor) - **Published:** 2026-05-29T00:00:00.000Z - **Tags:** openai, ipo, ai business, chatgpt, valuation ## TL;DR OpenAI confidentially filed its IPO paperwork with the SEC in late May 2026, targeting a Q4 2026 listing at a valuation between roughly $850 billion and $1 trillion, with Goldman Sachs and Morgan Stanley leading. It would be one of the largest tech debuts in history — and one of the least profitable, with the company still losing more than it earns. ## Key takeaways 1. OpenAI filed a confidential S-1 in late May 2026, targeting a Q4 2026 listing (possibly as early as September) at a valuation near $1 trillion. 2. Revenue is growing fast — past a roughly $25 billion annualized run rate by early 2026 — but losses are growing alongside it. 3. A confidential filing keeps the detailed numbers private until about 15 days before the roadshow, so the real risk disclosures are still hidden. 4. OpenAI is racing to list before Anthropic, which is reportedly planning its own late-2026 IPO. 5. The structural question for investors: can a company spending on infrastructure at OpenAI's scale grow into a trillion-dollar price? import PullQuote from '~/components/article/PullQuote.astro'; OpenAI [confidentially filed for an initial public offering in late May 2026](https://www.cnbc.com/2026/05/20/openai-ipo-filing.html), and the numbers attached to it are the kind that stop being numbers and start being weather. A valuation near $1 trillion. A target to raise at least $60 billion. Goldman Sachs and Morgan Stanley on the deal. A listing window in the fourth quarter of the year, possibly as soon as September. If it lands as reported, it would be one of the largest technology debuts in market history. It would also be one of the least profitable. That tension — record valuation, record losses — is the entire story, and most coverage buries it under the trillion-dollar headline. This piece is about the gap between those two facts. ## What was actually filed A confidential S-1. That phrasing matters more than it sounds. A company filing confidentially submits its full prospectus to the SEC but keeps the contents private until roughly 15 days before its public roadshow. So the documents that contain OpenAI's audited financials, its margin structure, and — most importantly — its full list of material risk factors are written and submitted, but you can't read them yet. What's public is the framing: a Q4 2026 target, a valuation range [reported between about $852 billion and north of $1 trillion](https://fortune.com/2026/05/22/openai-ipo-filing-1-trillion-may-finally-answer-these-big-questions/), and the two underwriters. Everything else circulating right now is reporting and analyst estimate, not prospectus fact. Keep that distinction in mind, because the headline valuation is doing a lot of work that the actual disclosures haven't yet backed up. ## The revenue is real, and so are the losses The growth is not in dispute. OpenAI went from a roughly $2 billion annualized revenue run rate in 2023 to over $20 billion by the end of 2025, and reporting puts it [past a $25 billion annualized run rate by early 2026](https://fortune.com/2026/05/22/openai-ipo-filing-1-trillion-may-finally-answer-these-big-questions/), driven by paid ChatGPT subscriptions, enterprise contracts, and API usage. Enterprise reportedly accounts for a growing share of that — a healthier mix than a pure consumer-subscription story, because enterprise revenue tends to be stickier. The problem is what sits underneath it. OpenAI is spending to build the thing it sells — compute, data centers, research headcount — faster than the revenue arrives. Reporting indicated the company was losing more than a dollar for every dollar of revenue in early 2026, and internal projections cited in the press point to multibillion-dollar losses for the year. Sam Altman has publicly floated infrastructure spending plans measured in the trillions over time. You do not get to a trillion-dollar valuation on $25 billion of loss-making revenue through current cash flow. You get there on a bet about the future. ## Why go public now Two reasons, one offensive and one defensive. The offensive one is capital. An IPO gives OpenAI access to public markets and, just as usefully, public stock as acquisition currency — shares it can spend to buy companies without burning cash. For a business that needs to fund infrastructure at the scale Altman describes, both matter. The defensive one is timing. Anthropic, OpenAI's closest competitor and the maker of Claude, is reportedly preparing its own late-2026 listing. Being first to market as the marquee pure-play AI IPO is worth something — it sets the comparison, captures the institutional appetite, and frames the narrative before a rival can. OpenAI going early is partly a race to be *the* AI listing rather than one of several. ## What it means for the market The scale here isn't contained to OpenAI. A cluster of mega-IPOs — OpenAI, SpaceX, potentially Anthropic — could collectively demand far more from public markets in a single year than the entire US IPO market raised in 2025. When that much capital rotates into new listings, institutional portfolios have to sell something to buy in. That "something" tends to be existing large-cap tech. So even an investor who never touches an IPO share could feel this as a quiet headwind in positions they already hold. That's the underappreciated mechanic, and it's the part the trillion-dollar headline hides. A trillion-dollar listing isn't just one company's debut. It's a reallocation event. ## The question to actually ask Strip away the headline and the real question for any prospective investor is narrow: do you believe OpenAI grows into the price? At a near-trillion valuation against loss-making revenue, you are not buying current performance. You are buying a forecast — that AI demand compounds, that OpenAI keeps its lead, that the infrastructure spend converts into durable margins before the losses become the story. That's a defensible bet. It is not an obvious one. And until the confidential prospectus goes public and the risk factors are visible, anyone claiming certainty in either direction is working from the same incomplete picture as everyone else. The smart move between now and the roadshow is to watch for the moment those disclosures surface — because that's when the trillion-dollar number stops being weather and starts being math.