OpenAI confidentially filed its IPO prospectus with the SEC on Friday, May 22, 2026 — and if you were wondering whether Sam Altman’s bet on converting from a nonprofit to a for-profit would actually pay off, Goldman Sachs and Morgan Stanley just answered that question with their checkbooks. The two most powerful investment banks on Earth are underwriting what could become the largest technology IPO in history, targeting a public listing as early as September 2026.
This isn’t a company testing the waters. This is a company that raised its last private round at an $852 billion post-money valuation — with Amazon, Nvidia, SoftBank, and a dozen other deep-pocketed investors fighting for allocation — and decided the private market still wasn’t giving it enough capital to win the AI race. That should tell you everything about how expensive this war has become.
The Numbers That Made Goldman Say Yes
Here’s the thing most people are missing: OpenAI isn’t going public because it needs the money. It’s going public because the money it needs is so astronomically large that even the private markets — the same markets that happily handed it $40 billion in a single round — can’t keep up.
The company’s annualized revenue rate exited 2025 above $20 billion, according to CFO Sarah Friar. Full-year actual revenue for 2025 landed around $13.1 billion. That’s a company growing at a rate that makes Salesforce’s early years look pedestrian. But here’s the catch: OpenAI is still hemorrhaging cash on compute. Training GPT-5 reportedly cost north of $4 billion. The inference costs for 400 million weekly ChatGPT users aren’t exactly cheap either.
Goldman Sachs and Morgan Stanley looked at those numbers and saw what every banker dreams of: a company with explosive top-line growth, a clear path to profitability, and an insatiable appetite for capital that will generate fees for years. This is the kind of IPO that builds careers on Wall Street.
The Confidential Filing Is the Quiet Part Out Loud
OpenAI chose a confidential filing, which means the actual prospectus won’t be public for weeks. That’s standard for companies this size — it lets them negotiate with institutional investors behind closed doors before retail investors ever see the numbers. But the timing is deliberate.
SpaceX filed its public S-1 just two days ago, targeting a June Nasdaq listing at a $1.75 trillion valuation. Cerebras already went public this month at a $95 billion valuation, 20x oversubscribed. The IPO window is wide open, and OpenAI is sprinting through it before the market mood changes.
Think of it this way: the last time the IPO market was this hot for tech, it was 2021, and half of those companies are now trading below their listing price. OpenAI doesn’t want to be the company that waited too long. Neither do Goldman and Morgan Stanley — their fees on an IPO this size could top $500 million.
What This Means for the AI Power Structure
An OpenAI IPO doesn’t just change OpenAI. It reshapes the entire competitive landscape of artificial intelligence.
For Anthropic: The pressure just doubled. Anthropic hit $10.9 billion in revenue and posted its first-ever profit, but it’s still private. Once OpenAI is public with a war chest of fresh capital and public market accountability, Anthropic either matches the pace or risks falling behind on the infrastructure arms race. Expect an Anthropic IPO filing within 12 months.
For Google and Meta: The AI labs they’re competing against now have access to public market capital. That’s a structural disadvantage for in-house AI divisions that have to compete internally for budget. Sundar Pichai can’t issue stock to fund Gemini development the way a public OpenAI can.
For Microsoft: This is complicated. Microsoft owns a significant stake in OpenAI, which means it’ll book a massive unrealized gain when the IPO prices. But a public OpenAI is also a more independent OpenAI — one that answers to public shareholders, not just to Satya Nadella’s strategic vision. The exclusive Azure partnership is already loosening; an IPO accelerates that.
The $852 Billion Elephant in the Room
Let’s talk about that valuation. At $852 billion, OpenAI would be the seventh most valuable company in the world on day one — ahead of Berkshire Hathaway, ahead of TSMC, ahead of every company that actually makes physical things. And that’s just the private market valuation. IPO pops in this market have been averaging 30-40% on day one. That puts a realistic first-day market cap above $1 trillion.
For a company that was literally a nonprofit six years ago, that’s not just impressive — it’s one of the most dramatic corporate transformations in business history. Sam Altman didn’t just convert OpenAI from nonprofit to for-profit. He converted it from a research lab into a trillion-dollar weapon.
But here’s what should make investors nervous: OpenAI’s revenue, while massive, is heavily concentrated in ChatGPT subscriptions and API access. The advertising business — which was supposed to be the next growth engine — has seen CPMs collapse from $60 to $25 in ten weeks. The enterprise business is growing but still represents a small fraction of total revenue. And the competitive moat? Every major tech company and at least three well-funded startups are building comparable models.
The Verdict: This IPO Is a Bet on AI Itself
Here’s what Goldman Sachs and Morgan Stanley are really selling: not just OpenAI stock, but a proxy bet on the entire future of artificial intelligence. Retail investors who missed Nvidia at $20 will see OpenAI as their second chance. Institutional investors who underweighted AI will see it as their entry point. And the FOMO machine that drives every mega-IPO will do the rest.
The filing is confidential, the roadshow hasn’t started, and the pricing is months away. But the decision has been made. OpenAI is going public, and every other AI company on Earth just got put on a clock.
The question isn’t whether this will be the biggest tech IPO of the decade. It’s whether the company behind it can grow fast enough to justify a price tag that would make even the most bullish venture capitalist flinch. If it can’t, a lot of people are going to learn an expensive lesson about the difference between building the future and betting on it.
