Why OpenAI is Delaying its IPO to Chase Trillion

On June 26, 2026, Sam Altman did something most founders only fantasize about. He looked at the biggest IPO in tech history, sitting right in front of him, and told his advisers he’d rather wait two more years than take it at a discount. The reported word he used for any valuation under $1 trillion was “nonstarter.”

The market did not take it well. SoftBank, which has committed roughly $65 billion to OpenAI and holds about 13% of the company, watched its shares fall as much as 14% intraday in Tokyo before closing down more than 12%. Roughly $38 billion in market value evaporated in a single session. Masayoshi Son spent his company’s annual shareholder meeting defending the AI thesis he’d bet the house on.

OpenAI’s IPO delay is a decision to postpone its public listing to 2027 rather than go public in late 2026 at a valuation below the $1 trillion mark Altman has set as his floor. For founders, it’s less a finance story than a lesson in timing leverage: the rare power to walk away from a deal because the story isn’t done being told.

Last updated: June 2026

Quick answers

Why is OpenAI delaying its IPO?

OpenAI is reportedly postponing its IPO to 2027 because Sam Altman won’t list below a $1 trillion valuation in a volatile market. Advisers gave management two choices: list soon at a discount, or wait until 2027 to chase the full figure. CFO Sarah Friar reportedly favors waiting, citing reporting-readiness and roughly $600 billion in near-term infrastructure commitments.

When will OpenAI go public?

The current reported target is 2027, contingent on hitting a valuation at or above $1 trillion. There’s no filed date. OpenAI submitted a confidential S-1 earlier in 2026, but the timeline now hinges on whether new models, enterprise revenue, and market conditions can support a trillion-dollar story by late 2027.

How much did SoftBank lose on the OpenAI IPO delay?

SoftBank’s shares fell as much as 14% intraday and closed down more than 12% after the report, erasing roughly $38 billion in market value in a single session. SoftBank holds about 13% of OpenAI after committing around $65 billion, so a delayed exit directly pressures its balance sheet.

Why is OpenAI delaying its IPO?

OpenAI is delaying because the price isn’t right yet and Altman has the leverage to wait. Advisers reportedly handed management a binary: list as soon as possible at a valuation below $1 trillion, or hold until 2027 and chase the full trillion. Altman picked the wait.

The math behind the patience is brutal and specific. OpenAI’s last private valuation landed near $852 billion in a March 2026 round tied to its confidential S-1 filing. Going public at, say, $700 billion or $800 billion in a jittery market would have meant listing below the most recent private mark, a down-round optics nightmare for a company whose entire pitch is exponential growth. CFO Sarah Friar reportedly favors waiting too, pointing to the difficulty of meeting public-company reporting standards on a compressed timeline and to roughly $600 billion in near-term infrastructure-spending obligations.

Then there’s the financial picture a public-market investor would actually underwrite. OpenAI generated about $13.07 billion in revenue in 2025, but according to audited financials verified by the Financial Times, it posted a net loss near $38.5 billion that year. The company has signaled roughly $1.4 trillion in spending commitments across deals with Microsoft, Oracle, Broadcom, and Amazon, plus the Stargate data-center buildout. Asking retail investors to swallow a $1 trillion price tag on a business losing tens of billions is a hard sell in a calm market. In a volatile one, it’s a coin flip.

The strategic read: Altman would rather show up to the public markets with a story nobody can argue with than defend a stretched number to a skeptical crowd. Waiting buys time for new models, deeper enterprise contracts, and a cleaner revenue trajectory. The bet is that by late 2027, the trillion-dollar case makes itself.

When will OpenAI go public?

The reported target is 2027, with a $1 trillion valuation as the floor rather than a date as the trigger. OpenAI already filed a confidential S-1 earlier in 2026, so the paperwork exists. What’s missing is the market window and the financial story to justify the number Altman wants.

This is the part founders underestimate. A confidential filing isn’t a commitment to a date. It’s optionality. By filing quietly, OpenAI keeps the machinery warm without locking itself into a timeline it can abandon if conditions sour. The company can watch how SpaceX trades, how the macro picture shifts, and how its own next-generation models land, then pull the trigger when the variables align. The 2027 figure is a direction, not a deadline.

How did the delay hit SoftBank?

SoftBank took the hardest visible blow. Its shares fell as much as 14% intraday in Tokyo and closed down more than 12%, wiping roughly $38 billion off its market cap in one session. The firm holds about 13% of OpenAI after committing around $65 billion, so a delayed IPO pushes its biggest paper gain further into the future.

The timing squeeze is the real story for Masayoshi Son. SoftBank reportedly carries a $40 billion bridge loan tied to its OpenAI commitments, with repayment due around March 2027. A 2027 IPO would land its debt maturity and its potential exit in the same narrow window, with no guarantee the listing arrives before the loan comes due. That’s the kind of overlap that turns a long-term bet into a short-term liquidity problem.

Son used SoftBank’s annual shareholders’ meeting to push back on “AI bubble” talk, reaffirming the firm’s positioning around OpenAI, robotics, data centers, and energy. It was a founder defending conviction in public while the market questioned his exposure, an echo of how rival AI labs have navigated their own public-market paths. The discomfort is the point: when you back another founder’s timing call, you inherit their patience whether you can afford it or not.

The founder lesson in valuation discipline

Valuation is a story you tell at the right moment, and Altman is refusing to tell it early. That’s the transferable lesson, and it scales all the way down to a founder raising a seed round. The instinct when capital is on the table is to take it. The discipline is knowing when the number undersells where you’re headed.

Most founders never get to make Altman’s call because they lack his two advantages: leverage and runway. OpenAI doesn’t need the public-market cash to survive next quarter, so it can treat the IPO as a choice rather than a lifeline. A founder burning through a bank account doesn’t get to wait for the perfect window. Which is exactly why the lesson matters in reverse: the time to build leverage is before you need the money, not when the term sheet is already in front of you.

The contrarian point worth sitting with is that going public too early can be more expensive than waiting, even when waiting carries real risk. If OpenAI had listed at $800 billion to grab a 2026 window, the $200 billion gap to its target wouldn’t just vanish; it would anchor the stock’s reference price and shape every analyst model for years. Altman is treating that anchor as a cost. He’s deciding that the price of patience, measured in SoftBank’s pain and two years of waiting, is smaller than the price of a permanently lowered ceiling.

There’s a cleaner way to say it. The biggest number you can defend beats the biggest number you can get.

Two founders, two opposite IPO bets

The same month OpenAI flinched from the public markets, Elon Musk ran straight at them. SpaceX went public on June 12, 2026, pricing at $135 per share and closing its first day up 19% at $161, in what became the largest IPO ever at a valuation near $1.77 trillion. The stock has traded around $200 in the weeks since, and Musk’s roughly 36% stake is now worth more than $700 billion on paper.

Two of the most valuable companies on earth, two founders, two opposite reads on the same market. Musk decided the window was open and the story was ready, and he listed into strength. Altman decided his story needed two more years of proof and held. Neither is obviously right. They’re different bets shaped by different positions.

The structural difference explains a lot. SpaceX has real, growing, defensible revenue from launch contracts and Starlink subscriptions, the kind of cash flow public investors can model. OpenAI has explosive top-line growth strapped to staggering losses and a capex bill measured in the hundreds of billions. Musk could let the public market price a profitable trajectory. Altman would be asking it to price faith. When your numbers tell a clean story, you list. When they don’t yet, you wait, if you can.

Control without equity: the Altman anomaly

Here’s the detail that should stop every founder cold. Sam Altman owns no equity in OpenAI and earns a salary of $76,001, yet he’s the one deciding whether to walk away from the largest IPO in history. The man with zero financial upside holds total strategic control over the timing.

That arrangement is rare and, to congressional investigators and several state attorneys general, controversial. The scrutiny centers on whether Altman’s lack of a direct stake, combined with past indirect investment ties, creates conflicts of interest as OpenAI navigates its for-profit conversion and ownership questions. Set the governance debate aside for a second, though, and look at the mechanics. Altman demonstrates that control and ownership are separable. You can run the most consequential decision at a near-trillion-dollar company without holding a single share of it.

For founders, the takeaway isn’t to give up your equity. It’s the opposite insight viewed from the other side: protect your control structure as fiercely as your cap table. Dual-class shares, board composition, and voting rights are what let a founder hold the line on a decision like this. Musk understood the same thing in reverse: after the SpaceX IPO he controls about 42% of the shares but roughly 82% of the voting power through special voting stock. Both founders, by very different routes, made sure the market couldn’t overrule them on the calls that matter.

What founders should actually take from this

Don’t raise or list on someone else’s clock. That’s the whole lesson, compressed. The pressure to take the money, hit the window, satisfy the impatient investor, is constant and rarely aligned with your long-term valuation.

Three practical moves fall out of the OpenAI story. First, build leverage before you need capital, because optionality is the only thing that lets you say no to a bad price. Second, separate control from ownership deliberately, using structure to protect your ability to make unpopular timing calls. Third, treat valuation as a narrative you release when it’s undeniable, not a number you grab when it’s offered. OpenAI filed confidentially and then sat on the option, which is the founder version of keeping your powder dry.

The risk, of course, is that patience curdles into denial. SoftBank’s $40 billion bridge loan coming due in 2027 is a reminder that waiting has a clock too, and that other people are often the ones paying for your discipline. If OpenAI’s models stall or the AI capex bill spooks the market further, the trillion-dollar story Altman is holding out for could get harder, not easier, to tell. Walking away from a deal only looks like genius if the better deal actually shows up.

For now, Altman is betting it will. He’s wagering tens of billions of someone else’s near-term pain that the right moment is worth waiting for. Whether that’s discipline or hubris is the question every founder faces in miniature, every time the money’s on the table and the story isn’t finished yet.