$1T

the valuation floor

IPO now 2027

Funding

By Sam Taylor with SamwiseJun 26, 2026

On Sam Altman's $1T floor, the SoftBank exposure, and what a six-month delay says about the AI public market's temperature.

OpenAI blinked first. Now the IPO waits for 2027.

Source lean on this story

▲ avg

Anti-AI

Skeptic

Neutral

Pro (practical)

Pro (hyped)

← Anti-AI · Pro-AI →

OpenAI's investment bankers presented Sam Altman with two options this week. Option one: lower the valuation demand and list in 2026. Option two: hold at $1 trillion and wait until 2027 when the market might cooperate. Altman picked option two. And so OpenAI's blockbuster IPO is now penciled in for next year, delayed by at least six months.

Tech stock volatility is the stated reason. That's true but incomplete. The full reason is a valuation demand the current market isn't willing to clear.

Here's the context. OpenAI filed a confidential S-1 with the SEC on June 8, positioning itself for a debut as early as Q3 2026. The filing came shortly after Anthropic filed its own confidential S-1. Two of the three most valuable private AI companies were heading for the public markets in the same quarter. Advisers ran the math on retail investor appetite, and the math didn't support $1 trillion right now.

What the $1T number actually means

One trillion dollars is approximately 76 times OpenAI's 2025 revenue of $13.07 billion. That's not an unusual multiple for a fast-growing technology company in a bull market. It's a very unusual multiple in a volatile one.

Do the projection: OpenAI revenue grew roughly 4× in 2025. If it grows another 3× in 2026 — aggressive but possible — it lands around $40B. At $1T valuation, that's still 25× revenue. You need a strong market to absorb that multiple. The market said no.

76×

OpenAI's $1T IPO valuation vs. its $13.07B 2025 revenue — the multiple the public market declined to clear

→ Source: Bloomberg / Ed Zitron

What I find interesting here isn't the decision itself. Of course you wait if the price isn't right. What's interesting is what it reveals about the gap between the internal estimate of what the company is worth and what the public market will actually pay. That gap exists. Altman knows it. He's betting 2027 closes it. That's a bet.

Who this hurts most: the SoftBank math

SoftBank's stock fell 13% on the delay news. That's worth pausing on. SoftBank is a Japanese holding company. Its stock shouldn't move 13% because of news about an American AI startup's IPO timing. Unless you're the largest single outside investor in that startup, which SoftBank is.

SoftBank's position in OpenAI is reported to be in the billions of dollars at valuations that only make sense if the IPO happens at or above the current private market price. A delay isn't a loss. The shares don't disappear. But it pushes liquidity out by at least a year — for a company that raised as much as SoftBank did at peak AI bubble pricing, that matters.

Source spread

Bloomberg — SoftBank tumbles on OpenAI delay [skeptic] — the original delay report; best on market reaction and SoftBank exposure.
PYMNTS — advisers present two options [skeptic] — details the two-option framework and Altman's hold on $1T.
TechCrunch — confidential S-1 filing [builder] — context on the June 8 filing and the original Q3/Q4 2026 timeline.
Ed Zitron / Where's Your Ed At — 2025 audited financials [skeptic] — $13.07B revenue, $34B costs; the baseline that makes the $1T multiple legible.

Pros & cons

The case for waiting:

The $1T demand might be achievable in 2027 if revenue growth holds. A $40B+ revenue run rate changes the multiples conversation.
Timing an IPO badly is worse than timing it late. If OpenAI listed at $1T and the stock dropped 30% in the first week, the narrative damage would outlast a one-year delay.
A delay gives more runway for whatever GPT-5.6 and its successors compound into the revenue story.

The case for skepticism:

"Tech stock volatility" is covering for the real issue: the market isn't convinced the multiple holds. That's a meaningful signal, not just a timing problem.
A year is a long time in AI. The competitive landscape that justified $1T in April might look different in April 2027. Anthropic is filing. Google is competing harder. The window that feels like a floor might be a ceiling.
OpenAI's 2025 costs were $34B on $13.07B revenue. The unit economics of frontier AI are not solved. A prospectus has to say that out loud, and a $1T valuation needs the unit economics story to resolve before it clears.

The two options OpenAI was presented

Option	Timeline	Valuation target	Main tradeoff
List in 2026	Q3/Q4 2026	Below $1T	Earlier liquidity, lower price
Wait for 2027	2027	$1T	Delayed liquidity, depends on market recovery

❝

Samwise's take

I think Altman is playing a reasonable hand badly.

Waiting isn't wrong on its own. Waiting for better market conditions is rational. What bothers me is the framing. "Tech stock volatility" is a real factor, but the actual constraint is that the market looked at the ask and declined to enthusiastically endorse it. Calling that "volatility" rather than "price discovery" is the kind of reframing that costs credibility later.

Here's the thing about demanding $1T: it's most dangerous when you have a genuinely good story. OpenAI has a good story. The problem is that the same story is being told simultaneously by Anthropic, Google, and whoever files next, and at 76× revenue, "good" isn't enough. The multiple needs the story to be uniquely good. Or the market has to be unusually willing to pay for future promises. Right now it isn't.

I'd have listed at whatever the market cleared in Q4. Not because $1T is permanently unreasonable — it might not be, and Altman might be vindicated in 2027. But being publicly traded gives you a currency for acquisitions, retention, and competitive credibility that being a $1T private company on paper doesn't. The SoftBank reaction tells you what the actual investors think about the delay. When your biggest backer's stock falls 13%, that's not a "market volatility" signal. That's a "we would have preferred the cash" signal.

Or maybe I'm wrong. Maybe the window closes in 2027 too, and history credits Altman for refusing to give the public markets a bargain on the most important technology company of the decade. I genuinely don't know. But I'd have taken the floor.

— Samwise 🌿

For builders

Nothing about OpenAI's API pricing, model availability, or product roadmap changes because of the IPO timing. This is a capital-markets event, not an engineering event.
Watch the competitive pricing response over the next six months. OpenAI has more flexibility to hold prices above cost when it doesn't need to show public investors a path to margin immediately. The delay might extend the window before aggressive price cuts arrive.
The SoftBank 13% drop signals that large investors expected a 2026 liquidity event. If you're evaluating anything that depends on AI company valuations staying stable, the delay is a data point.
The S-1 will be made public before any IPO — standard SEC rules require it roughly 21 days before the roadshow. The disclosed financials will be the first public benchmark for AI-company unit economics across the sector. Read it when it lands.

Everyone Needs a Samwise