OpenAI Draws Its IPO Line: Public in 2027, and Don't Panic About Anthropic

The most important number in artificial intelligence right now might not be a parameter count. It’s a date on a calendar, and OpenAI just told its employees what it is.

At an all-hands meeting on Wednesday, August 19, OpenAI Chief Financial Officer Sarah Friar gave the company’s roughly 5,000 employees the clearest timeline yet for one of the most anticipated stock market debuts in history: OpenAI “will be a public company in 2027” — and potentially sooner if the business keeps compounding at its current pace.

Her message came with a deliberate dose of stage management. According to two sources familiar with her comments, Friar framed the listing not as an exit but as a financing event: “The IPO is not a finish line, it is a milestone, another fundraise.” She reminded staff that OpenAI raised a staggering $122 billion in committed capital in its March 2026 round, which closed at an $852 billion post-money valuation — giving the company, in her telling, no urgent need to rush to the ticker tape.

The elephant (and the rival) in the room

What made the all-hands unusual was how much of it was about someone else’s IPO.

Anthropic, OpenAI’s chief rival in the frontier-model race, confidentially submitted a draft S-1 registration statement to the SEC on June 1, 2026, and is widely reported to be targeting a listing as soon as September — literally next month. OpenAI followed with its own confidential filing a week later, on June 8, and reports have swirled for months that the ChatGPT maker could target a valuation north of $1 trillion, which would make it one of the largest IPOs ever executed.

Friar’s advice to employees worried about being beaten to the public markets: don’t be. She acknowledged there’s “a chance” Anthropic pulls off its listing first — and told staff that’s fine. Being second isn’t losing if your business is bigger, growing faster, and you don’t need the money yet.

That confidence rests on genuinely aggressive numbers. OpenAI’s annualized revenue run rate has now surpassed $40 billion, up roughly 35% quarter-to-date. In July alone, the company’s overall run rate grew 20% month over month, while business customers grew even faster at 32%, per investor slides viewed by CNBC. Just a week earlier, Friar had told employees that July’s ARR topped all of Q2 — a single month out-earning an entire quarter from the recent past.

The uncomfortable arithmetic

Of course, the same disclosures that showcase OpenAI’s growth also expose its costs. Audited financial documents obtained by Quartz show OpenAI posted a net loss of $38.5 billion in 2025 on $13.07 billion in revenue. The gap between a $40 billion run rate and an $852 billion private valuation — to say nothing of a $1 trillion public one — is the central question public-market investors will have to answer.

And Anthropic is no longer the plucky challenger. The rival’s revenue run rate has been reported surpassing $65 billion ahead of its own IPO. By EpochAI’s math, Anthropic generates roughly $9 million in revenue per employee, outperforming OpenAI’s $5.6 million and even Nvidia’s $5.1 million. Prediction markets, for what they’re worth, have priced Anthropic’s 2026 listing as a near-certainty, with one major platform showing 78% odds of a debut by December 31.

Why the timing dance matters

The sequencing of these two listings has become a genuine strategic chess match. Going first carries risk: the debutant sets the valuation benchmark, absorbs the market’s skepticism about AI economics, and hands its rival a free read on public-market appetite. Going second carries a different risk: if the first listing soars, the follower looks late; if it stumbles, the follower gets to recalibrate.

Friar’s framing suggests OpenAI has chosen option two — let Anthropic probe the defenses, then list into a market that has already priced the sector once. It’s the same playbook mature companies have used for decades, now applied to the most hyped asset class in a generation. The “2027 or sooner” hedge keeps the door open if Anthropic’s September debut triggers a land-grab for AI exposure and OpenAI decides to accelerate into the demand.

There’s also an internal-relations dimension. OpenAI’s employee compensation is heavily weighted toward equity, and the gap between an $852 billion private mark and a public-market verdict is, for many staff, the difference between paper wealth and real wealth. A CFO publicly committing to “2027” gives those employees something concrete to plan around — and gives recruiters at Anthropic, Meta, and Google a date to argue against.

What to watch

Three signals will determine whether “2027” holds. First, Anthropic’s September debut — its first-day performance and post-listing multiple will effectively beta-test the market for OpenAI’s file. Second, OpenAI’s S-1 going public, which will force the company to disclose revenue concentration, Microsoft revenue-share mechanics, and the true cost of its compute commitments in exhaustive detail. Third, the run-rate trajectory itself: at 35% quarter-to-date growth, OpenAI could enter a 2027 roadshow with an annualized revenue figure approaching $60–70 billion — a very different story from the $13 billion in booked 2025 revenue.

For now, the line has been drawn in sand that everyone acknowledges might shift: public in 2027, unbothered by whoever rings the opening bell first. In a year where OpenAI paused frontier training runs for safety reviews and Washington forced countries to pick AI blocs, a CFO calmly talking IPO milestones may be the most quietly confident signal the company has sent all year.