OpenAI's CEO and CFO Are Telling Different IPO Stories. That's the Point.
There is a precise division of labor emerging inside OpenAI as the company approaches what may be the most anticipated public listing in technology history. Sam Altman is the face of the product. Sarah Friar is the face of the finances. And on the subject of when or whether OpenAI might go public, they are not quite saying the same thing.
Altman has been pushing for a Q4 2026 listing, The Information reported. Friar, by contrast, has spent months telling anyone who will listen that an IPO is not imminent. "IPO is not on the cards right now," she said at the Wall Street Journal's Tech Live conference in November. She has also told associates the target is 2027, with some advisers suggesting late 2026 as a possibility. The divergence is almost certainly deliberate.
This is a standard pre-IPO communications playbook. The CEO projects confidence and momentum. The CFO manages expectations downward so that when the S-1 finally lands, the numbers look like a pleasant surprise rather than a disappointment. Friar is doing CFO work. Altman is doing CEO work. Neither is lying. They are performing different roles in the same theater.
What makes this moment different from a simple he-said-she-said is the underlying constraint driving the timeline debate. That constraint is compute.
The Information reported that Friar told ARK Invest in an interview released this week that OpenAI is currently "making some very tough trades" and turning down opportunities because we don't have enough compute. She said she spends significant time hunting for last-minute compute availability in 2026. Business Insider reported that the company has discontinued its Sora video app and pulled back on other initiatives specifically because compute allocation is the binding constraint.
This matters for the IPO timeline in a specific way. If compute limits suppress revenue growth in 2026 and 2027, the growth trajectory that would support a large public offering starts to look questionable. A CFO who lets a CEO's Q4 2026 narrative go unchallenged is a CFO who gets blamed when the numbers disappoint.
Meanwhile, Altman is operating on a different clock. He stepped down as chairman of Helion Energy, the fusion startup he backed for more than a decade, on March 23, a move that clears a potential conflict as OpenAI explores a large-scale power purchase agreement with the company. The next day, OpenAI discontinued Sora. These are the actions of a CEO stripping away distractions ahead of a public listing. Altman wants Q4 2026. The compute crunch may give him 2027 instead.
One concrete signal that the IPO is genuinely being prepared: OpenAI has arranged a $4.7 billion credit line from JPMorgan, Goldman Sachs, Citi, Morgan Stanley, and Wells Fargo, according to om.co. Those are not just lenders. Those are the banks that would underwrite the offering. A credit facility from your prospective underwriters is a relationship-building exercise that gets them familiar with your books before the mandate is official.
The company has also been included in several ARK Invest ETFs, giving retail investors indirect exposure to the private company before any public shares exist. Cathie Wood's funds will channel retail capital into OpenAI shares ahead of the IPO, creating demand and distributing the story in advance. This is a pre-IPO marketing mechanism that did not exist in this form five years ago.
The competitive pressure adds another dimension. Anthropic is also preparing for a Q4 2026 listing, targeting a $60 billion raise, according to techi.com, which cited Bloomberg reporting. OpenAI and its rival are effectively racing to be the first AI company to anchor a major index fund. Whoever lists first gets first access to trillions of dollars in passive capital. That race does not pause for compute shortages.
The Altman-Friar divergence is not a story about two executives who disagree. It is a story about what happens when a company with an $852 billion private valuation, $122 billion in recent fundraising, and a product that cannot currently be made fast enough to meet demand tries to go public. The CFO is lowering the bar. The CEO is clearing the runway. Both are right. The question is whether the compute shows up in time for Q4 2026.
