Executive Summary
- OpenAI CFO Sarah Friar has, per The Information, said that OpenAI is not ready to go public in 2026, in part because of the "risks from its spending commitments" and not being sure whether the company's revenue growth would support its spending commitments.
- Friar (CFO) no longer reports to Sam Altman (CEO) and hasn't done so since August 2025.
- OpenAI's margins were lower in 2025 "...due to the company having to buy more expensive compute at the last minute."
News out of The Information's Anissa Gardizy and Amir Efrati over the weekend - OpenAI CFO Sarah Friar has apparently clashed with CEO Sam Altman over timing around OpenAI's IPO, emphasis mine:
She told some colleagues earlier this year that she didn’t believe the company would be ready to go public in 2026, because of the procedural and organizational work needed and the risks from its spending commitments, according to a person who spoke to her. She said she wasn’t sure yet whether OpenAI would need to pour so much money into obtaining AI servers in the coming years or whether its revenue growth, which has been slowing, would support the commitments, said the person who spoke to her.
I cannot express how strange this is. Generally a CFO and CEO are in lock-step over IPO timing, or at the very least the CFO has an iron grip on the actual timing because, well, CEOs love to go public and the CFO generally exists to curb their instincts.
Nevertheless, Clammy Sam Altman has clearly sidelined Friar, and as of August last year, the CFO of OpenAI doesn't report to the CEO. In fact, the person Friar reports to (Fiji Simo) just took a medical leave of absence:
In an unusual move for a large company, where CFOs almost always answer directly to the CEO, Friar stopped reporting directly to Altman in August last year and instead began reporting to Fidji Simo, who had joined as head of OpenAI’s applications business. Simo last week told staff she would take a short medical leave.
It is extremely peculiar to not have the Chief Financial Officer report to the Chief Executive Officer, but remember folks, this is OpenAI, the world's least-normal company!
Anyway, all of this seemed really weird, so I asked investor, writer and economist Paul Kedrosky for his thoughts:
Having been around this industry as an investor for decades, I cannot recall an example of the CFO of a major pre-IPO tech company (allegedly) taking issue with their own company's IPO plans. It doesn't happen.
Very cool! Paul is also a guest on this week's episode of my podcast Better Offline, by the way. Out at 12AM ET Tuesday.
Anyway, The Information's piece also adds another fun detail - that OpenAI's margins were even worse than expected in 2025:
In another sign of its financial pressures, OpenAI told investors that its gross profit margins last year were lower than projected due to the company having to buy more expensive compute at the last minute in response to higher than expected demand for its chatbots and models, according to a person with knowledge of the presentation.
Riddle me this, Batman! If your AI company always has to buy extra compute to meet demand, and said extra compute always makes margins worse, doesn't that mean that your company will either always be unprofitable or die because it buys too much compute?
Say, that reminds me of something Anthropic CEO Dario Amodei said to Dwarkesh Patel earlier in the year...
So when we go to buying data centers, again, the curve I’m looking at is: we’ve had a 10x a year increase every year. At the beginning of this year, we’re looking at $10 billion in annualized revenue. We have to decide how much compute to buy. It takes a year or two to actually build out the data centers, to reserve the data center.
Basically I’m saying, “In 2027, how much compute do I get?” I could assume that the revenue will continue growing 10x a year, so it’ll be $100 billion at the end of 2026 and $1 trillion at the end of 2027. Actually it would be $5 trillion dollars of compute because it would be $1 trillion a year for five years. I could buy $1 trillion of compute that starts at the end of 2027. If my revenue is not $1 trillion dollars, if it’s even $800 billion, there’s no force on earth, there’s no hedge on earth that could stop me from going bankrupt if I buy that much compute.
My Quick Take
It is extremely strange that the CFO of a company doesn't report to the CEO of a company, and even more strange that the CFO is directly saying "we are not ready for IPO" as its CEO jams his foot on the accelerator. It's clear that both OpenAI and Anthropic are rushing toward a public offering so that their CEOs can cash out, and that their underlying economics are equal parts problematic and worrying.
Though I am entirely guessing here, I imagine Friar sees something within OpenAi's finances that give her pause. An S-1 - one of the filings a company makes before going public - is an audited document, and I imagine the whimsical mathematics that OpenAI engages in - such as, per The Wall Street Journal, calculating profitability without training compute - might not match up with what actual financiers crave.
Enjoy This Piece? Subscribe To Premium
If you like this piece and want to support my independent reporting and analysis, why not subscribe to my premium newsletter?
It’s $70 a year, or $7 a month, and in return you get a weekly newsletter that’s usually anywhere from 5,000 to 18,000 words, including vast, detailed analyses of NVIDIA, Anthropic and OpenAI’s finances, and the AI bubble writ large. I just put out a massive Hater’s Guide To The SaaSpocalypse, as well as last week’s deep dive into How AI Isn't Too Big To Fail.
Supporting my premium supports my free newsletter.
