Two sets of Numbers
The trial's in recess. Musk has, shockingly, followed the judge's orders so far. The prediction markets are not impressed but the lawyers appear to be winning for now.
The trial opened Monday. Tuesday evening the Wall Street Journal published a story about missed revenue targets. Friday the Journal dropped a full CFO profile. Saturday the rest of the press caught up.
Musk was on the stand the entire time.
The Wall Street Journal broke this in two parts — Tuesday's revenue miss story and Friday's full CFO profile — both paywalled at the source but covered freely elsewhere. OpenAI's CFO Sarah Friar wants to delay the IPO from 2026 to 2027. Her reasoning: the company isn't yet ready to meet the rigorous reporting standards required of public companies. She has also, according to the Journal, been frozen out of financial planning meetings for raising concerns about whether the spending commitments are supportable.
If that second part sounds familiar it's because I wrote about it. In Burn It Down. Months ago. Before the trial. Before the CFO went to the Journal. Before any of this was sworn testimony.
That post wasn't a prediction. It was just paying attention to the timing.
Here's the timeline as it actually stands.
Altman once called advertising a "last resort." It hit $100 million in annual recurring revenue within six weeks of launch. The CFO who closed the largest private fundraising round in history got frozen out of financial planning meetings for asking whether the math worked. She now reports to the head of applications instead of the CEO. The same week that came out Altman published a 13-page manifesto about saving capitalism. The same morning that dropped Ronan Farrow published Ilya Sutskever's confidential memo — first word of his concerns: "Deception" — and Dario Amodei's 200-page private notes titled "The problem with OpenAI is Sam himself" and Microsoft executives on record saying he "distorts, twists, renegotiates, and violates agreements."
Now the CFO is in the Wall Street Journal during trial recess saying the company isn't ready for public reporting standards.
And then there's this: the Journal reports OpenAI has been telling banks its revenue targets are different from those its investors know about. The same week Friar quietly walked back Altman's public claim of $1.4 trillion in computing commitments — the actual planned spend, she told investors, is $600 billion through 2030. Two sets of revenue numbers. Two sets of spending numbers. And ChatGPT, which OpenAI told investors would hit one billion weekly active users by end of 2026, has been stuck at 900 million since February with growth flattening.
Two sets of numbers. One for the banks. One for investors. During active IPO preparation. While the CEO is simultaneously being sued over the company's financial structure in federal court.
The SEC has very specific opinions about material misrepresentation to investors. Those opinions tend to arrive in the form of enforcement actions.
And then Altman and Friar jointly denied the internal friction publicly. Called the Journal report "ridiculous." Said they were "totally aligned." Morningstar's analyst noted dryly that when a CEO and CFO feel compelled to publicly assert strategic alignment, the report is usually close to accurate. The denial is more informative than the original story.
This is the part where Musk's lawyers get a gift they didn't have to go looking for.
They don't need to subpoena the CFO. They don't need a deposition. They don't need a leak. The CFO of the company their client is suing just told the Wall Street Journal — on the record, during active litigation — that the company isn't ready for public reporting standards and has been operating with two sets of revenue numbers.
It's in the paper. With her name on it.
Even if the judge rules it inadmissible as direct evidence the jury heard the question. Nine people from the greater Bay Area who sat through voir dire already know this trial exists. You can instruct a jury to disregard. You cannot un-ring a bell.
And Musk's lawyers don't need it admitted as evidence anyway. They need the timeline. If you've been following this series the timeline isn't new. It's been documented here stop by stop. The CFO going to the Journal is just the latest entry.
There's a comparison that keeps getting more accurate as this week develops.
Theranos had a charismatic founder with a reality distortion field. A product that worked differently in demos than in practice. A culture where asking hard questions was career limiting. People who raised concerns quietly documenting themselves out of liability before things fell apart.
Elizabeth Holmes didn't go to prison for building a bad product. She went for telling investors the product worked when she knew it didn't.
OpenAI's technology genuinely works. The product is real. Which makes the financial structure around it more interesting not less. You can build something genuinely useful and still construct a misleading investment case around it.
The CFO just built her exit ramp. She went to the Journal instead of the boardroom. She documented herself out.
Call it TheranasAI. If that lands you already know where this is going. Holmes didn't get indicted the day the Journal story dropped either. That post is coming. So, eventually, might other things.
The SEC doesn't need to go looking for any of this. It's in the Wall Street Journal — and Gizmodo if the paywall's in the way. Named sources. A spokesman quote that confirmed the two sets of numbers without apparently realizing how that sentence sounds. A joint denial that made things worse.
Altman testifies this month. Brockman too.
The CFO went to the Journal. Musk's lawyers read the Journal. The judge has opinions. The jury has ears.
Read the terms. They're more honest than the marketing.