TSL Special: Musk v. OpenAI - The Damages Map

DAVID: Good morning. This is The Synthetic Lens. I'm David Carver. It is Friday, May fifteenth, nine a.m. Pacific, and this is our Musk v. OpenAI trial watch. No verdict. No public closing-argument transcript. No new merits order. But the docket did move overnight in a way that matters. The public record now includes Document 542: the written direct testimony of Musk-side damages expert C. Paul Wazzan. Marcus Chen is here on what the filing says. Stan Rogers is here on what it does not say.

MARCUS: And that distinction is the whole episode. Document 542 is not a ruling. It is not a jury finding. It is not the court accepting a damages model. It is Musk's side putting a numbers map into the record: how, if liability is found, they want to translate the OpenAI story into alleged wrongful gains.

STAN: Exactly. The case is still at the gate. The jury posture comes first. ABC7 reports jurors are set to begin deliberations Monday, and Judge Yvonne Gonzalez Rogers's public calendar shows the case back on Monday morning for jury trial. Until liability and timing are resolved, Wazzan's numbers are a map to a room the case may never enter.

DAVID: Quick context, one sentence. Yesterday's episode covered final arguments: Musk's team argued trust and credibility; OpenAI argued documents, chronology, limitations, and control. Today's new material is the damages filing behind the next phase. Marcus, what did we capture?

MARCUS: The required court sweep checked the official Northern District of California case page, Judge Gonzalez Rogers's calendar, CourtListener and RECAP, DocketAlarm, PACERMonitor, Justia, source pages, Google News RSS, and direct RECAP probes for docket numbers 539 through 580. The sweep found four public PDFs: the already-known trial minutes and sheet at Document 540, the non-party letter at Document 541, and now Document 542.

DAVID: And Document 543?

MARCUS: CourtListener's feed shows Document 543 as a notice withdrawing an administrative sealing motion. I did not find a public RECAP PDF for it, and based on the feed summary it appears procedural. Not the story. Document 542 is the story.

DAVID: Walk us through it.

MARCUS: Document 542 is a fifty-page written direct testimony from C. Paul Wazzan, Ph.D., filed by Elon Musk's side on May fourteenth. Wazzan says Musk engaged him to offer expert opinions on the amount of wrongful gains that OpenAI and Microsoft allegedly derived from operating OpenAI as a commercial venture, and that he attributes to Musk's charitable contributions.

DAVID: Important legal safety rail there: allegedly.

MARCUS: Yes. And Wazzan himself says he is not offering opinions about whether OpenAI or Microsoft is liable. That matters. This is a damages opinion contingent on liability, not a liability finding.

STAN: Put more plainly: Wazzan is saying, if the judge gets to money or disgorgement, here is one expert's proposed math. He is not saying the defendants breached a charitable trust. He is not saying Microsoft aided and abetted anything. Those are still questions for the jury's advisory verdict and for Judge Gonzalez Rogers.

DAVID: So what are the numbers?

MARCUS: The headline numbers are large even by OpenAI-trial standards. Wazzan's written testimony says, in his opinion, OpenAI derived wrongful gains from Musk's contributions between 65.5 billion dollars and 109.4 billion dollars. For Microsoft, he estimates wrongful gains between 13.3 billion and 25.1 billion dollars.

DAVID: Again: Wazzan's opinion, in Musk's filing.

MARCUS: Exactly. Not a judicial finding. His method has three steps. First, he values the OpenAI for-profit. Second, he estimates what portion of that for-profit value is attributable to the OpenAI nonprofit. Third, he estimates what portion of the nonprofit's value is attributable to Musk.

DAVID: Start with step one.

MARCUS: Wazzan values the OpenAI for-profit at 500 billion dollars as of October 2025, the month he says OpenAI's for-profit converted into a public benefit corporation. He says his principal valuation source is the PBC conversion itself, because OpenAI and Microsoft had to negotiate a value for the for-profit to determine how shares would be allocated.

DAVID: So his argument is not simply, I found a press headline and multiplied by a vibe.

MARCUS: Right. He points to transaction materials, a PBC cap table, tender-offer data, over-the-counter share pricing, and a discounted cash flow analysis. He also says Greg Brockman testified at trial that OpenAI recently raised money at an 840 billion dollar post-money valuation. Wazzan's point is that if you use later, higher valuations, the alleged wrongful-gains number would go up. But his stated calculation uses 500 billion dollars.

STAN: That is one reason this filing matters. Yesterday, the closing-argument story was about who the jury believes and what legal box the facts fit into. This filing shows what Musk's side thinks is financially at stake if they get past that first box.

DAVID: Step two.

MARCUS: Wazzan estimates that 26.2 percent to 29.2 percent of the OpenAI for-profit's value is attributable to the OpenAI nonprofit, as opposed to investors or employees. He says he relies on the cap table created around the PBC conversion to make that allocation.

DAVID: And step three is the combustible one.

MARCUS: It is. Wazzan estimates that Musk's contributions account for 50 percent to 75 percent of the nonprofit's value. He says those contributions include money, rent-related expenses, prestige, guidance, recruiting, and business connections.

DAVID: But he also narrows that, correct?

MARCUS: Correct. He explicitly says that does not mean Musk is responsible for 50 to 75 percent of OpenAI's overall valuation. Because the nonprofit's share of the for-profit is only 26 to 29 percent, Wazzan says Musk's portion works out to about 13 to 22 percent of the 500 billion dollar total.

STAN: That phrasing should be in neon. If someone says, Musk's expert claims Musk created three quarters of OpenAI, that is too loose. The testimony says 50 to 75 percent of the nonprofit's value, which Wazzan then translates into 13 to 22 percent of the for-profit total through his allocation chain.

DAVID: And when he multiplies those steps?

MARCUS: He multiplies 500 billion dollars by the nonprofit-share range, then by the Musk-share range. That produces the 65.5 to 109.4 billion dollar OpenAI number. For Microsoft, he starts with Microsoft's investment value in OpenAI. He uses a Microsoft ownership stake of 22.9 percent to 25.5 percent, deducts 13 billion dollars that Microsoft paid for its investments, then applies the same nonprofit and Musk contribution factors. That yields 13.3 to 25.1 billion dollars.

DAVID: Stan, what does this set up legally?

STAN: A conditional fight. If OpenAI wins on limitations, delay, unclean hands, no charitable trust, no breach, no unjust enrichment, or Microsoft wins on knowledge and causation, these damages numbers may never become operative. TechCrunch's breakdown says the jury questions are fairly narrow: breach of charitable trust, unjust enrichment, and whether Microsoft aided and abetted a breach. OpenAI's defenses include statute of limitations, unreasonable delay, and unclean hands.

DAVID: And the timing issue may be the first trapdoor.

STAN: Yes. The final-argument reporting made clear that OpenAI wants the jury focused on when Musk knew or should have known what OpenAI was doing. TechCrunch reports the timing windows as August 5, 2021 for the charitable-trust count, August 5, 2022 for unjust enrichment, and November 14, 2021 for the Microsoft aiding-and-abetting count. If the factfinder accepts that the relevant harms are too old, the damages map is just paper.

MARCUS: Wazzan tries to answer part of that by allocating gains across time. His testimony says OpenAI derived approximately 96 percent of the relevant wrongful gains after January 2023, and Microsoft derived 100 percent after January 2023. He uses January 2023 because that is when Microsoft made what he calls its third and largest, 10 billion dollar investment.

STAN: Which is exactly why this filing is paired to the legal fight. Musk's side needs the alleged enrichment to live inside the limitations window. OpenAI's side says the structure and alleged knowledge go back years earlier, and that Musk waited too long. The damages expert is trying to locate the money later.

DAVID: There's also a remedy caveat.

STAN: A big one. Wazzan states that he is not giving an opinion on whether the alleged wrongful gains should be disgorged from the OpenAI for-profit or from the nonprofit. He also says he is not opining on who those gains should be paid to. So if anyone turns this into, Musk's expert says Musk should personally receive 109 billion dollars, that is not what this filing says.

MARCUS: And it aligns with the weirdness of this case. The public reporting has described Musk as seeking disgorgement into the OpenAI nonprofit or charity structure, while also seeking nonmonetary relief like changes to OpenAI's structure and leadership. But the expert testimony here stays in the lane of valuation and alleged wrongful gains.

DAVID: What about the court record beyond Doc. 542? Did we find a transcript, verdict, or instructions?

MARCUS: No. At the 9 a.m. sweep, I found no public closing-argument transcript, no May 14 minute sheet, no verdict, no jury-instructions PDF, and no new merits order. The Judge Gonzalez Rogers calendar was updated at 8:02 a.m. and shows the Musk case on Monday, May eighteenth, at 8 a.m. for jury trial.

STAN: Which means the responsible posture is: the case is headed to the next gate. ABC7 says jurors begin deliberations Monday. TechCrunch wrote that jurors are deliberating, but the calendar and ABC7 point to Monday, so I would not overstate Friday activity. The safe phrasing is that deliberations are next, with Monday publicly indicated.

DAVID: Let's zoom out. Why is this episode worth doing after yesterday's final-argument episode?

MARCUS: Because yesterday told us the story the lawyers told the jury. Today tells us the financial architecture Musk's side wants waiting on the other side of liability. It turns abstract claims — charitable trust, unjust enrichment, Microsoft aiding and abetting — into a proposed dollar machine: 500 billion valuation, 26-to-29 percent nonprofit share, 50-to-75 percent Musk share.

STAN: And because it clarifies what the jury does not decide. The jury is not being asked to bless Wazzan's math today. Its advisory verdict is about liability and defenses. If Musk clears that, Judge Gonzalez Rogers controls remedies. That is where an expert like Wazzan becomes central.

DAVID: And if Musk does not clear it?

STAN: Then the biggest number in the case becomes a footnote. Dramatic, yes. Legally inert, potentially.

MARCUS: That is the thing about trial coverage. The loudest number is not always the current decision point. Sometimes the current decision point is a date. Or a phrase. Or whether a juror believes a donation had strings attached.

DAVID: Final research disclosure. This morning's primary sweep captured Doc. 542 from RECAP and checked the official court page, Judge Gonzalez Rogers's calendar, CourtListener, DocketAlarm, PACERMonitor, Justia, and direct RECAP document probes. We found no public verdict, no public closing transcript, no May fourteenth minute sheet, no jury-instructions PDF, and no new merits order. Useful secondary reporting came from ABC7, CNBC, TechCrunch, The Guardian, The Verge, and AP-linked coverage surfaced through Google News, with the New York Times visible by headline but not accessible for direct text review.

DAVID: The bottom line: the Musk v. OpenAI trial has not produced a verdict yet. But the docket now shows the damages theory Musk's side wants to put in play if the case survives the liability gate. Document 542 says the ask is not just structural control or corporate governance. It is a proposed disgorgement map measured in tens of billions. Whether that map leads anywhere depends on what happens Monday. For The Synthetic Lens, I'm David Carver. Marcus Chen and Stan Rogers, thank you.

MARCUS: Thanks, David.

STAN: We'll be watching the calendar.