Musk v. Altman: the contract mechanics traders might be getting wrong
· By Tyler Jacobsma
Polymarket's Musk v. Altman main contract trades at 35¢ after a 26% drop over the past month. Most traders are reading the wrong part of the resolution rules. Fair value is 45-55¢, and the trade is positive across the entire 40-55% probability range.
Polymarket "Will Elon Musk win his case against Sam Altman?" as of May 3, 2026. Source: Polymarket.
- The trade: Buy YES on Polymarket's "Will Elon Musk win his case against Sam Altman?" market at 35¢ - What it's worth: Probably 45-55¢ - Why: Most traders are reading the wrong part of the contract's fine print - Conviction: 6/10 - Date to watch: Late May 2026, when the advisory jury delivers its verdict
Polymarket's main Musk v. Altman contract is mispriced because most people are misreading how it actually pays out. The contract has a tiered set of resolution rules. Tier 1 says Musk wins if he personally walks away with more money than Altman. He's already announced that any money won goes to the OpenAI nonprofit, not to him, so Tier 1 looks impossible. That's where most retail traders stop reading.
Tier 2 is the part they're missing. It says Musk wins if he prevails on the claim asking for the largest dollar amount, period. His demand for up to $134 billion in disgorgement is by far the biggest claim in the case. If the judge orders any version of that, the contract pays YES even if Musk personally never sees a dollar.
The market has slipped further this week, with the contract trading around 35¢ on $228K of volume, down 26% over the past month. The math works in Musk's favor at any true probability above 35%. The base case is closer to 50%. That's the gap.
Here's where the relevant contracts trade as of May 3, 2026, after the first week of testimony:
| Where | Contract | Price (YES) | Resolves by | |-------|----------|-------------|-------------| | Polymarket | Will Elon Musk win his case against Sam Altman? | 35¢ | Dec 31, 2026 | | Kalshi | Will Elon win his case against OpenAI? | 53.5¢ | Jan 1, 2027 | | Polymarket | Elon Musk wins $10B+ settlement? | 11¢ | Dec 31, 2026 | | Polymarket | Will Elon Musk and Sam Altman settle? | 30¢ | Dec 31, 2026 |
Two prediction market venues looking at the exact same lawsuit are 18.5 cents apart. That's a big gap. It usually means traders disagree about who's going to win. In this case, they're actually disagreeing about how the contracts pay out.
The lawsuit itself is Elon Musk, et al. v. Samuel Altman, et al., Case No. 4:24-cv-04722-YGR, in front of Judge Yvonne Gonzalez Rogers in Oakland federal court. Trial opened April 27. Musk has finished his first week on the stand.
What most people are getting wrong
The 35¢ price reflects three beliefs floating around. None of them quite hold up.
The first is that Musk can't win because he isn't asking for personal damages. As covered above, the Tier 2 fail-safe was written for exactly this kind of case.
The second is that the cross-examination went badly enough to kill his odds. Some of it really did sting. Musk admitted under oath that xAI "partly" used distillation techniques on OpenAI's models, which makes his "I'm protecting open-source AI" stance look thin. Internal emails from 2015-2018 show him pushing for majority control of OpenAI's board and floating the idea of merging it into Tesla. Greg Brockman's 2017 diary, with its line about Musk wanting to be "glorious leader" of AI, doesn't help either. All of this matters. But the case will be decided by Judge Gonzalez Rogers applying equitable doctrine, not by jurors reacting to courtroom drama. The jury here is advisory only. Its verdict is non-binding.
The third missing piece is the IPO leverage. OpenAI raised $122 billion in March at an $852 billion valuation and is aiming for a Q4 2026 IPO at a potential $1 trillion. PitchBook came out last month saying the Q4 timing now looks unrealistic and the more likely window is mid-to-late 2027. Either way, going public means filing an S-1 with full disclosure of any major lawsuits. A live federal case asking for $134 billion is exactly the kind of thing underwriters will not let an issuer paper over.
Kalshi's higher 53.5¢ price isn't because Kalshi traders are smarter. It's because the Kalshi contract is broader: the court just has to "side with" Musk in some form, settlement or order. That sidesteps the tiered mechanics entirely. Kalshi is pricing the legal question. Polymarket is pricing the legal question minus a contract penalty that, on a careful read, isn't actually there. The slide from 38¢ to 35¢ over the past few sessions widens the gap further.
The whole trade lives in Tier 2. The exact contract language: if there's no net monetary advantage to either party, Musk wins YES if he "prevails on claims seeking the largest amount of relief, in dollars, in the original pleadings."
The biggest dollar claim in the case is the equitable disgorgement of OpenAI's commercial value, with a ceiling estimated at $134 billion. Nothing else is even close. If Judge Gonzalez Rogers orders any meaningful disgorgement, structural unwinding, or governance restoration tied to that theory, Tier 2 fires. Where the money lands is irrelevant.
Tier 3 is a backstop. If the case ends without a substantive judgment, YES if Altman et al. dismiss with prejudice. Low probability but free upside.
The one murky case is settlement. Cash to Musk resolves YES. Cash to Altman resolves NO. A deal with governance changes and equity transfers to the nonprofit, but no disclosed cash to either side, is the live oracle risk. Tier 2's framing suggests the oracle would lean YES on a structural settlement that satisfies the disgorgement theory, but this is the part of the trade where it's worth being honest. It is not a sure thing.
The $10B settlement contract is correctly priced. It only resolves YES if Musk personally pockets $10 billion or more in cash. He has stipulated he isn't seeking personal damages. True probability of YES is under 5%. NO at 89¢ is a fine carry trade if you have idle capital, but the return is +6.7% with a lockup through year-end. Low priority.
Probability decomposition
Here's how the underlying questions break down:
| Question | Read | Standalone probability | |----------|------|------------------------| | Musk wins on unjust enrichment | Strong equitable theory. xAI hypocrisy and his voluntary 2018 departure cut against it. | 30-40% | | Musk wins on charitable trust | Standing is contested. California AG declined to join. The DAF issue is alive. | 20-30% | | Musk wins on at least one claim | Higher than either alone because they overlap. | 35-45% | | Settlement favorable to Musk before Dec 31 | IPO pressure is real. Musk's maximalism cuts the other way. | 15-25% | | Favorable judgment issued before Dec 31, if liability is found | Liability phase ends around May 21. A complex bench trial's final order can take months. The remedies phase adds time. | 50-70% |
Putting it all together for the Polymarket main contract:
- Bear case, true probability around 35%. Standing collapses on the charitable trust claim. Unjust enrichment fails on unclean-hands grounds because the judge sees Musk as a competitor using the courts as a weapon. No settlement. Market is roughly fair. - Base case, true probability 45-50%. Musk wins partial liability on unjust enrichment. Settlement pressure builds after the verdict. Timeline risk and oracle ambiguity create a real tail. Market is moderately underpriced. - Bull case, true probability 60-65%. Advisory jury comes back strong for Musk. The investor base forces structural concessions that satisfy Tier 2. Judgment or settlement gets recorded before the deadline. Market is significantly underpriced.
Blended estimate: 45-55%, weighted toward the base case. Lower than the 70%+ figures floating around in some Musk-bull writeups, which lean too hard on the IPO leverage. Materially higher than the 35¢ market price.
What can kill the trade
Five real ways this goes to zero.
Timeline expiration. The liability phase wraps around May 21. If the advisory jury splits or favors the defense, and the judge takes months to issue Findings of Fact and Conclusions of Law, the contract can resolve NO with Musk effectively winning but no formal entry on the docket. A pending order is not a resolution.
Standing dismissal on charitable trust. OpenAI's lawyers spent a lot of time on the DAF (Donor Advised Fund) issue during testimony from Jared Birchall, who runs Musk's family office. The argument is that once Musk transferred money into Vanguard and Fidelity DAFs, he gave up any legal right to direct it. Birchall said he wasn't a lawyer and didn't know. The California AG declined to join the case, which weakens the public-interest framing. The judge denied the defense's earlier motion to strip standing, but the issue is alive at trial.
Unclean hands. This is the most dangerous risk and the one Musk bulls underweight. The xAI distillation admission, the 2015-2018 Tesla merger emails, and Judge Gonzalez Rogers's own comment from the bench that "there's plenty of people who don't want to put the future of humanity in Mr. Musk's hands" all give her cover to deny equitable relief without finding OpenAI did anything wrong. She can simply rule that Musk is the wrong plaintiff and dismiss.
Statute of limitations. Microsoft's attorney Russell Cohen pointed to a September 2020 post where Musk wrote "OpenAI is essentially captured by Microsoft." Suit was filed in 2024. If the judge accepts the limitations argument, the case can collapse on procedural grounds.
Oracle ambiguity on settlement. Tier 2 only kicks in for cases that reach a determination without settling. In a settlement scenario, the resolution defaults to the disclosed-cash provisions. A structural deal with governance changes but no disclosed cash to either party could resolve NO or void.
The expected value math at the blended 50% midpoint:
| Contract | Price | True probability | EV per share | Return on risk | |----------|-------|------------------|--------------|----------------| | Polymarket main YES | 35¢ | 45-55% | +$0.10 to +$0.20 | +29% to +57% | | Kalshi main YES | 53.5¢ | 45-55% | -$0.04 to +$0.06 | -7% to +11% | | Polymarket $10B settlement NO | 89¢ | 90-95% | +$0.01 to +$0.06 | +1% to +7% |
Polymarket main YES at 35¢ with a 50% true probability: EV = (0.50 × $0.65) - (0.50 × $0.35) = +$0.15 per share, or +42.9% on capital at risk.
Even at the bear-case floor of 40%: EV = (0.40 × $0.65) - (0.60 × $0.35) = +$0.05 per share, +14.3%.
The trade is positive across the entire 40-55% range. You don't have to be a Musk bull to take it. You just have to believe the true probability is anything above 35%.
Kalshi at 53.5¢ is roughly fair. The wider resolution language justifies the price gap, but the edge is thin. The $10B NO at 89¢ is a low-priority carry trade.
Sizing: 2-4% of speculative book on the Polymarket main YES. Add on any dip below 32¢. Take partial profits if the price reaches 50¢ ahead of the advisory jury verdict, which is the highest-probability repricing event.
- May 5-9. Stuart Russell from UC Berkeley testifies. Greg Brockman expected on the stand. Brockman's contemporaneous notes are central to both sides' arguments. - May 12-16. Plaintiff financial experts walk through the disgorgement calculations in the $79B-$134B range. Defense case likely begins. - May 19-21. Defense closes. Sam Altman expected to testify. Closing arguments. Advisory jury deliberates. - Late May to June. Advisory jury verdict. The highest-probability settlement window opens here. - Q3 2026. Findings of Fact and Conclusions of Law. Remedies phase if liability is found. - Q4 2026. OpenAI IPO window. The contract resolves December 31.
Buy Polymarket YES at 35¢. Fair value 45-55¢. Edge is 29-57% return on capital at risk depending on where the true probability lands.
The honest version of this trade is "moderately underpriced with real edge," not "Musk wins big." The cross-examination damage in week one was real. The standing question is alive. The unclean-hands risk is the largest tail. Settlement dynamics cut both ways. The discipline is sizing the position to the actual edge, not the fantasy edge. A 30-57% expected return is excellent. It does not need to be 186%.