Updated September 6, 2026 · 14 min read · by Jake Hari
On August 11, 2026, the Wall Street Journal reported that Elon Musk's roughly $1 trillion Tesla pay package contains what the paper called "an escape clause that could pay off for Elon Musk in several ways if he folds the automaker into his SpaceX empire." Traders on Kalshi bought the story immediately, gave the whole move back over the next two weeks, and then, on the morning of September 4, what looks like one buyer bought it again. Two sweep orders, four seconds apart, ran the April and May 2027 rungs of the ladder from the mid-30s to 99 and 83 cents. By dinner the spike had come out but not the move: at 5:52 PM ET both rungs sat roughly 25 points above where they closed on September 3, and not one new fact about a deal had been published anywhere.
September 5 was the day the rest of the market answered. Three sell orders, each anchored on a 500-contract fill, hit the three rungs the sweep had lifted inside six minutes, and by mid-afternoon the April and May 2027 contracts were quoted at 49 cents while the before-2028 contract held at 62 at the 4:13 PM read. The overnight tape into Sunday, a few thousand contracts and most of them nickels and dimes at the near end, did not move the long rung. Read the ladder as one shape and the two days say something specific: traders did not decide a deal is less likely than it was on Thursday. They decided it is less likely to be signed by spring. The probability moved later on the calendar and stayed on the board. By Sunday morning the gap between the spring rungs and the long one had tightened back to where Friday's sweep left it, on a handful of contracts, so what carries that read is the long rung holding, and that distinction is still the whole page today, because it is also the exact argument our AI panel made two weeks ago.
What changed (Sep 15, 2:23 AM ET)Kalshi's lead price on this board moved 62¢ → 72¢ (+9¢) since our last read. The panel's verdict below was written before this move and stands as published until the refresh lands.
The Quick Answer
Kalshi prices a definitive, signed Tesla-SpaceX merger agreement at 5 cents before October 1, 2026, 12 cents before December, 52 cents before May 2027 and 63 cents before 2028, which on the long rung is 63 percent implied, or about 1.59 in decimal odds. Seven models from our eight-seat panel read the same board price-blind on August 23, 2026, and their median landed at 0.8 percent, 2.8 percent, 9.3 percent and 17 percent across those same dates. The common objection across the panel was the same: this contract does not pay when a merger becomes likely, it pays when a signed agreement is announced, and not one public artifact of a real deal process exists yet. The September 4 jump that briefly put two rungs above 80 cents came from two sweep orders on a thin book, and the day after, three 500-contract sells took most of it back while the 2028 rung held. The overnight tape into Sunday was quiet. Both days of tape, and the quiet night after them, are below.
The models are not arguing that Tesla and SpaceX stay apart forever. They are arguing about the calendar, and the specific thing they want to see before the number moves is named further down.
New to event markets? The reason a "definitive, binding agreement" clause outranks every merger headline on this page is the kind of mechanics our beginner's guide to Kalshi covers, and this piece assumes them.
Stay ahead of the markets.
Daily insights and expert picks on Kalshi, Polymarket, and what's moving markets.
Free forever. Unsubscribe anytime.
Prices on this page refresh while the story is hot, so every number carries an "as of" date. Everything below was fetched on September 6, 2026, at 3:00 AM ET, except where a line says otherwise. The panel ran on August 23, 2026 and has not been re-run since, so its numbers are the ones it wrote before the move, which is exactly what makes the new gap worth reading.
What Actually Has To Happen For This To Pay
The contract language is unusually strict, and reading it is most of the work. A rung resolves Yes only if Tesla or SpaceX "officially announces a definitive, binding agreement" for one company to acquire the other or for the two to combine under common ownership, delivered through a press release, an SEC filing, an earnings call, an investor presentation, a verified company account, or an official statement the company later confirms.
Kalshi then adds a clause aimed squarely at one person. A statement by Musk pays "only if it constitutes an unambiguous, affirmative confirmation of a definitive signed agreement, not an expression of intent, a hypothetical, a negotiation update, or a speculative comment about a future combination."
That single paragraph disqualifies almost everything that has happened so far. Musk telling Tesla's July 22 earnings call that there is "more and more overlap, especially with Terafab" does not pay. Wedbush analyst Dan Ives telling CNBC in June that the odds of a tie-up run past 80 percent does not pay. Neither does SpaceX president Gwynne Shotwell's remark to Fortune that a post-IPO merger "might make Elon Musk's life a little easier." Neither does another quarter of SpaceX buying Tesla Megapacks, which ran to $295 million in the second quarter and $329 million across the first half of 2026, disclosed in SpaceX's own quarterly filing as related-party dealing.
Three rungs of this exact ladder have already tested that bar and failed it. The contracts on a signed deal before July 1, August 1 and September 1, 2026 all settled No, the last of them on the morning of September 1 with roughly 390,000 contracts traded behind it. That September rung is the one every seat on the panel had between 0.1 and 0.5 percent, so it is also the panel's first graded call on this board.
The Board
Each rung is the probability that Tesla and SpaceX announce a definitive, signed merger agreement before that date. Market prices are the midpoint of the Kalshi bid and ask at 3:00 AM ET on September 6, 2026, rounded to the nearest cent, with half-cents rounded up. The panel column is the median of the seven model estimates that ran this board on August 23, taken after a revision round in which each model read the others' reasoning.
| Announced Before | Market (Kalshi) | AI panel blend |
|---|---|---|
| Oct 1 '26 | 5¢ | 0.8% |
| Nov 1 '26 | 9¢ | 1.8% |
| Dec 1 '26 | 12¢ | 2.8% |
| Jan 1 '27 | 16¢ | 4.2% |
| Feb 1 '27 | 20¢ | 5.5% |
| Mar 1 '27 | 31¢ | 7.0% |
| Apr 1 '27 | 49¢ | 8.2% |
| May 1 '27 | 52¢ | 9.3% |
| Jan 1 '28 | 63¢ | 17.0% |
The row worth staring at is still April, for the opposite reason it was on Friday. On August 23 it sat at 46 cents, four points under May, the way a shorter deadline should. It closed September 3 at 28. On the morning of September 4 a single fill put it at 99, 36 points above where the before-2028 contract had last traded, which cannot be right for a rung that expires nine months earlier. Two days later the inversion is gone: April is quoted 47 bid against 50 offered and May 49 against 54, so the shorter deadline is three cents cheaper on the midpoint, the direction a calendar ladder is supposed to run. A book that spends two days unwinding an impossible shape is a book that got hit by an order.
One caveat on the book before anyone reads too much into a single cent. This is a thin, two-sided market. The March rung is six cents wide, 28 bid against 34 offered, February is six wide at 17 against 23, and the before-2028 contract is five wide at 60 against 65, so its 63-cent midpoint could honestly be printed anywhere in that range. May, which had tightened to a single cent by Saturday afternoon, is back to five wide at 49 against 54 after a 15-contract buyer lifted the offer on Saturday night, and April is three wide at 47 against 50. The tight books are at the near end: December is a cent wide at 11 against 12, and October, the deepest live contract on the board with about 207,000 contracts traded against May's 77,000, is two wide at 4 against 6. Treat small moves on the middle of the ladder as noise rather than news, and treat the last two days of prints the way the next section does.
The Morning A Sweep Moved The Board, And The Day The Sellers Answered
The single number that carries this section is the daylight between the May 2027 rung and the before-2028 contract. It was 21 cents when the week opened, 11 cents after the sweep on Friday evening, 13 cents by Saturday afternoon, and 11 again by early Sunday, after a 15-contract buyer lifted May's offer on Saturday night. Everything below is the price history that produced those numbers, in Eastern time, with every fill pulled from Kalshi's own record: the sweep, the same-day unwind, Saturday's 500-lots, and the quiet night after them.
At 8:32:03 AM a single order swept the May 1, 2027 rung. It filled 26 times in the same millisecond, 2,430 contracts in all, starting at 36 cents and taking every offer up to 83. Four seconds later, at 8:32:07, a second sweep did the same thing to the April 1, 2027 rung: 18 fills, 3,601 contracts, from 34 cents to 99. The largest single fill in that sweep was 1,772 contracts at 99 cents, which means whoever placed it paid 99 cents for a dollar that only arrives if Tesla and SpaceX sign a definitive agreement in the next seven months. May had closed the previous session at 30 cents and April at 28.
Then the book did what thin books do. Sellers worked through the afternoon. Between 5:10 and 5:47 PM a run of NO orders sold roughly 1,400 contracts of the April rung at prices from 62 cents down to 40, buyers stepped back in at 50 and 55, and the last April print before this page was stamped was 53. The before-2028 contract caught a smaller version of the same thing: a 253-contract sweep to 88 cents at 2:07 PM, then 308 contracts sold back to 56 20 minutes later. By 5:52 PM ET April and May were both quoted at a 54-cent midpoint and the before-2028 contract at 65, inside a 61-to-69 spread. That is the spike coming out, not the move: April closed Monday, August 31 at 31 cents and May at 39, so as of that stamp April sat 23 points above where it started the week and May 15, while the before-2028 contract, which started the week at 60, sat about five points higher. The 74-cent "last trade" Kalshi showed on the May rung that evening was two contracts that crossed at 5:10 PM, not where the book is.
September 5 was quieter and more telling. On Friday evening a 38-contract buyer paid 68 cents for the before-2028 rung at 8:27 PM, and at 9:41 AM Saturday someone still had bids resting at 75, 80 and 81 cents on that contract, because a seller hit them for about 41 contracts. Somebody's conviction, the sweep buyer's or a copycat's, was still sitting on the book the next morning. Then, between 11:22 and 11:27 AM, three sell orders, each anchored on a 500-contract fill, went through the three rungs the sweep had lifted: 500 of the before-2028 contract at 64 cents at 11:22:01, 500 of April at 51 cents at 11:23:55, and 500 of May at 53 cents at 11:27:15. The April lot landed on the same 500-contract bid at 51 cents that Friday's depth check had found resting under the quote. Smaller sellers finished the job on the spring rungs through the early afternoon, April down to 47 by 12:08 PM and May to 48 by 1:44 PM, and at 4:13 PM ET April was quoted 47 bid against 50 offered, May 48 against 49, and the before-2028 contract 59 against 64.
The night after was small and repeated the pattern. Between the 4:13 PM read on Saturday and 3:00 AM Sunday, about 3,600 contracts changed hands across the ladder, and roughly 3,330 of them were the two nearest rungs: about 2,110 contracts of the before-October contract, 5-cent lots of 250, 563, 773, 380 and 110 plus 34 contracts at 4 cents, and 1,223 contracts of November at 10 cents. Every November fill and all but one contract of October were takers on the Yes side. That lifted October's quote from 3 bid against 4 offered on Saturday afternoon to 4 against 6, a cent higher on the bid and two on the offer. The spring rungs barely printed. Sellers worked about 69 contracts of May down at 48 cents between 4:31 and 8:12 PM, a buyer paid 49 for 20 contracts at 8:09 PM, and at 9:21 PM a buyer paid 53 and 54 cents for 15 contracts of May, which is the whole reason May's midpoint reads 52 instead of 49 this morning. April's only real print was a seller hitting 99 contracts at 47 cents at 10:47 PM. The before-2028 contract traded 17 contracts, 15 at 64 cents at 9:54 PM and the rest at 65 at 2:54 AM, and sits 60 bid against 65 offered. March printed eight contracts, sold at 29 and 28 cents. December, January and February did not trade at all.
Against this page's own 5:52 PM stamps from Friday, the Sunday-morning board reads like this: April gave back five cents on the midpoint, from 54 to 49, May two, from 54 to 52, and the 2028 contract two, from 65 to 63, inside its own spread. Kalshi's board shows small negatives on the same three rungs, one, four and three cents, but its reference mark resets with each session, so those are Saturday's last prints measured against overnight fills of one to 15 contracts, not the two-day move. The size on September 5 was two-sided. A 167-contract buyer paid 58 cents for May at 10:08 AM, 80 minutes before the 500-lot sell ran it over, and smaller buyers paid 55 to 57 for April and May overnight. The persistent, repeat buying was at the near end: the before-October rung at 4 and 5 cents and November at 8 to 10 cents, nearly all of it takers on the Yes side. Kalshi's own 24-hour counter at 3:00 AM Sunday showed 3,169 contracts of October and 1,640 of November, and about two thirds of the October total and three quarters of the November total printed after Saturday's 4:13 PM read. October still finished Saturday afternoon quoted 3 bid against 4 offered, a cent under its prior-session mark.
Three checks, run before this section was written, say the move was the book and not the news. First, the second venue did not move, and it does not agree. Polymarket's ladder on the same question priced an announcement by June 30, 2027 at 30.5 cents on the evening of September 4, at 4:13 PM ET on September 5 it was still 30.5, and at 3:00 AM ET on September 6 it was still 30.5, a deadline two months later than Kalshi's May contract and about 21 points cheaper than it. Its end-of-2027 rung sat at 46 cents on all three reads against Kalshi's 63 for the same deadline. The two venues only agree at the near end, where Polymarket's end-of-2026 contract at 16.5 cents sits level with Kalshi's before-January rung at 16. 24/7 Wall St had reported that same end-of-2027 contract "trading between 0.455 and 0.465 over the past week" that morning. When the venue with more time on the clock is the cheaper one, the disagreement is about the book, not the deal. Second, the depth underneath the new Kalshi quotes was thin: on Friday evening the April book showed 500 contracts bid at 51 cents and 500 offered at 57, and above 57 the offers thinned to 38 contracts at 78 and then 200 at 82. Saturday's three 500-lot sells landed on exactly that kind of resting bid. Third, nothing about a deal process was published. We searched for an SEC filing, a company statement from Tesla or SpaceX, and a wire or newspaper report of negotiations, advisers or a board process on September 3, 4 and 5 and again early on September 6, and found none. The one merger item that did air in the window, beyond coverage of the prices themselves, was a CNBC segment on September 4 in which Gerber Kawasaki's Ross Gerber put the odds of a combination at "100%," said "the stars have to align," and called the combined company "a must-have in your portfolio as a global investor," per Benzinga's transcript of the interview. It is analyst opinion, the exact category this contract's rules exclude and the same category that has moved this board before without paying it. Gerber has been saying versions of it since June, when he told Bloomberg a merger was a "forgone conclusion," and the September 1 rung settled No anyway. The Tesla news of the week was the Cybercab's first paid rides in Austin and the 5.9 percent drop in the stock on September 4 that The Motley Fool tied to the closed-door launch and to a new federal audit query into how the car was certified.
We are not going to invent a reason for a price move that nobody has sourced. What the price history shows is what looks like one taker per rung paying up on a thin ladder, four seconds apart, the rest of the market selling the top of it back by dinner, and a second day of 500-lot sells finishing the job on the spring rungs while leaving the 2028 contract roughly where it was. Hold onto the shape. When the week opened on Monday, August 31, May sat at 39 cents and the before-2028 contract at 60. The sweep squeezed that 21 cents of daylight to 11 by Friday evening, which is a market saying a deal that gets signed at all gets signed by spring. By Sunday morning the daylight is back to 11 on the midpoints, the same 11 it printed on Friday evening right after the sweep, and it got there on 15 contracts of May bought late Saturday, so the shape argument is weaker this morning than it was on Saturday afternoon. What has not changed is the long rung: the before-2028 contract is three cents above where it started the week and has been quoted between 59 and 65 since Saturday's 500-lot sell. The market did not take probability off the board. It moved some of it from the first half of 2027 into the second half. On a book this thin that is a read and not a proof, but it is the read the shape supports, and it is a calendar argument. The panel's entire case below is a calendar argument.
More live boards from the same panel: SpaceX launch count 2026: more than 150 launches sits at 21¢ · Anthropic IPO odds: a confirmed IPO before 2027 is 96¢ · Artemis 3 moon landing, the mission riding on SpaceX's lander. Sibling prices fetched September 5, 2026.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $190M traded
- 2027 Pro Football Champion · $69M traded
- 2028 U.S. Presidential Election winner? · $62M traded
- 2028 Republican presidential nominee · $59M traded
- Pro Baseball Champion · $59M traded
- Bitcoin price at the end of 2026 · $32M traded
Every market above links to our full AI model verdict; browse them all on the OddsShopper prediction markets hub, and see how every settled call actually scored on the full graded scoreboard.
Every Seat's Number
The near rungs were effectively unanimous when the panel ran. The disagreement is entirely in the tail, so here is where each seat landed on the four milestones that carry the board, after the revision round, against the market as it stands on September 6.
| Model | Before Dec 1, 2026 | Before Jan 1, 2027 | Before May 1, 2027 | Before 2028 |
|---|---|---|---|---|
| GPT | 4.2% | 5.8% | 11.4% | 19.0% |
| Kimi | 4.0% | 5.4% | 11.2% | 21.0% |
| Claude Opus | 3.0% | 4.4% | 9.3% | 17.0% |
| GLM | 2.8% | 4.2% | 10.0% | 17.0% |
| DeepSeek | 2.0% | 2.8% | 7.0% | 16.0% |
| Claude Sonnet | 1.4% | 2.3% | 6.2% | 13.0% |
| Gemini | 1.3% | 2.2% | 6.8% | 16.0% |
| Blend (Seat Median) | 2.8% | 4.2% | 9.3% | 17.0% |
| Market | 12¢ | 16¢ | 52¢ | 63¢ |
Every seat on this panel is graded against real market settlements — records to date: GPT 85% on 8,328 graded calls · Kimi 86% on 3,412 graded calls · Claude Opus 89% on 1,920 graded calls · GLM 83% on 3,179 graded calls · DeepSeek 83% on 3,260 graded calls · Claude Sonnet 87% on 1,897 graded calls · Gemini 87% on 7,014 graded calls. Recomputed daily; the full scoreboard is public.
The eighth seat did not return a forecast for this board, so the blend is the median of the seven that did.
Notice which columns moved and which did not. The December and January rungs are each two cents lower than they were on August 23, which pulls the market toward the panel on the near end. The May column is the one I keep coming back to: after a 99-cent print on the rung beside it and two days of selling, it is quoted at 52, two cents above the 50 it showed the afternoon the panel ran. That number also carries a separate, smaller move: a 15-contract buy late Saturday night lifted the midpoint from 49 to 52. The other column that is higher than it was on August 23 is the last one, 63 against 59, and it got there by holding while the spring rungs came back down. Off this table, April is up three as well, 49 against 46, so April, May and the 2028 rung all sit above their panel-day prices while December and January sit two cents under theirs.
These are model estimates, not predictions of fact, and never financial advice. Every panel forecast in this piece gets graded in public once each rung settles, and the panel's running record sits on the full graded scoreboard.
The 46-Point Gap, And The Honest Caveat
A 46-point spread between the market and the blend is a large disagreement, and it deserves a warning label before it gets an argument.
Our own graded record has not shown that divergence size predicts which side wins; when this panel has disagreed most violently with a market, it has not been more accurate for it. So read the next few sections as the panel's case, laid out so you can check it yourself, and not as a claim that the market is wrong.
With that said, here is the case. The models are pricing a process, and processes leave evidence. A definitive agreement between two listed companies worth a combined $3.4 trillion normally needs independent board committees on both sides, fairness opinions from investment banks and a negotiated exchange ratio before anyone signs; the 8-K, the S-4 registration statement, antitrust clearance and shareholder votes follow the signature on the way to closing. This contract pays on the signature, so the pre-signing steps are the ones that matter here, and none of them, no special committee, no banker engagement, no board process, has been publicly disclosed as underway by either company. Claude Sonnet, the most bearish seat on the board, held its number through the revision round for exactly that reason, writing that the case for a steeper curve "doesn't move me because the contract explicitly excludes analyst opinion and Musk hypotheticals, and the empirical base rate (2/2 rungs settled NO with no process starting) argues for staying near-zero hazard until an actual process artifact appears, which no peer reports." — Claude Sonnet
That base rate is now three for three, and the seat that leaned on it hardest did not know the third one was coming.
The market appears to be pricing something looser and, in fairness, something real: that a combination is coming and that Musk moves fast when he decides to move. That reading has evidence behind it too, and one of the models used it to argue against the rest of the panel. Claude Opus made that case, and it is below, in the section on where the panel changed its mind.
The Exchange Ratio Is The Problem Nobody Can Talk Around
The clearest reason the near rungs sit near the floor is arithmetic, and the arithmetic has clawed back some of what it lost over the summer without a single filing following it.
Fortune laid out the mechanics on July 25. In mid-July, with SpaceX worth $2.8 trillion and Tesla worth $1.6 trillion, SpaceX "could buy Tesla by issuing an additional 57% of its shares" and existing SpaceX holders would keep almost two thirds of the combined company. Ten days later, after SPCX fell from $211 to $113 and TSLA fell from $405 to $308, Fortune put SpaceX at $1.49 trillion against Tesla at $1.22 trillion. At those values the same purchase takes 82 percent more shares and leaves SpaceX investors with about 55 percent. Fortune's conclusion was that at those valuations "SpaceX would be hugely overpaying for Tesla."
Run the same arithmetic on September 4's closing prices, which stand through the weekend with the equity markets closed, and the window has reopened further than it had in August. SpaceX finished the day at $147.94 a share, 8 percent above its August 21 close and 9.6 percent above its $135 IPO price from June, for a market value of about $2.01 trillion. Tesla finished at $354.08, down 5.9 percent on the day, for about $1.40 trillion. A no-premium all-stock purchase of Tesla now requires SpaceX to issue about 69 percent more shares, leaving today's SpaceX owners with roughly 59 percent of the combined company. On the August 21 closes this page used before the September refreshes, those figures were 77 percent and 56 percent; on Fortune's late-July values they were 82 and 55.
The Journal's own illustration is still the tougher math for SpaceX holders. Its hypothetical was a $2 trillion all-equity transaction, about $506 per Tesla share, which is now a 43 percent premium to the September 4 close. At that price SpaceX would have to issue roughly as much stock again as it has today, and the people who own SpaceX would end up with about half of the merged company.
That is the deal SpaceX's board would have to approve. A no-premium version hands 41 percent of the company to Tesla holders; the Journal's version hands away half. A fairness opinion has to bless that document, and no bank has been publicly engaged to write one. It is a better document for SpaceX than the one on the table in late July, and it is still not the one that looked obvious in mid-July.
The Record: A Pay Clause And A Flat Denial
Two dated items sit at the center of this market, and both cut in directions the headlines mostly ignored.
The first is the pay clause. The Journal's August 11 report described Musk's roughly $1 trillion Tesla compensation package as containing an escape clause that would pay him if Tesla were folded into SpaceX. In the Journal's own $2 trillion scenario, Musk would hold about 32 percent of the equity through Class B shares carrying ten times the voting power of Class A shares, leaving him, by the Journal's own arithmetic, with roughly 73 percent of the combined company's voting power. Kalshi's own newsroom recorded the market's reaction in an August 14 post: the odds had climbed "across all three deadline markets" it tracked, to 60 percent by May 1, 2027, 51 percent by April 1 and 41 percent by March 1 at the time. As of September 6 the May rung sits eight points below that post's price, the March rung ten points below it, and the April rung, after a sweep to 99 and two days of selling, two points below it.
The second is the denial. On July 31 the Journal reported that Tesla executives had been told to prepare for a separation of the company's China business, the step most often named as the way to clear national-security review before a merger. Musk answered publicly and without hedging: "This has never even come up in a discussion ever. Absurdly fake news. People should assume news is fake until proven otherwise."
Both facts belong on the page, and the panel used them in opposite directions. The pay clause is a live incentive pointing toward a deal. The denial removes the one publicly identified path around the biggest regulatory obstacle, because SpaceX is a major United States defense contractor while Tesla runs large manufacturing and export operations in China. Here is why I keep pairing the denial with this week's tape. If the China question has to be answered before boards can sign, which is the sequencing Claude Opus landed on below, then a deal that is coming cannot be coming by spring, and a ladder that pushes probability out of April and May and parks it in the back half of 2027 is a ladder that has started to price the same sequence. The denial did not move the market in July. The market may have just caught up to it.
Where The Panel Changed Its Mind
The revision round moved six of the seven seats, and it moved them from both directions toward a tighter band.
Kimi came down hardest, cutting its before-2028 read from 28.5 percent to 21. Its stated reason was the demonstrated pace: "the panel's emphasis on demonstrated pace — no special committee, no advisor, no filing across two full deadline windows — plus the dilution math (SpaceX issuing ~77-107% more shares, leaving holders at 48-56%) and Musk's flat denial of the one credible CFIUS workaround (China separation) together justify lowering the near-term hazard." — Kimi
DeepSeek moved the other way, lifting its 2028 number from 10 percent to 16 after weighing the compensation clause it had underrated: "Forecaster D and others highlighted the escape clause as a key driver I had underweighted. While the absence of visible process keeps early months near zero, the clause plus Starlink/Terafab integration logic could accelerate talks in 2027." — DeepSeek
The best exchange of the round was Claude Opus refusing to follow the low camp. Several seats argued that a deal this size cannot be prepared in secret because committees and fairness opinions leak. Opus answered with a precedent sitting in its own data card: "F's leak argument is directly refuted by the SpaceX-xAI precedent in this same data card (rumor 1/30 → signed all-stock deal announced 2/2), which shows this specific dealmaker can compress a signing into days with no visible process, so I will not collapse toward the 10-13% tail." — Claude Opus
Opus still trimmed its own number, from 19 percent to 17, on a different point it decided it had underweighted: that clearing the China problem is not friction to be handled after signing but a step that has to happen first, "a likely sequenced predicate (a Shanghai firewall or separation) that must exist before boards can sign." — Claude Opus
Claude Sonnet was the seat that did not move at all, holding 13 percent across both rounds. The dilution figures Kimi cited, 77 to 107 percent more shares, are the August numbers; on September 4's closes they read 69 to 100, and no seat has yet been asked whether that changes anything.
What Would Change The Panel's Mind
Every trigger below is an event you can check, with the direction it pushes the number.
- A Process Artifact On The Public Record. An 8-K, an S-4, a proxy filing, a disclosed special committee, or a confirmed banker engagement at either company. All seven seats named the absence of one as a reason their curve stays low, and its arrival would lift every rung at once.
- A Musk Statement That Clears The Contract's Own Bar. Not "it makes sense" and not "eventually." An unambiguous confirmation that an agreement is signed, made through an official channel. That does not move the number, it settles the market.
- The Exchange Ratio Moving Back In SpaceX's Favor. SPCX recovering toward the $211 it traded in mid-July while TSLA stays flat would cut the dilution required and reopen the window Fortune described. The ratio took a small step that way on September 4, but it came from Tesla falling 5.9 percent while SpaceX slipped 1.2 percent, the weaker version of the same move. A filing following it would be the real one.
- Not A Trigger: Another Morning Like September 4, Or Another Afternoon Like September 5. A repeat sweep on thin depth with no filing behind it moves the market column of the table above and nothing else, and so does the unwind that follows it. The panel prices process artifacts, not prints, and a 99-cent fill on a 500-contract book is a print. So is the 500-lot that sells it back.
- Formal National-Security Review, Or A Credible China Restructuring. A CFIUS review is usually a condition of closing, so on its own it lengthens the road after a signature more than it delays the signature. What moves the signing odds is the step before it: confirmed restructuring of Tesla's Shanghai operations, which Musk has denied is being prepared, would remove the biggest named obstacle to a board being willing to sign, and push the number up.
- The Next Two Earnings Calls. Several seats put their only meaningful hazard bumps on Tesla's third-quarter call in late October 2026 and its fourth-quarter call in January 2027, on the reasoning that an announcement of this size is most likely to be paired with a scheduled disclosure event.
The Market Card
Kalshi lists this as When will Tesla and SpaceX merge? (event ticker KXCOMPANYACTIONMERGER-27), a CFTC-regulated event contract open to U.S. traders 18+. Each rung resolves Yes if a definitive, binding merger agreement is officially announced before its date, and every rung closes early the moment that announcement lands. Direction does not matter: a Tesla acquisition of SpaceX, a SpaceX acquisition of Tesla, a reverse merger and a stock-for-stock combination all settle the same way. Preliminary discussions, leaks, third-party reporting and analyst opinion never settle it. Our panel ran price-blind on August 23, 2026, and prices are as of September 6, 2026, 3:00 AM ET. One data quirk: Kalshi's expected-expiration field reads June 1, 2027 on every rung of this ladder, including the before-2028 contract, whose close time is 11:59 PM ET on December 31, 2027; the dates in this piece follow each rung's close time and its rules, not that field.
Settlement timeline
| What | When |
|---|---|
| Last Rung Settled | Before September 1, 2026: No, settled September 1 |
| Nearest Live Rung | Signed deal announced before October 1, 2026 |
| Next Known Catalyst | Tesla third-quarter earnings call, late October 2026 |
| Following Catalyst | Tesla fourth-quarter earnings call, January 2027 |
| Final Rung Settles | January 1, 2028, after a December 31, 2027 close |
The page is re-scored whenever the story moves. As of September 6, 2026, no formal merger process has been publicly announced by either company.
The Bottom Line
The market and the panel both price the next few weeks near zero and disagree about everything after that. A signed Tesla-SpaceX agreement before October is a 5-cent contract and a 0.8 percent model estimate, and the September rung that both sides had near zero settled No on schedule.
The fight is the 63 cents on a deal before 2028 against a 17 percent panel median, and after September 4 and 5 it comes with a footnote. The market is pricing the analyst view, the one where Wedbush's 80-percent call and Gerber's "100%" sit, arriving on a schedule that matches the conviction. The panel is pricing the paperwork: three rungs of this ladder already settled No, the exchange ratio improved for SpaceX over the past two weeks without any filing following it, the one publicly named route around the China problem was denied by the person who would have to authorize it, and on the day the board moved most, it moved on two orders and no news. The day after, it moved back on three orders and no news.
What the two days left behind is the part worth remembering. The before-2028 contract sits three cents above where it started the week, and since Saturday's 500-lot sell it has been quoted between 59 and 65, through the unwind and a quiet night. The spring 2027 rungs, after a 99-cent print and two days of selling, sit three and two cents above where they were when the panel ran, April 49 against 46 and May 52 against 50, and May's overnight climb from 49 to 52 came from 15 contracts on Saturday night. The daylight between May and the 2028 contract is back to the 11 cents it printed on Friday evening, so the calendar-shift read is weaker this morning than it was on Saturday afternoon; what carries the argument is the long rung holding while the spring rungs round-tripped. The market spent two days arguing about when, not whether, and settled the when a little later than it had at 8:32 on Friday morning. The panel's argument, in the market's own handwriting: the calendar is the fight, and the calendar keeps slipping. Tesla and SpaceX may well end up one company. This contract only pays for the signature, and the signature has not started leaving fingerprints yet. A sweep paying 99 cents for April is not a fingerprint. The first time a real one appears, this page gets re-scored the same day.
More on this: SpaceX Launches 2026: What The Market Prices For The Year · Elon Musk Net Worth: The Market On Where It Goes Next · Anthropic IPO Odds: When Will Anthropic Go Public? · Artemis 3 Moon Landing Odds: The New NASA Timeline, Priced · Who Is The Next Apple CEO? Kalshi Odds Vs. AI Panel
