Artemis 3 Moon Landing Odds: The New NASA Timeline, Priced
If you searched for Artemis 3 moon landing odds, or the Artemis 3 launch date, the first thing to know is that NASA quietly took the landing out of the mission. Artemis 2 flew its crew around the Moon in April 2026 and came home clean; a month earlier, NASA had already restructured what comes next, turning Artemis 3 into an orbital docking rehearsal and pushing the first boots-on-the-surface flight to Artemis 4, targeted for early 2028. Kalshi, a CFTC-regulated exchange, prices the whole question as a ladder: crewed NASA landing by end-2026, end-2027, end-2028, end-2029. We handed that ladder to our AI panel, price-blind, with a card carrying the full restructure record and this program's unbroken history of slipped dates. The panel came back more optimistic than the traders on the far rungs, 25.2% by end-2028 and 55.9% by end-2029 against 20.5 and 47.5 cents, and the reason is a date-math detail half the seats admitted they almost missed.
The Quick Answer
The first thing the odds say: Artemis 3 is no longer the landing. NASA restructured it in March 2026 into a late-2027 orbital docking test, and the first crewed landing now rides Artemis 4, targeted early 2028. Kalshi prices a crewed NASA landing by end-2028 at 20.5 cents and by end-2029 at 47.5 cents; our eight price-blind models blended higher on both, 25.2% and 55.9%, while agreeing 2026 and 2027 are dead rungs (0.2% and 1.1%). The panel's whole disagreement with the market is one observation: NASA's target says early 2028, but the contract runs to December 31, so the 2028 rung quietly tolerates nearly a year of slip. The full ladder, the serial dependency chain, and the single milestone every seat is watching are below.
Free: The Weekly PM Market Brief — the panel's graded record, the week's biggest market-vs-model gaps, and what's spiking next. Every rung of this ladder gets graded in public as it settles; the Brief is where the receipts land. Free, Sundays. The signup box is at the bottom of this page.
New to event markets? How to bet sports on Kalshi covers the exchange mechanics; this page is the lunar layer. Prices carry an as-of stamp and refresh on our standard market-pulse cycle as the program moves.
The Market
| Venue | Kalshi, a CFTC-regulated exchange (18+, availability varies by state, as of August 2026) |
| The event | Kalshi event MOON: "NASA lands on the moon?", a cumulative deadline ladder |
| The contract | Official rule, verbatim: "If an announcement indicates that a manned NASA mission landed on the Moon after Issuance and by December 31, [year], then the market resolves to Yes." Settlement source: NASA. Flybys and orbital missions do not count, so Artemis 2's April flyby settled nothing |
| Prices | By end-2026: 1¢ · end-2027: 2.4¢ · end-2028: 20.5¢ · end-2029: 47.5¢ — as of August 7, 2026, 15:20 UTC. The 2026 and 2027 rungs have matched about 277,000 and 197,000 contracts respectively; the far rungs are thinner |
Read that far column again: the market says a crewed NASA Moon landing is an underdog even through the END of 2029, more than three years and several target dates away.
Hottest Prediction Markets Right Now
- 2028 Democratic presidential nominee · $168M traded
- 2028 U.S. Presidential Election winner? · $57M traded
- 2027 Pro Football Champion · $54M traded
- Pro Baseball Champion · $45M traded
- More tech layoffs in 2026 than in 2025? · $31M traded
- Bitcoin price at the end of 2026 · $30M 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.
What Changed In March: The Restructure
The card behind the panel carried the dated record. Artemis 2 launched April 1, 2026, carried Reid Wiseman, Victor Glover, Christina Koch, and Jeremy Hansen around the Moon on the first crewed lunar mission since Apollo 17, and splashed down April 10. It worked. And it no longer leads where the mission patches said it would.
In March 2026, following a safety-panel report, NASA Administrator Jared Isaacman announced a major overhaul framed explicitly as a move "to take down risk." Artemis 3, long marketed as the return to the surface, was redefined as a low-Earth-orbit demonstration targeted for late 2027: Orion will rendezvous and dock with uncrewed test versions of one or both commercial landers, SpaceX's and Blue Origin's, validating docking, life support, and spacesuits a few hundred miles up, where problems are survivable, instead of a quarter-million miles out. NASA named the crew on June 9, Bresnik, Parmitano, Douglas, Rubio, and published the docking and software protocols in July. The landing itself moved to Artemis 4, targeted for early 2028, with a potential second landing later that year.
Upstream of all of it sits the pacing item: SpaceX's uncrewed Starship lander demonstration, currently targeted for mid-2027, which requires orbital propellant transfer from a depot before it can even attempt the surface. NASA's own safety advisers have criticized the demo plan for deviating from "test like you fly" principles. And the program's schedule history is its own base rate: the landing was targeted for 2024, then 2025, then 2026, and now 2028; Artemis 2 itself slipped from 2024 to the April 2026 flight that finally flew.
Reading The Ladder: A Worked Example
Work the 2028 rung as the example: 20.5¢ for a $1.00 payout implies about a 1-in-5 chance the chain holds well enough to land crew by December 31, 2028; the panel's 25.2% blend says 1-in-4. Every seat modeled the same object: a three-link serial chain. First SpaceX's uncrewed Starship lander demo, targeted mid-2027 and gated on orbital cryogenic propellant transfer, which has never been demonstrated at scale. Then the Artemis 3 docking demonstration in late 2027. Then Artemis 4 itself. Serial means slips add instead of average, and this program's own slip ledger, landing targets of 2024, then 2025, then 2026, now 2028, is the base rate every model priced from.
"A serial chain — HLS refueling demo, Artemis 3 LEO demo, then Artemis 4 — must hold with under a year of aggregate slip; this program has slipped four years in five." (Claude Fable)
So why does the panel sit ABOVE the market on the far rungs? The date-math point: "early 2028" versus a December 31, 2028 deadline leaves roughly eleven months of buffer, so the 2028 rung does not require the target to hold, only to not collapse. Several seats raised their numbers in revision specifically on that argument. The 2029 rung then absorbs exactly one major slip, which multiple seats called the modal outcome for a chain like this, and that is how the blend crosses 50% there while the market sits under 48 cents. The half of the distribution beyond 2029 is not pessimism about rockets; it is what compounding three first-of-kind gates does to a calendar.
Prices refresh during the news cycle. A Starship demo result, a new NASA date, or a budget headline moves this board; we re-read the market and update the stamp when it does. The shape of the ladder carries the argument, and the dated quotes are the evidence.
The Panel: Eight Models, One Hazard Curve
Each seat received the same dated card, price-blind, and priced all four deadlines from one view of the dependency chain, monotone by construction. Then every seat read the others' anonymized reasoning and could revise. Post-revision numbers below.
| Model | By Dec 31, 2026 | By Dec 31, 2027 | By Dec 31, 2028 | By Dec 31, 2029 |
|---|---|---|---|---|
| Claude Fable | 0.2% | 1.0% | 25.0% | 53.0% |
| Claude Opus | 0.1% | 0.8% | 12.0% | 38.0% |
| Claude Sonnet | 0.1% | 1.5% | 28.0% | 58.0% |
| GPT | 0.1% | 0.8% | 24.0% | 58.0% |
| Gemini | 0.1% | 0.5% | 22.0% | 55.0% |
| Kimi K3 | 0.5% | 1.5% | 28.0% | 55.0% |
| GLM 5.2 | 0.1% | 1.5% | 33.0% | 65.0% |
| DeepSeek V4 | 0.1% | 1.5% | 30.0% | 65.0% |
| Blended Verdict | 0.2% | 1.1% | 25.2% | 55.9% |
Model estimates generated 2026-08-07. These are model estimates, not predictions of fact and not financial advice.
Every seat on this panel is graded against real market settlements — records to date: Claude Fable 84% on 334 graded calls · Claude Opus 85% on 385 graded calls · Claude Sonnet 83% on 366 graded calls · GPT 83% on 2,013 graded calls · Gemini 85% on 2,062 graded calls · Kimi 83% on 2,099 graded calls · GLM 81% on 2,138 graded calls · DeepSeek 79% on 2,145 graded calls. Recomputed daily; the full scoreboard is public.
More live boards from the same panel: 2028 presidential election odds · Elder Scrolls 6 release date · before 2028: 16¢ · Kalshi tech layoffs market
The interesting seat is the low one. Claude Opus put 12% on 2028 and 38% on 2029, far under the 22-33% cluster, and its reasoning is worth quoting because it is the honest bear case: the higher numbers "require believing this program's stated date is roughly as likely as not," which its own four-slips-in-five-years history argues against. The high seats, GLM at 33% and DeepSeek at 30%, are betting the restructure did what NASA said it was for: converting hidden schedule risk into visible, serial, testable steps. The spread is the real uncertainty, and it all keys on one flight.
Where The Panel Changed Its Mind
This board produced the most two-way revision round of the batch, with movement in both directions and a clear reason each way.
Claude Opus (2028: 7% → 12%): moved on the deadline-buffer argument "I had not weighted: ... the deadline is Dec 31, not 'early 2028', so the rung tolerates ~12 months of slip rather than the near-zero margin I assumed," then explicitly refused to join the higher cluster.
Gemini (2028: 12% → 22%): conceded its round-one number "was excessively punitive for minor slips" once the same buffer point landed.
DeepSeek (2028: 45% → 30%): the panel's optimist cut hard the other way: the 22-28% cluster and the base-rate argument "convinced me my 45% was underweighting the slip risk on the unprecedented orbital refueling milestone."
GLM (2028: 40% → 33%): moved on compounding: "three serial gates each with independent slip risk multiply."
Claude Sonnet (held 28%): the disagreement "is about interpretation weight ... not about any surfaced fact," so it held. Convergence from both directions toward the mid-20s is what an actual argument, rather than an anchoring drift, looks like.
What Would Change The Panel's Mind
- The uncrewed Starship lander demo flying, or visibly slipping, by mid-2027. Named by nearly every seat as the single milestone that reshapes the whole ladder; it is the chain's only never-demonstrated physics.
- A successful full-scale orbital propellant transfer. The prerequisite inside the prerequisite; a clean cryogenic depot-to-ship campaign before mid-2027 steepens the 2028 rung immediately.
- The Artemis 3 docking demo holding late 2027. The calendar link between the lander program and the crewed flight; a quarter's slip there consumes the 2028 buffer.
- Any new NASA date announcement. This program's landing target has moved three times already; another move in either direction settles the argument between the 12% seat and the 33% seat.
Settlement Timeline
| Date | What happens |
|---|---|
| Mid-2027 (target) | SpaceX uncrewed Starship lunar-landing demo, the chain's pacing item |
| Late 2027 (target) | Artemis 3 orbital docking demonstration |
| Early 2028 (target) | Artemis 4, the first crewed landing attempt |
| Dec 31, 2026 / 2027 / 2028 / 2029 | The four rungs settle in turn; every number here gets graded on the scoreboard |
More long-horizon boards from the same panel: 2028 presidential election odds · Elder Scrolls 6 release date · GTA 6 release date
The Bottom Line
The Artemis 3 that was sold as boots-on-the-Moon is now a docking rehearsal a few hundred miles up, and the landing itself belongs to Artemis 4 and a 2028 target that traders will only pay 20.5 cents for. Our panel's estimate sits a little higher at 25.2%, not because it trusts the date, no seat does, but because the contract's December deadline forgives most of a year of the slippage this program reliably produces. The next eighteen months of demos decide who was right, and every number on this page is on the scoreboard when they do.
To be explicit: nothing here is a prediction of fact or financial advice. It is a market price and a set of model estimates with a public graded record, built from dated, sourced facts. Every rung gets graded as it settles.
