An uncrewed SpaceX vehicle soft-lands intact on Mars by 2031
Mars lander a few years sooner.
Elon Musk, 2026-07-29 — X — posted directly beneath the crewed-Mars line
Original post
Sealed at 20%
Locked 2026-08-31. Confidence 55%. Uncertainty ±11 pts. Resolves 2031-12-31. Status: PENDING.
How the engine got there
Our published positions on this vehicle programme were sealed long before this statement: SPC-02 at 32% for an uncrewed Starship soft landing on the Moon by 2027-12-31, and CAL-16 at 68% that no uncrewed Starship HLS reaches the Moon before 2028. Mars is harder than the Moon in three specific ways rather than one. Transit is six to nine months instead of three days. Entry, descent and landing must cope with an atmosphere thick enough to demand heat shielding but too thin to land on aerodynamics alone, for a vehicle far larger than anything that has attempted it. And the whole architecture depends on orbital refuelling operating reliably, which is the pacing item for every ambitious Starship mission and remains undemonstrated at scale. The structural feature that lifts this well above our lunar number is the launch window. Earth–Mars transfer opportunities open roughly every 26 months, and the period to end-2031 contains two — late 2028 and early 2031. Two independent attempts at a hard problem is materially better odds than one, and that is the main reason this seals at 20% rather than near 10%. Against it: SpaceX has never operated a spacecraft beyond cislunar space, and Mars EDL has a poor success record even for agencies with deep planetary experience and far smaller vehicles.
What counts as correct
The statement is relative — "a few years sooner" than people on Mars in 5 to 7 years — so it is read as roughly 2 to 4 years out and resolved on the generous end of that window. CORRECT if, on or before 2031-12-31, a SpaceX vehicle achieves a controlled soft landing on the Martian surface with the vehicle surviving touchdown in a recoverable state, confirmed by SpaceX and corroborated by independent tracking or NASA. A hard impact, or a landing that destroys the vehicle, does not resolve this CORRECT.
We are wrong if
no SpaceX vehicle has soft-landed intact on Mars by 2031-12-31, including where one has launched and is in transit or in Mars orbit at that date.
Causal chain
Starship reached orbital flight and reuse testing
A vehicle of the required lift class now flies, which is the necessary precondition for any Mars payload of this size.
The launch-capacity objection to Mars ambitions largely dissolves.
Orbital refuelling remains the unproven pacing item
Trans-Mars injection for a fully fuelled Starship requires multiple tanker flights and reliable cryogenic transfer in orbit, never yet demonstrated operationally.
Every deep-space date depends on a capability with no flight record.
Launch windows are fixed by orbital mechanics
Transfer opportunities recur roughly every 26 months and cannot be advanced by funding or urgency; missing one costs slightly over two years.
Two discrete attempts exist inside the resolution period, which is what makes the probability non-trivial.
Historical precedent
1971-2021 — Mars landing attempt record
Roughly half of all Mars landing attempts have failed, across agencies with decades of planetary experience and vehicles orders of magnitude lighter.
2011-2020 — Falcon Heavy and Starship schedule history
SpaceX programmes have consistently delivered the capability and consistently missed the announced date, often by several years.
2018-2026 — Artemis lunar schedule
A nationally funded, urgency-backed lunar programme has slipped repeatedly — the basis for our own 68% seal that the uncrewed lunar step misses 2028.
Directly related seals
More on this desk
Browse the Space desk, the full projection registry, the calibration ledger, or the record by subject. Scoring is explained in how a prediction is made.