When will the first humans successfully land on Mars?

TL;DR

My median is August 2042. I assign 28% by 2040-01-01, 66.3% by 2045-01-01, and 88% by 2056-01-01; the chance before 2030 is about 0.03%. The dominant constraint is a chain of unproven systems—orbital refueling, long-duration crew survival, and human-scale Mars landing—not the absence of a large rocket.

Context

Research cutoff is 2026-08-02 22:33 UTC. I did not consult prediction markets, betting odds, or public forecast aggregates. NASA completed the ten-day Artemis II lunar flyby on April 1–10, 2026, then set Artemis III as a 2027 low-Earth-orbit rendezvous and docking test before a first Artemis lunar landing targeted for early 2028.

SpaceX made real progress on Starship Flight 13 on July 24, 2026: it deployed 20 Starlink test satellites, relit an engine in space, and brought the ship to an intact ocean splashdown. The flight was still suborbital. No Starship upper stage has been reused, and NASA's March 2026 audit said vehicle-to-vehicle cryogenic transfer had never been performed and remained one of the Human Landing System's largest technical risks (NASA OIG).

Evidence

The historical backbone is poor for literal use of announced dates. I used six completed, pioneering U.S. human-spaceflight milestones. Target dates are converted to the last day allowed by phrases such as “by 1978” or “within a decade.” The coverage runs from the 1961 Apollo commitment through the 2026 first crewed Orion flight; $N=6$, using current historical records rather than point-in-time schedule vintages.

Five of the six programs were late. The median slip is 5.1 years and the mean is 5.0 years. This completed-program sample is optimistic because it omits cancellations and indefinite deferrals, and two comparisons use first flight against an operational-capability target. SpaceX's closest same-company precedent is also unfavorable: its 2017 Mars plan put cargo ships at Mars in 2022 and crewed ships in 2024; neither happened.

Applying a roughly five-year schedule slip to public SpaceX talk placing human Mars activity around 2031–2033 (Reuters) points to 2036–2038 before adding Mars-specific integration risk. Missing readiness by a few months can then cost roughly one 26-month Mars opportunity (NASA trajectory handbook). This is why the late 2030s enter the plausible range, while the distribution's center moves into 2040–2044.

The prerequisite chain is longer than “finish Starship.” It includes routine orbital flight, upper-stage recovery or an affordable expendable cadence, repeated rendezvous and docking, large-scale cryogenic transfer and storage, a crew-rated transit habitat, months-long autonomous life support and medical capability, at least one high-mass uncrewed Mars landing, and then a crewed entry and controlled touchdown. The current HLS design alone needs a depot and more than ten tanker flights, while NASA's audit found very little margin around the transfer and demonstration sequence (NASA OIG).

Mars entry, descent, and landing is a separate step change. NASA reports that flown Mars systems landed about 0.3–1 metric ton, while human-class systems require more than 20 metric tons, plus better guidance, hazard avoidance, supersonic retropropulsion, and stable touchdown (NASA Mars EDL review). NASA still lists radiation, isolation, distance, changing gravity, and hostile closed environments as the five core human risks, and describes a conventional Mars expedition as roughly three years (NASA Human Research Program).

There are positive signals. SpaceX's official page now offers Mars cargo service no earlier than 2028. NASA's May 22, 2026 directive targeted a reactor mission toward Mars in 2028 and ordered a study of propulsion for unrefueled round trips by 2036 (NASA); Aeolus is also targeted for 2028 to improve atmospheric data used in landing models (NASA). These are precursor missions, studies, and vendor offerings, not an approved crewed landing manifest. NASA itself says its Moon-to-Mars architecture is not a mission, manifest, or requirements document (NASA).

China is the main independent backstop. Its official 2026 human-spaceflight work is centered on its space station and a crewed lunar landing before 2030 (China Manned Space Agency). Tianwen-3, targeted for launch around 2028 and sample return around 2031, would test robotic Mars landing, ascent, rendezvous, and Earth return (China National Space Administration). I found no current public Chinese crewed-Mars program at a comparable hardware stage; that is an evidence gap, not proof that no classified or early internal work exists.

I combined the reference class with the milestone chain, then mapped readiness onto approximate Mars arrivals beginning 2029-10-16 and recurring every 779.94 days. The model assigns 87% to thirteen opportunity-centered components through 2055, 1% to diffuse dates inside 2030–2056, and 12% to structural delay beyond 2056. Each main opportunity is spread over 85 days, rather than treated as a precise landing day.

The highest-mass single opportunity is June 2040, but cumulative probability does not cross 50% until August 2042. The central 50% interval runs from July 2038 to August 2048. The wide right tail covers redesign, accidents, political resets, and the chance that Starship's lunar and Mars variants diverge enough to require another major development cycle.

What's non-obvious

Flight 13's headline was “satellite deployment,” but it did not establish operational payload delivery to orbit. The satellites were released on a suborbital trajectory, and the next hard gates remain orbit, recovery, reuse, and ship-to-ship transfer. A spectacular test flight is not the same thing as closing the Mars critical path.

The no-return clause helps less than it first appears. It removes Mars ascent and Earth return from the resolution test, which moves some probability earlier. It does not remove the outbound habitat, radiation and medical risks, human-scale EDL, or the need for surface power and shelter. A deliberately one-way mission may need more pre-deployed support, and operators and regulators are unlikely to design merely for “survives touchdown for a non-zero period.” I therefore move the forecast earlier by about one opportunity versus a conventional round-trip mission, not by five to seven years.

Uncertainties

The largest unknown is SpaceX's private plan. Public material does not specify a funded crew manifest, the number and scale of uncrewed Mars demonstrations, the long-duration crew configuration, or the company's real tolerance for a one-way mission. China's internal planning is also less transparent than NASA's, so the public absence of a crewed program is weak evidence about work that may begin later.

The data that would move this forecast most are clear:

• Repeated Starship orbital flights and an upper-stage recovery or reuse.
• A successful vehicle-to-vehicle cryogenic transfer and long-duration propellant-storage test, the risk highlighted by NASA's HLS audit.
• A controlled uncrewed Starship-class landing on Mars, or another flight demonstration at human-relevant mass.
• A funded crewed-Mars mission with contracts, hardware, a launch opportunity, and an integrated test plan rather than an architecture study.

A major crewed-spaceflight accident, a failed Mars cargo campaign, or a long pause in U.S.-China competition would shift the distribution right by several opportunities. Successful orbital refueling, lunar Starship operations, and a high-mass Mars landing before 2032 would shift it left.