When will the first humans successfully land on Mars?

TL;DR

My median forecast is March 2042. The distribution gives a 13% chance before 2037, a 61.2% chance before 2043, and a 13% chance after January 1, 2056. The decisive test is not another Starship splashdown; it is orbital ship-to-ship refueling, followed by a successful human-class Mars landing demonstration, without losing one or more 26-month Mars opportunities.

Context

SpaceX remains the main route to an early landing. Its thirteenth full-stack Starship test flew on July 24, 2026, deployed operational satellites, relit an engine in space, and left the upper stage intact after its ocean landing. But it was still a space-skimming test, not a normal orbital mission. SpaceX's June 5, 2026 prospectus described upper-stage recovery and in-orbit propellant transfer as future milestones and expected orbital payload service only in the second half of 2026.

NASA has resumed crewed deep-space flight and now plans a 2027 Earth-orbit lander demonstration before an early-2028 lunar landing. Yet NASA's current Moon-to-Mars Architecture explicitly says it is not a mission, manifest, or requirements document. China's public program focuses on a crewed lunar landing before 2030 and the robotic Tianwen-3 Mars sample return, planned to launch around 2028 and return around 2031. I therefore see no actor with a funded, integrated crewed-Mars landing mission at the cutoff.

Evidence

The cleanest development-duration reference class comes from a NASA Inspector General review published November 15, 2021. Its complete sample contained five recent programs: SpaceX Cargo, Orbital Cargo, Boeing Crew, SpaceX Crew, and SLS. The mean time from contract award to first operational flight was about 8.5 years, and the mean gap from test flight to operations was about two years. The same report says Apollo's lunar lander took roughly six years, but under far higher inflation-adjusted funding. This is a small, heterogeneous sample with projected values in its 2021 vintage, so it is an anchor rather than a statistical law.

The lunar Starship program is a live test of that reference class. A March 2026 NASA Inspector General audit found that SpaceX's contracted lunar lander work had already slipped at least two years. The same architecture requires a depot and more than ten tanker launches, with aggregation starting more than 200 days before crew launch and a target rate of one tanker every six days. The audit said vehicle-to-vehicle cryogenic transfer had never been done and the needed launch-pad turnaround had not been demonstrated. A July 23, 2026 GAO assessment still listed cryogenic fuel management among the top HLS risks. (techport.nasa.gov)

Flight 13 is real progress. It shows that Starship development is not vaporware and that commercial payload demand can finance rapid iteration. It does not yet show orbital rendezvous, docking, large-scale refueling, long-duration propellant storage, upper-stage reuse, crewed operation, or Mars entry and landing. Those gates are sequential. A delay at one gate often delays the whole mission rather than merely adding a few months.

Mars landing technology is the strongest reason not to extrapolate directly from the Moon. NASA's human Mars entry, descent, and landing review records 12 successes in 19 attempted robotic landings. All flown Mars systems landed between 0.3 and one metric ton, while NASA's human-class concepts require more than 20 metric tons. The report also says Mars EDL cannot be fully tested under Earth-analog conditions. Lunar landings provide useful propulsion and surface-operations data, but they do not reproduce Mars's atmosphere, heating, wind, dust, or supersonic retropropulsion.

The crew and surface systems are also unfinished. NASA identifies radiation, isolation, distance from Earth, altered gravity, and hostile closed environments as interacting human-spaceflight hazards. NASA's Earth Independent Operations Crew Interaction project, updated May 27, 2026, was at technology-readiness level 3, targeting level 6 by September 2030. Its Fission Surface Power project, updated July 15, 2026, was at level 4, targeting level 8 by September 2028. These levels describe individual projects, not total mission readiness, but they show that basic Mars-enabling systems are still being developed.

There are credible positive weak signals. NASA awarded a Mars Telecommunications Network contract on September 1, 2026, with a maximum value near $700 million, delivery by December 31, 2028, and planned Mars operations in 2030. NASA's fiscal-year 2027 request also includes $200 million for a near-term human-class Mars-lander demonstration. The first is a real contract; the second is still a budget request. I treat both as evidence that Mars preparation is becoming concrete, but neither closes the transportation, crew, or landing architecture.

SpaceX's own public messaging also separates cargo from crew. Its current Starship page offers Mars cargo flights no earlier than 2028 but gives no committed human date. Elon Musk said in February 2026 that the Moon had become the overriding near-term priority, while Mars work would begin in roughly five to seven years. I read that as evidence of durable Mars intent and near-term lunar prioritization, not as a calibrated landing schedule.

China reduces the chance of indefinite delay but adds little early probability. Its official crewed-space program is building toward the Moon, while Tianwen-3 would demonstrate robotic Mars landing, ascent, rendezvous, and Earth return at far smaller scale. China's 2024–2050 space-science plan supports lunar and Mars research but does not publish a funded crewed-Mars surface manifest. Public silence does not prove internal planning is absent, so I assign China substantial probability in the 2040s and 2050s rather than treating it as a near-term favorite.

I modeled landings around approximate arrival opportunities separated by 779.94 days. NASA says efficient alignments recur about every 26 months and a typical cruise lasts about 200 days. For opportunity center $c_k$, probability weight $w_k$, and timing width $\sigma_k$, the finite-horizon cumulative distribution is:

$$F(t)=\sum_k w_k\Phi\left(\frac{t-c_k}{\sigma_k}\right),$$

where $\Phi$ is the standard-normal cumulative distribution. I used widths of 75 to 130 days because actual touchdown dates depend on trajectory, vehicle performance, and landing season. The model allocates probability as follows:

The lower tail represents unusually fast, risk-tolerant SpaceX execution. The middle represents lunar Starship maturity, one or more uncrewed Mars landings, and eventual NASA or international support. The upper tail covers failed precursors, architecture redesign, fatal accidents, funding reversals, corporate reprioritization, or a later Chinese-led mission. The resulting 10th, 50th, and 75th percentiles are approximately March 2036, March 2042, and October 2046.

What's non-obvious

The permissive resolution rule does not make an early landing easy. A safe return is unnecessary, and nobody even has to leave the spacecraft. But the decision to launch a crew occurs before touchdown. A government or major company will still demand reliable transit life support, landing confidence, communications, surface power, supplies, and a credible survival plan. I moved the forecast earlier than I would for a successful round trip, but only modestly.

The Moon-first strategy cuts both ways. It diverts near-term effort from Mars, but it also forces SpaceX and NASA to solve high-cadence launch operations, orbital refueling, docking, deep-space crew interfaces, surface power, and large-lander operations. The new Mars communications contract and human-class lander-demonstration request mainly reduce the post-2056 tail. They do much less for 2031–2036 because Mars EDL and long-duration crew systems remain separate critical paths.

Uncertainties

• SpaceX publishes no detailed internal schedule for Mars crew systems, surface equipment, or certification. The public Mars offer gives an earliest cargo date, not the maturity or testing plan behind it.
• NASA's 2028 lunar target and fiscal-year 2027 Mars request remain exposed to technical reviews, appropriations, and future administrations.
• China's public material gives firm lunar and robotic-Mars goals but little usable information about a crewed Mars program. The absence of a published manifest limits how precisely its long-run contribution can be modeled.
• The opportunity centers are approximations, not trajectory solutions. Higher-energy transfers could shift arrivals by months, while a failed precursor could move the crew mission by one or several whole 26-month cycles.

The most valuable new evidence would be successful Starship-to-Starship propellant transfer, actual upper-stage reuse, an uncrewed lunar Starship landing, and then a controlled human-class uncrewed Mars landing. Until those events occur, the forecast should remain broad and centered in the early 2040s.