Why Mars hasn't happened yet

For most of my professional life there has been an unbroken stream of confident announcements about when humans will set foot on Mars. Bush 41 said the 2010s. Bush 43 said 2030. Obama said the mid-2030s. SpaceX has said, variously, 2024, 2026, 2029 and "before the end of the decade". None of these has come close. There is, in 2026, no funded program in the world that has a credible architecture, a serviceable transit-vehicle design, a working surface habitat, and the political will to deploy any of it. There are several ambitious slideware decks. The interesting question isn't when this gets fixed but why it keeps not getting fixed.

The architecture problem

Every credible Mars architecture study of the past thirty years — NASA's Design Reference Mission 5.0 (2009), the various Mars Architecture Study Team revisions, the Aldrin-Mars-cyclers white papers, the Zubrin Mars Direct architecture and its descendants, the SpaceX Starship-based plans — converges on roughly the same logistics envelope. To put a single small crew on the Martian surface for a meaningful stay (the typical "long-stay" mission profile is 500-600 days at Mars, with launch windows opening every 26 months), you need to deliver somewhere between 100 and 250 tonnes of useful mass to the Martian surface across multiple cargo flights, plus get the crew there and back.

The Mars surface delivery problem is more difficult than the Lunar surface delivery problem by roughly two orders of magnitude. Mars has an atmosphere thick enough to make aerodynamic heating a serious problem (every previous successful Mars lander has used a heat shield, parachute, and either airbags, retropropulsion, or a sky-crane), but thin enough that parachutes alone don't slow a heavy lander to a survivable touchdown. The current state of the art for Mars EDL — entry, descent, and landing — caps out at around one tonne of usable payload on the surface for the larger US missions (Curiosity, Perseverance), or much less for everyone else. Nobody has ever landed anything heavier than that on Mars, and the engineering between a one-tonne rover and a fifty-tonne crew habitat is not a smooth scaling problem.

Starship is the first realistic candidate vehicle for the heavy-mass Mars EDL problem. The proposed solution — a fully fuelled Ship doing supersonic retropropulsive descent through the Martian atmosphere — has been simulated extensively and tested zero times. The vehicle has not yet completed a clean orbital flight at Earth. The on-orbit refuelling architecture that the Mars version of Starship depends on has not been demonstrated. The Martian propellant production needed for the return trip — methalox synthesised from atmospheric CO₂ and either local water ice or imported hydrogen — has been demonstrated only at gram-scale on Perseverance's MOXIE experiment. The gap between any of these capabilities and "we are ready to launch a crew" is large enough that not even SpaceX's most optimistic public timelines suggest pre-2032 humans-on-Mars.

The science problem

The case for crewed Mars missions has always run on a mix of three rationales — science, settlement, and inspiration. Settlement and inspiration are not really arguable on cost-benefit terms; people either subscribe to them or they don't. The science argument, on the other hand, is genuinely arguable, and it has been quietly losing ground for a decade.

Robotic missions have gotten dramatically more capable. Curiosity (operating since 2012) and Perseverance (since 2021) have between them executed roughly the science programme that NASA's 1990s human-Mars planning expected an astronaut crew to deliver, at substantially less than 1% of the cost of a crewed mission. The robotic programs do this with fewer life-support constraints, lower launch mass, longer mission durations, and far better risk tolerance — Perseverance is rated for 1.5 sols of operations and is in its sixth year. The Mars Sample Return programme that's intended to bring back the Perseverance cache will deliver the first ever samples of unaltered Martian rock returned to Earth — exactly the science deliverable that justified crewed missions in the original justification papers.

Crewed missions still have one genuinely unique science capability: a human geologist on the Martian surface can, in a couple of weeks of fieldwork, do more interpretive geology than a rover can do in a decade. The on-the-ground judgement, the ability to find unexpected interesting rocks and choose what to sample, the rapid hypothesis-test loop — these are all things robots can't do well. But Mars Sample Return delivers most of what those geologists would have been used for at far lower cost. And if you want to put humans on a planetary surface for fieldwork purposes, the Moon is two orders of magnitude cheaper, three days from Earth, and offers thousands of unsampled square kilometres of unexplored terrain. The science case for Mars over Moon, dollar for dollar, is no longer obvious.

The political problem

The third reason this hasn't happened is that nobody has been willing to pay for it. The realistic cost of a NASA-led crewed Mars program, including the development of new EDL hardware, life support, surface systems, transit vehicle, and the multiple precursor cargo flights, is somewhere between $300 billion and $1 trillion in 2024 dollars over a fifteen-year program. NASA's total budget over the same period would be about $370 billion at current spending levels. The math has never worked.

Apollo cost roughly $260 billion in 2024 dollars and ran 1961-1972. NASA's budget peaked at over 4% of federal spending and stayed above 1% for the entire decade. Today NASA's budget is below 0.5% of federal spending and has been roughly flat in inflation-adjusted dollars for thirty years. The political conditions that made Apollo possible — a Cold War crisis, an explicit superpower rivalry, a Kennedy-era domestic policy environment, a willing Congress — do not exist now and are not on the visible political horizon.

SpaceX's Mars timeline depends on Elon Musk's willingness to fund the program from Starlink revenues for as long as it takes. That's a real funding source — Starlink is generating something like $9-10 billion in annualised revenue and growing — but the SpaceX architecture also depends on hundreds of Starship flights worth of orbital refuelling, the development of a Mars surface infrastructure from scratch, and the engineering and operational discipline to keep crew alive for a 1,000-day round-trip. The cost to SpaceX of doing this on its own balance sheet, with no government revenue, is plausibly well above $200 billion. That's a meaningful fraction of the entire SpaceX market valuation. It is not impossible, but it is not happening on the schedule that has been publicly communicated, and the company has been increasingly quiet about the operational timeline over the past two years.

What's actually plausible

What is plausible over the next fifteen years:

Mars Sample Return delivers the first samples of returned Martian material to Earth around 2033 (the current programme schedule, after the 2024 architecture descope). This is genuinely a flagship science achievement and would make the strongest scientific case for the planet that has ever existed.

SpaceX flies one or more uncrewed Starships to Mars surface in the 2028-2031 launch windows. Whether they survive landing is the open question. Even if they don't, the precursor flights generate engineering data that no other organisation has at substantial scale.

NASA and ESA continue robotic programmes at Mars — possibly Mars Life Explorer, possibly an Ice Mapper, possibly a sample-cache rover — but at much lower cadence than the Lunar program because the institutional priorities have shifted firmly toward Artemis and the cislunar economy.

China lands robotic missions at Mars on its own cadence. Tianwen-3 (sample return, mid-2030s) is officially in development. China is not currently committed to a crewed Mars program, and is unlikely to make that commitment before completing the crewed Lunar landing it has scheduled for the late 2020s.

What is implausible: a NASA crewed Mars landing before 2040. A SpaceX crewed Mars landing before 2035. A Chinese crewed Mars landing before 2040. Any "Mars settlement" of more than a small precursor crew before 2050.

None of this is a bad thing. The Mars exploration programme is, despite all the unmet timelines, in better shape today than it has ever been in the history of the field. Eight active spacecraft are operating at Mars. The science returns of the past decade are spectacular. Robotic sample return is on the books. Starship's success at Earth, when it comes, will reset the entire conversation about deep-space architecture. The reason humans haven't walked on Mars yet is not that the field is failing. It's that the field has correctly recognised which Mars activities are worth funding and which are not. The crewed surface mission, on every honest cost-benefit analysis, has been the latter for thirty years.

That can change. It might change. The political conditions that produced Apollo took roughly four years to materialise and four years more to be funded. If something similar emerges — a Chinese crewed lunar program acceleration, a Cold War-style geopolitical pivot — Mars could move from twenty-year time horizon to twelve. But the default trajectory, today, points to a 2040s human landing at the very earliest. The honest answer to "when will humans walk on Mars" is "later than the people who keep promising it would have you believe."