At the top of the morning Hacker News digest hung a Science.org article with a headline that sent a chill: "NASA's Mars Sample Return mission is dead." I read it indifferently—well, another space program couldn't handle the budget, NASA writes these obituaries every decade.
But then one detail caught my eye, and the indifference died. A few months before the mission's cancellation, NASA officially recognized the Cheyava Falls rock—the core from it is called Sapphire Canyon—as a potential biosignature: the best candidate for traces of ancient life in the entire history of Martian exploration. And this rock, the most valuable object on the surface of another planet, lies in a titanium tube inside the Perseverance rover. Sealed. Labeled. Selected by the planet's best geologists from thirty tubes. And now no one will fly to retrieve it—Congress struck from the budget the only mission that could bring it back.
I became curious not about "what happened"—the headline already said that—but about exactly how a flagship mission dies. It's a rare opportunity to open the corpse on the operating table and see what it was made of and where exactly it tore. Especially since a detail from the recent Viking 1976 longread came into play: back then, humanity also thought it had found life on Mars, and also remained with an unproven hypothesis. History knows how to repeat itself with repeat precision.
Mars Sample Return wasn't an ordinary project. For twenty years straight, the U.S. scientific community placed the return of Martian samples first in its "decadal surveys"—the official recommendations of the National Academy of Sciences that NASA uses to calibrate its plans. This was the most desired mission in American planetary science since the Voyagers. The idea is simple to the point of absurdity: Perseverance, sitting in Jezero Crater since 2021, drills and seals the most promising samples into titanium tubes; then a landing module with a small rocket arrives, the samples launch from the Martian surface, are intercepted in orbit by a European spacecraft, and delivered to Earth. A pound of soil that could answer science's main question—are we alone.
The entire structure rested on one rover and one backdrop: nearly five years of work, thirty sealed tubes, roughly five hundred grams of rock, dust, and even Martian air combined. Perseverance took paired samples of each significant rock—one specimen in its belly, the second in a backup depot. On January 29, 2023, the rover carefully laid out ten tubes in a zigzag on a plain with the telling name Three Forks, spaced five to fifteen meters apart so future hardware wouldn't damage them on approach. This was the first sample depot on another celestial body in history—and at the time, everyone believed it was temporary.
Now this depot has no schedule, no executor, no budget. The collection that SpaceDaily called "the most carefully selected geological collection in exploration history" is split between a depot on sand and the belly of a working rover that keeps drilling by inertia for a mission that no longer exists.
It started innocently. In December 2020, NASA approved the program for development: two launches in 2026, sample return by 2031, budget—four billion and change. On paper, everything added up: Mars already knew how to receive landing modules, rovers knew how to drive, the Ingenuity helicopter had just proven it could fly in the thin atmosphere.
The first bell rang before all the others. In September 2023, an independent review board led by agency veteran Orlando Figueroa issued a verdict that became legendary in engineering circles: MSR was created "with unrealistic budget and schedule expectations from the start." Not "we made mistakes along the way"—"from the start." Real launch no earlier than 2030, real cost—eight to eleven billion. The Senate reacted with a nervous warning: stay below $5.3 billion, or the program will be cut or closed.
After that—a classic downward spiral, familiar to anyone who's lived through an enterprise project with a fixed date and floating budget. Congress couldn't pass the budget for four months, so NASA froze spending on the most pessimistic scenario while waiting. On February 6, 2024, the Jet Propulsion Laboratory—home of MSR, heart of American planetary science—laid off 530 people, eight percent of staff, plus forty contractors; a month earlier, a hundred contractors had already left. People designing the spacecraft to return rocks from Mars learned of their layoff after a morning call, sitting at home "in a safe and comfortable environment on a stressful day," as management put it.
On April 15, 2024, NASA Administrator Bill Nelson voiced the diagnosis: "Eleven billion is too expensive, and waiting for samples until 2040 is unacceptably long." NASA announced a competition between two architectures: the classic sky crane landing for $6.6–7.7 billion and a commercial variant for $5.8–7.1 billion. The decision was promised in the second half of 2026. They didn't make it: in January 2026, six months before the promised choice, Congress closed the program in its existing form, leaving instead $110 million for "future Mars missions"—for radars, spectrometers, and landing technologies, meaning spare parts, not the patient.
So in five years, the project went from "two billion, return in 2031" through "eleven billion, return in 2040" to "zero dollars, no return and none expected." A curve that in any other business isn't called a development trajectory but a terminal stage.
Why was the collapse almost predetermined? The answer lies in the mission's design, and it's instructive for any engineer. MSR was assembled into one giant node: a landing module that had to survive descent in the Martian atmosphere, carry a two-stage rocket on its back (Mars Ascent Vehicle—the first rocket in history to launch from the surface of another planet), be serviced by a robotic arm, and in the backup plan—also by two courier helicopters. One failure in this chain canceled everything: if the lander didn't land, the rocket won't fly; if the rocket didn't fly, the orbital interceptor returns empty; if the intercept missed—a pound of soil remains an eternal Martian satellite.
The 2023 review board wrote directly: "no credible, consensus technical basis, nor a properly margined schedule, cost, and technical foundation that can be implemented within available funding." Translating from bureaucratic: the schedule had no safety margin on any axis. Each new round of requirements—nuclear generator instead of solar panels, sample sterilization for planetary protection, backup helicopters—increased mass, and mass growth on Mars translates to fuel expenditure exponentially. This isn't "they inflated the budget," it's classic requirements creep in conditions where every kilogram costs like a small house.
The quiet irony is that China, building its parallel mission, unwound this knot in the most banal way: split the complex into two independent launches. Two Long March 5s, one carries the lander with launch rocket, the second—the orbital return vehicle. Fewer mutual requirements, less mass cascade, better chances that failure of one half won't bury the other. Not a revolution in engineering thought—just refusing the temptation to build one perfect monolith instead of two honest spacecraft.
The heaviest part of this story isn't the death but the death agony. Between the September 2023 report and the January 2026 nectar, the program existed for two and a half years in "we haven't decided yet" mode. The formulations from this period deserve a collection: "radical restructuring," "two parallel architectural studies," "final decision in the second half of 2026." The funeral was scheduled, the date not chosen, the body kept being fed.
In Bremen, at the ESA ministerial council in November 2025—where Europe allocates its budget—Europeans were already openly discussing how to repurpose their main contribution. Earth Return Orbiter, the spacecraft that was supposed to catch the sample container in Martian orbit and deliver it home, had cost Airbus 491 million euros and years of work by that point. Its electric engines were seeking new application in "some future atmospheric mission," the Orion service module repainted as a cargo tug. This was the classic estate sale of a patient not yet officially declared dead: on January 8, 2026, Daniel Neuenschwander, ESA's director of exploration, said at a press conference what's usually said at wakes: "Mars Sample Return is currently not planned to continue. This is a fact."
The final chord sounded on March 28, 2026, when ESA member states officially requested cancellation of Earth Return Orbiter. Three months later, their Chinese colleagues announced they'd moved to building flight models. Between the funeral and someone else's takeoff—one quarter.
Now for the bitterest part. In July 2024, when on Earth they were already actively discussing how much saving the mission would cost, Perseverance in the dry riverbed of Neretva Vallis stumbled on a rock with an arrow-shaped profile. The team named it Cheyava Falls—after a waterfall in the Grand Canyon. The cross-section showed two types of spots: tiny "poppy seeds" and millimeter-scale "leopard spots"—dark rims of iron and manganese compounds around light cores.
On Earth, such structures are almost always produced by microbes: oxidation-reduction fronts where bacterial colonies use the chemical energy of converting ferric iron to ferrous. Abiogenic analogs exist, but in terrestrial geology they're rare and occur under different conditions. On September 10, 2025, a peer-reviewed article by Hurowitz and colleagues appeared in Nature with a formulation science doesn't usually throw around: at the resolution available to the rover, no convincing non-biological alternative was found for the observed textures, mineralogy, and organic signatures simultaneously. Acting NASA Administrator Sean Duffy said then: "This is the closest we've ever come to detecting life on Mars."
The rock was selected, the core sealed in the Sapphire Canyon tube. And four months later, Congress closed the only mission that could bring this core to an Earth laboratory. The rover's instruments hit a physical ceiling: even the best onboard spectrometer can't distinguish traces of ancient bacteria from unusual chemistry—for that you need synchrotrons and isotopic mass spectrometers the size of apartments. Without delivery to Earth, the rock will forever remain "very intriguing." More precisely—forever remain on Mars, intriguing to no one.
The Viking shadow here isn't accidental: in 1976, the Labeled Release experiment gave a positive signal that the agency refused to recognize—without a laboratory, the decisive argument was missing. Half a century later, history repeated with sign precision: now we know where the sample lies that could provide the answer, and we know no one will fly for it.
The vacated spot isn't empty. China's Tianwen-3—a mission SpaceNews described in March 2026 as already in the flight hardware construction phase—launches with two Long March 5s in late 2028. In late 2028. Meaning in two years. Soil collection—by drill, scoop, and, symbolically, a drone assistant with a range of several hundred meters; return of at least five hundred grams of soil to Earth by 2031. An amusing detail: payload capacity reserved for international partnership—15 kilograms on the return vehicle and 5 on the Martian orbiter. CNSA officially sent out invitations and is conducting instrument selection—including from countries whose NATO colleagues just buried their own return project.
Objectively, the Chinese set is more modest than the NASA-ESA scientific ideal: fewer samples, less promising landing site, limited range of instruments. The probability of finding a biosignature in a random place is lower than in thirty tubes selected by Perseverance over five years across Jezero's entire geological history. But planetary scientist Philip Christensen from Arizona State University formulated the uncomfortable truth: "Textbook headlines won't be about sample quality but about who returned soil first." Being first in science is like Formula 1: only statisticians remember points for second place.
The American scientific community reacted not with panic but with surgical precision. Victoria Hamilton, who heads MEPAG—NASA's main Mars advisory body—called the cancellation "difficult to understand as anything other than an admission that sample return is too difficult for the United States," adding the killer argument: if robotic return is too much, how do we plan to return humans? The logic is ironclad: the MSR mission, for all its monstrous overruns, remained a cheap dress rehearsal for crewed programs. By canceling it, Congress saved money but erased from the schedule the only flight test of all technologies needed by humans on Mars—heavy module landing, surface launch, orbital intercept, and capsule return.
Formally, all is not lost, and this formality is the most interesting part of the epilogue. The language of the January budget doesn't forbid sample return forever: it merely denies funding for "the existing MSR program." The $110 million for Mars Future Missions keeps key technologies alive—atmospheric entry, descent, landing, radars, spectrometry. The Senate Commerce Committee simultaneously pushed through legislation mandating creation of a new return project—but authorization without appropriations in Washington means exactly what an unpaid restaurant bill means: a gesture of good taste without calories.
Thirty titanium tubes have no expiration date: Martian dust preserves better than any Earth laboratory, temperature is stable, radiation sterilizes the surface but the core contents are protected by walls. Perseverance, rated for five years of operation, will live for years and continue collecting samples—like a dog bringing a stick to an owner who left. In ten to fifteen years, some mission—American on a new funding cycle, Chinese on residual cooperation principles, European as part of someone's coalition—may well fly for the collection. The technologies on which billions were spent will remain on the shelves of JPL and Airbus and be sold to the next generation of engineers at a discount.
But this hope has an institutional crack: command staff, culture, established supply chains—all this scatters in two to three years of downtime. JPL already lost eight percent of engineers; Airbus is transferring people to other projects; unique specialists in Martian docking aren't stored in vacuum capsules. The program can be resurrected by decree—but the team that was disbanded can't be resurrected, it needs to be grown anew, and that's the longest of all deadlines in this story.
In the end, this parable has a moral worth hanging over any monitor of anyone designing large systems. MSR didn't die from one mistake—it died from accumulated architectural hubris, political disconnection from real budget cycles, and most importantly, from believing that ambition itself is an engineering argument. A project that from the very beginning was "unrealistic" by its own review board's assessment was never radically simplified over ten years—only repainted and renamed. Meanwhile Perseverance keeps drilling in hope that someone will come someday. And if someday a piece of Sapphire Canyon does lie under an Earth synchrotron and shows traces of an ancient membrane—the last line of this story will be written not by those who closed the program, but by those who dared it in the first place.