The Stack: Every Layer Needed to Become Multiplanetary

The first problem is the hardest physics. Earth's gravitational well requires roughly 9.4 km/s of delta-v to reach low orbit. That number is fixed by the mass of the planet. Chemical rockets, burning a fuel and oxidizer to produce hot exhaust, are the only proven method of generating that delta-v.

The Tsiolkovsky rocket equation governs everything here. For a given exhaust velocity, the ratio of propellant mass to payload mass grows exponentially with the required delta-v. For a typical kerosene-oxygen rocket reaching orbit, roughly 85 to 90 percent of liftoff mass is propellant. The payload is a tiny fraction of what you start with on the pad.

For decades, the response to this was expendable rockets: build the vehicle, fly it once, throw it away. The Saturn V that took Apollo to the Moon cost roughly $185 million per flight in the dollars of the day, on the order of $1 billion once adjusted for inflation to today. The Space Shuttle ended up costing on the order of $54,000 per kilogram to orbit once all program costs were amortized across every flight. SpaceX changed the equation with Falcon 9. By landing and reflying the first stage booster, they demonstrated that a rocket's most expensive component could be recovered and reused. The reused-booster cost is now on the order of $2,700 per kilogram to low Earth orbit. Starship, currently in flight testing, is targeting full reusability of both stages with a stated goal of costs in the range of $67 to $100 per kilogram to orbit at high flight rates, a target SpaceX has repeatedly pushed back and has not yet demonstrated.

The difference between $54,000 and roughly $100 per kilogram, if Starship's target is actually reached, is the difference between a technology only governments can afford and one that can support an industrial economy in space. Every subsequent layer assumes launch costs continue to fall, though how far and how fast remains genuinely uncertain.

Reaching low Earth orbit is step one. From there, the solar system requires different propulsion for different missions. A transfer to the Moon requires roughly 3 to 4 km/s beyond LEO. Mars requires 3.6 to 4.3 km/s depending on the launch window, which opens for about 30 days every 26 months when Earth and Mars align. Missing a window means waiting two years.

Chemical propulsion works for these transfers but is inefficient. The efficiency of a rocket engine is measured in specific impulse: seconds of thrust per unit of propellant consumed. The best chemical engines (hydrogen-oxygen) achieve around 450 seconds. Electric propulsion, ion thrusters and Hall-effect thrusters, can reach the low thousands of seconds, and some designs push higher still. The tradeoff is thrust level: electric propulsion generates millinewtons to newtons rather than the meganewtons of a chemical engine. It is extremely efficient but extremely slow. For cargo and satellites it is ideal. For crewed missions on tight timelines it is not sufficient alone.

Nuclear thermal propulsion is the compelling middle ground. A nuclear reactor heats propellant, typically hydrogen, and expels it at high velocity. The NERVA program of the 1960s and 70s demonstrated working nuclear thermal engines on the ground, reaching a specific impulse of roughly 825 to 850 seconds in testing, close to double chemical performance, at thrust levels useful for crewed missions. None have ever flown. Regulatory and political barriers have kept nuclear propulsion grounded for decades despite the engineering being substantially proven.

For Mars, transit time matters enormously. A chemical transfer takes roughly 6 to 9 months. Crew are exposed to deep space radiation, microgravity bone loss, and muscle atrophy for the entire duration. A nuclear thermal transit is generally projected to cut that meaningfully, commonly cited in the 3-to-4-month range, reducing radiation exposure, consumables mass, and the psychological burden of confinement, though this remains a projection, not a flown result.

Three days from Earth, with a return trip possible within days if something goes wrong, the Moon offers a proving ground for every capability that Mars will require. A failed ISRU system on the Moon means a mission abort and a lessons-learned document. The same failure on Mars means many months before any help can arrive.

The most important lunar resource is water ice, confirmed in permanently shadowed craters at the poles by NASA's LCROSS impactor mission in 2009 and by subsequent orbital measurements. Water ice can be mined, melted, and electrolyzed into hydrogen and oxygen. Hydrogen and oxygen are rocket propellant. Oxygen is breathable air. A lunar propellant depot changes the economics of the entire solar system: instead of launching propellant from Earth's deep gravity well, you refuel at the Moon's much shallower one, roughly 2.4 km/s to escape the lunar surface versus 11.2 km/s from Earth.

The Artemis program's goal of returning humans to the lunar surface is explicitly framed as proving the technologies needed for Mars. ISRU, surface mobility, dust mitigation, radiation management, in-suit operations: all of these get tested on the Moon before a crew is committed to a months-long transit from which there is no early return. A parallel commercial lunar economy is emerging through NASA's CLPS program, seeding a private-sector supply chain for lunar operations.

On Earth, the atmosphere provides pressure, filters radiation, moderates temperature, and supplies oxygen. None of these are free off-Earth. Every one must be engineered, powered, and maintained continuously. A habitat failure on Mars carries no rescue window for at minimum several months.

Radiation is the first-order problem most habitat designs underestimate. Surface measurements from Curiosity's RAD instrument put the Martian surface dose on the order of a few hundred millisieverts per year, commonly cited around 300. The occupational limit for radiation workers on Earth is 50 millisieverts annually. The only practical solutions are burying habitats under regolith, building substantial above-ground shielding, or locating underground. The sci-fi image of glass domes on the Martian surface is a long-term aspiration sitting on top of an engineering problem that first-generation construction cannot yet solve.

Pressure is the second constraint. Mars atmospheric pressure is approximately 600 pascals, roughly 0.6% of Earth sea level. A pressurized habitat must maintain its pressure differential continuously through daily temperature swings on the order of 80 to 90 degrees Celsius, through dust storms lasting months, through the mechanical stress of airlocks cycling as crew and equipment move in and out.

Psychology may be the most underestimated constraint of all. ISS crew data, Antarctic winter-over studies, and submarine analog research all point in the same direction: the quality of the living environment has measurable effects on crew performance, immune function, and psychological stability that no pharmaceutical fully substitutes. Window access, spatial variety, natural light cycles, the ability to grow plants: mission infrastructure, not amenities.

Mars Dune Alpha, a 3D-printed analog habitat at NASA's Johnson Space Center designed by Bjarke Ingels Group and built by ICON, has run CHAPEA crew analog missions of roughly a year each to generate baseline data on exactly these questions. What volume does a crew member need? What spatial transitions matter? What light conditions preserve circadian rhythm?

Engineering a habitat keeps humans alive. Biology determines whether a colony is a permanent presence or a rotating outpost. The two questions are distinct, and the second is less understood than almost any other layer in this chain.

Reproduction in reduced gravity is an open problem, and it is one of the few layers in this piece with a genuinely new 2026 data point. Researchers at the University of Adelaide's Robinson Research Institute published a study in Communications Biology showing that simulated microgravity impairs sperm navigation through channels modeling the female reproductive tract, without affecting motility itself, producing roughly a 30 percent drop in fertilization rates in mice. Adding progesterone, a hormone the egg itself releases to help guide sperm, measurably improved navigation in human sperm under the same simulated conditions. Separately, a 2023 study published in iScience used a purpose-built device to thaw and culture frozen two-cell mouse embryos aboard the ISS: 720 embryos were split between real microgravity and an onboard 1g control, and the microgravity group developed into blastocysts with normal cell counts and gene expression, indicating that gravity is not required for the earliest stages of mammalian cell differentiation.

Whether a full-term pregnancy is viable in Martian gravity (0.38g) or Lunar gravity (0.16g) remains genuinely unknown. No mammal has been carried to term in reduced gravity, and the uterine and placental interactions involved in gestation have not been tested. Popular-science writing has floated various speculative names for a hypothetically space-adapted human lineage, but no such term is an established part of the scientific literature, and this piece will not invent one either. What is established is narrower and more useful: reproduction in altered gravity is unresolved, and it is the one layer in this chain whose failure mode cannot be engineered around the way a habitat or a power system can.

Population thresholds for technical survival are better understood, though the figures below come from two different bodies of research answering two different questions, not one continuous model. Jean-Marc Salotti, a professor at Bordeaux INP, published the first quantitative engineering estimate of a minimum settlement size in Scientific Reportsin 2020: 110 people, the point at which task-sharing lets individuals specialize instead of each person needing to master the full survival skill set. That figure is specific and well sourced. The much larger numbers sometimes cited for long-term genetic resilience come from a separate literature on minimum viable population size for isolated, multi-generational groups, most notably anthropologist Cameron Smith's work proposing a founding population in the 20,000-to-40,000 range for a multi-generational interstellar voyage. Other population-genetics estimates for a colony that must remain genetically healthy for centuries without immigration range more broadly, roughly 500 to 10,000 depending on the assumptions used. A Mars settlement fed by continuing arrivals from Earth is not the same problem as a sealed generation ship, so these numbers should be read as bounding the question, not as a single agreed answer.

Every system in a Mars habitat runs on electrical power. Life support, lighting, heating, computing, communications, ISRU processing, robotic systems, food production: all of it. There is no grid to connect to and no utility company to call. Power generation and storage must be entirely local, entirely reliable, and sized to handle peak demand without margin failure.

Solar power works on Mars but is substantially degraded relative to Earth. Mars receives about 43% of the solar flux Earth does, a direct consequence of sitting roughly 1.5 times farther from the Sun. Dust opacity reduces this further: during the 2018 global dust storm, atmospheric optical depth over the Opportunity rover's location reached roughly 10, an extreme haze level that blocks the overwhelming majority of surface sunlight. Opportunity, which relied on solar power, entered hibernation during the storm and was never heard from again.

Nuclear power is the answer for Mars. A fission reactor produces constant, weather-independent, 24-hour power regardless of dust, night, or seasonal variation. NASA's Kilopower project demonstrated a small fission reactor in 2018, the KRUSTY test, validating a design intended to scale to roughly 1 to 10 kilowatts continuously. Microreactor companies targeting terrestrial markets in remote locations and defense applications are building related technology, driven by commercial incentives that no longer depend solely on NASA contracts.

In-Situ Resource Utilization is the layer that separates an outpost from a civilization. An outpost consumes what it brings from Earth. A civilization produces what it needs from what it finds. The economics of Mars make ISRU existential: even under the most optimistic launch-cost projections, any consumable that can be produced locally probably should be.

The Mars atmosphere is about 95% carbon dioxide at low pressure. The MOXIE experiment on the Perseverance rover demonstrated in April 2021 that oxygen can be produced from the Martian atmosphere via solid oxide electrolysis, running at roughly 6 grams of oxygen per hour in its early tests and continuing to operate reliably across seven runs and multiple Martian seasons through the rest of 2021, toward a design target around 12 grams per hour. A full-scale system sized for a human mission would need to run continuously at far larger scale for months before crew arrive.

Water is more complex. Mars has water ice in the polar caps and subsurface deposits at mid-latitudes. Extracting it requires drilling, heating, and collecting vapor in a low-pressure environment with abrasive dust. Water covers three needs simultaneously: drinking water, oxygen production via electrolysis, and hydrogen production for rocket propellant.

Regolith, the loose rock and soil covering the surface, can be used as a construction material if properly processed. The perchlorates must be removed first. Processed regolith can be sintered or 3D printed into structural elements. ICON's research into printing structures from basalt analogs is directly aimed at developing a system adaptable to Martian feedstock.

The seed factory concept, a small robotic package that uses in-situ resources to replicate itself, is an appealing idea for driving the cost of local infrastructure down over successive generations of self-replication. Cost projections for delivering mass to the Martian surface vary enormously depending on architecture, from SpaceX's own optimistic 2016 target of roughly $140 per kilogram for a fully reusable, ISRU-refueled system, to far higher figures for expendable, one-off missions. No credible, sourced figure exists for what a mature self-replicating factory would eventually bring the effective cost down to; any specific number attached to that idea, including figures that have circulated informally, should be treated as an illustrative back-of-envelope exercise, not a modeled result. Settlements will likely follow a vitamins-and-info model in the interim, importing high-complexity components such as computer chips from Earth while manufacturing bulk materials and fuel locally.

Robots should build the landing pad, construct the initial habitat shell, activate the power systems, and verify that life support is functional before a crew commits to landing. The sequencing matters: every task completed by a robot before crew arrival reduces time crew must spend in suits doing dangerous surface work.

Autonomy is non-negotiable for Mars robotics. The one-way communication delay between Earth and Mars ranges from about 3 to 22 minutes depending on orbital positions. Remote control of a robot digging a trench or assembling a structure is impossible at these delays. Mars robots must complete complex, multi-step tasks with minimal human oversight, detect and recover from failures autonomously, and communicate results back for human review rather than requesting instruction at every decision point.

A human requires roughly 2,000 to 3,000 calories per day depending on body mass and activity level, plus adequate protein, fat, vitamins, and minerals for long-term health. A crew of six on Mars for two years, at a conservative 2,000-calorie planning figure, requires on the order of 8.8 million calories across the mission, not counting the psychological and nutritional benefits of food variety that long-duration isolation studies consistently show matter to crew performance.

Controlled-environment agriculture, growing plants in sealed chambers with recycled water, artificial lighting, and atmospheric CO2, is the near-term solution. The MELiSSA project run by ESA, initiated in 1989, has spent over three decades developing closed-loop life support systems that recycle organic waste into nutrients for plant growth.

The longer-term solution is more radical. Solar Foods has demonstrated the ability to produce protein from CO2, water, and electricity using microorganisms, with no sunlight or agriculture required. Their product Solein is produced entirely from atmospheric gases. On Mars, with its CO2-rich atmosphere and nuclear electricity, this process requires no imported feedstock at all. Cultivated meat, grown from animal cell cultures without raising animals, closes the protein gap further. These technologies are being developed for Earth markets, but Mars is the environment they were effectively designed for.

For long-term habitation, food and life support systems are increasingly framed around a bioregenerative model. A January 2026 paper in The Innovation lays out a four-stage paradigm for exactly this: biological pioneering, using extremophile microorganisms to establish baseline survival conditions; biological conversion, using microbes to biomine metals from regolith; biological fabrication, synthesizing construction materials through biological processes; and integration into a self-sustaining life support system. Related work in Frontiers in Microbiology in 2026 notes that biomining experiments such as BioRock and BioAsteroid have proven microorganisms can leach metals from rock under space-relevant conditions, though scale-up to industrial output has not yet been demonstrated.

Communication with Mars is governed by physics that cannot be engineered around. Earth-Mars distance varies from roughly 54.6 million to 401 million kilometers, producing one-way communication delays of about 3 to 22 minutes. Every communication is asynchronous. Mars crews cannot consult Earth for real-time guidance in emergencies. Mission control cannot remotely operate Mars systems in real time.

The infrastructure requirements include relay satellites in Mars orbit to maintain contact when the planet rotates away from the direct Earth line, sufficient ground station capacity on Earth, and eventually a deep-space internet protocol capable of handling delay-tolerant networking. NASA's Deep Space Network currently handles Mars communications for robotic missions. Scaling it for crewed missions and a settlement with significant data requirements is an unsolved infrastructure problem.

Within Mars itself, a local communications network is needed: surface relay stations, satellite constellations for global coverage, and high-bandwidth links between habitat modules and surface vehicles. Starlink is the Earth layer. The Mars layer has not been designed.

Before landing humans anywhere on Mars, you need the surface in detail. Slope, composition, subsurface ice presence, dust storm frequency, thermal properties, hazard density: all of this must be mapped from orbit before a landing site is committed. Once settlement begins, ongoing observation supports site selection for expansion, resource prospecting, and monitoring of dust storm development that might affect solar power or EVA planning.

Synthetic aperture radar sees through dust and functions at night, which makes it particularly valuable for Mars where global dust storms can obscure optical imaging for months. Hyperspectral imaging identifies mineral composition from orbit, supporting ISRU site selection. Thermal imaging maps subsurface ice through diurnal temperature variations.

The commercial Earth observation industry has driven down the cost of satellite imagery dramatically. Planet Labs images most of Earth's landmass daily. Capella Space and Umbra provide commercial SAR. These companies are building the observation infrastructure and data processing pipelines that translate raw imagery into actionable intelligence: the same technology stack that reconnaissance for a settlement program requires.

A spacesuit is the minimum viable habitat. It contains everything needed to keep a human alive in vacuum or near-vacuum: pressure shell, oxygen supply, CO2 scrubbing, temperature regulation, communications, and enough power for a multi-hour EVA, while allowing enough dexterity for a human to perform useful work with their hands.

Martian dust is fine enough to penetrate most seal designs over repeated use cycles, and its perchlorate content makes contamination into the habitat a genuine health risk. Bearing seals, glove interfaces, and boot closures all degrade with dust ingestion in ways that are well documented from Apollo and still not fully solved. Dust mitigation through electrostatic repulsion, improved seal design, or decontamination airlocks is an active engineering problem.

Suit mobility, the ability to bend, kneel, reach, and grip in a pressurized suit, has always been the fundamental ergonomic challenge of spacesuit design. Apollo astronauts fatigued quickly during EVAs partly from the muscular effort of working against suit pressure. Next-generation suits using advanced materials and improved joint designs aim to substantially reduce this fatigue load.

Mars has iron, aluminum, silicon, magnesium, calcium, and sulfur in its surface rocks in quantities sufficient for large-scale construction. The asteroid belt contains metallic asteroids with iron-nickel cores and platinum-group metal concentrations far higher than Earth's richest surface deposits. The Moon has titanium in its mare basalts, helium-3 in regolith implanted by the solar wind, and water ice in permanently shadowed craters.

The economics of asteroid mining depend entirely on whether extracted materials can be used in space rather than returned to Earth. Returning platinum to Earth from an asteroid requires overcoming Earth's gravity well twice. Using iron-nickel from a metallic asteroid to build space structures in orbit never has to lift that material out of Earth's gravity well at all. The value proposition shifts entirely when the customer is the space economy rather than the terrestrial one.

True independence from Earth requires the ability to manufacture: to take raw materials and produce functional equipment, replacement parts, construction elements, and eventually new technology without Earth supply chains.

Microgravity manufacturing enables processes impossible on Earth. Without convection currents, crystal growth proceeds without gravity-driven defects. Fiber optic preforms grown in microgravity have demonstrated lower attenuation than Earth-grown equivalents in early commercial trials. Pharmaceutical crystals, high-performance alloys, and exotic semiconductor materials are all candidates for orbital manufacturing. Varda Space Industries is developing reentry capsules specifically to return microgravity-manufactured products to Earth customers.

On planetary surfaces, additive manufacturing from local materials is the path to construction at scale. 3D printing concrete analogs from Martian regolith can produce structural elements without importing building materials. ICON's Project Olympus work for NASA is developing the construction technology that a Mars settlement will require.

The full vision is a closed manufacturing loop: ISRU extracts raw materials, robotics processes them, manufacturing converts them into equipment and structures, and the settlement grows without requiring a proportional growth in Earth supply missions. That is the threshold condition for a self-sustaining civilization.

Every prior layer is an engineering problem. This one is not. The legal framework governing what happens in space was written in 1967 and has not been substantively updated since. It was not written for private companies building settlements or extracting resources at commercial scale.

The foundational tension is between two incompatible frameworks. The Artemis Accords, a US-led bilateral framework, permit the extraction and private ownership of space resources, and had grown to roughly 70 signatory countries by mid-2026. The Moon Treaty of 1979 defines space resources as the common heritage of mankind and effectively prohibits private ownership. The Moon Treaty lacks the signatures of any major spacefaring nation, which does not resolve the conflict so much as defer it. As resource extraction becomes commercially real, the legal gap between these two frameworks becomes an operational one. Which law governs a mining claim on the lunar surface? Which court adjudicates a dispute between operators from signatory and non-signatory nations? No current answer exists.

The communication delay creates a second governance problem that physics imposes without appeal. One-way delay between Earth and Mars ranges from roughly 3 to 22 minutes. A round-trip lag that can approach 45 minutes at maximum distance makes centralized Earth governance of a Martian settlement operationally unworkable. Emergency decisions cannot wait for Earth's input. Legal disputes cannot be adjudicated on Earth timescales. Martian settlements will of necessity develop independent governance structures capable of making binding decisions without reference to Earth. The legal literature describes this trajectory as polycentric governance: independent decision centers operating without a central authority, analogous to maritime law in its early development. The eventual outcome, on long enough timescales, is plausibly some form of political autonomy. That is not speculation about distant futures. It is the logical endpoint of the physics.