Building a Martian city with local materials

Building in a world without an industrial base

A Mars city must be able to lose one room without losing its atmosphere

The first principle of permanent habitat design is not “never fail.” No real system meets that standard. The goal is to keep a local failure local. On Earth a building can evacuate to the outside. On Mars, the outside is not a safe refuge. Compartments, pressure doors and independent life-support paths therefore become survival architecture.

The analogy with ships and submarines is useful but incomplete. A Mars settlement also faces abrasive dust, long rescue delays, partial gravity and the need to support ordinary civilian life. Quarters need multiple pressurized evacuation paths. Medical, refuge and command spaces should not all depend on the same electrical bus, air manifold or water line without isolation capability.

Compartmentation has a cost. Every pressure door adds seals, motors, sensors, inspections and failure modes. Yet as population grows, the alternative — one enormous common volume — becomes increasingly fragile. A robust city accepts some daily inefficiency in exchange for the ability to survive rare but severe events.

At one thousand residents, the right mental model is no longer an enlarged space station. It is a network of districts able to operate temporarily without one another. A power failure in one sector should not stop every pump. A decompression should not force evacuation of the whole city. Network topology becomes as important as wall material.

Compartment size is a trade. Very small cells contain leaks and fire well but add doors, seals, controls and logistical friction. Very large cells are comfortable and efficient in normal operation but allow one incident to threaten more people and equipment. The design therefore needs explicit fault-containment zones tied to evacuation routes, refuge capacity and independent environmental control.

A useful rule is to ask what the settlement is allowed to lose. If the answer is ‘one room’, pressure boundaries and utility isolation must make that physically possible. If the answer is ‘one district’, emergency shelter and cross-ties must support the remaining population. Architecture becomes a statement of tolerated failure.

The real city is underground: air, water, power, data and heat

Pictures of settlements show domes, towers and modules. Real cities work because of invisible networks between them. Every district needs breathable air, potable water, wastewater transport, electrical power, data, heat, emergency isolation and repair access. On Mars, a frozen pipe or severed cable can disable life-critical functions, so utility galleries matter as much as façades.

Buried or regolith-covered networks gain thermal and mechanical protection but must remain accessible. Modular architecture can use pressurized human corridors and separate unpressurized service galleries, with frequent isolation points. Pumps and pipe sections must be replaceable without shutting down an entire district.

Heat belongs in the same plan. Vertical farms, computing centres, reactors and industrial shops reject heat at different temperatures. Some of that energy can serve space heating, water processing or freeze protection instead of being rejected directly. The thermal network becomes a recovery system while retaining the ability to reject excess heat when consumers are unavailable.

Growth also needs reserved corridors. A pipe buried today where a pressure tunnel must run tomorrow is a planning failure. A Mars master plan therefore manages subsurface utilities, landing zones, rover routes, water reserves, industrial districts and future neighbourhoods as one territorial system.

The visible room is only the end of a network. Air ducts, water loops, electrical feeders, data fibres, fire detection, heat-rejection loops and waste lines have to cross pressure boundaries without turning every penetration into a leak path or maintenance trap. Utility corridors and accessible service volumes can therefore be more valuable than maximizing habitable floor area.

A city-scale design also needs graceful isolation. One failed pump or contaminated water branch should not require shutting down an entire district. Cross-ties, bypasses and sectional valves add complexity but allow maintenance while people remain inside. The invisible network is what turns a collection of pressure vessels into a settlement.

A city should fail by compartment, not by atmosphere

As a settlement grows, connecting every room to one undivided pressure volume becomes increasingly dangerous. One major breach, fire or contamination event could threaten the entire population. A mature pressure network therefore needs isolatable branches, valves or doors, independent sensing and enough duplicated life-support capability to keep other districts alive while one section is evacuated.

This resembles a ship more than a terrestrial open-plan building. Pressure boundaries create neighborhoods. Utility trunks need isolation points. Emergency routes have to remain usable when one corridor is closed. The layout is therefore driven by fault containment as much as by architectural convenience.

Compartmentation also affects everyday efficiency. Every pressure door adds hardware, inspection and operational friction. Too many boundaries make logistics difficult; too few create catastrophic common-mode risk. The design problem is to choose compartments large enough to live and work efficiently but small enough that losing one is survivable.

Build the city in phases that remain useful if growth stops

A settlement plan should not require the thousand-person city to justify the first twenty-person module. Each phase should deliver independent value: safer landing, better shielding, more workshop capacity, additional pressure volume or a second refuge. Phased construction limits stranded infrastructure and lets real Martian material properties change later designs rather than forcing the first assumptions to govern the whole city.

That principle also limits stranded infrastructure if population growth pauses below the optimistic scenario.

Modularity preserves that option without wasting the earlier investment.

That resilience has lasting value.

From four to one thousand people: a habitat becomes an urban network

Four people may live in a few tightly coupled modules with one refuge and limited redundancy. Twenty justify several compartments, a real workshop, stronger medical capability and separated crop and maintenance volumes.

One hundred require a pressure network that can lose one branch without losing the settlement. Workshops, stores, laboratories, sleeping zones and community areas become functional districts. One thousand require urban infrastructure: multiple power and life-support sources, distributed reserves, many airlocks, roads, landing areas and several shielding levels.

At that point “Mars habitat” is almost too small a phrase. The object is a network of pressurized and unpressurized buildings, utility corridors, refuges, roads and protective works. Resilience comes from losing a part without losing the whole.

At four people, one pressure module can be both home and refuge. At twenty, isolation zones and independent utilities become much more important. At one hundred, a single pressure volume or one central life-support plant creates unacceptable common-cause exposure. At one thousand, the settlement behaves like an urban network with districts, utility trunks, isolation valves, maintenance teams and planned outage capability.

Growth should therefore be modular. Every new district should add useful capacity without making existing residents dependent on an unfinished expansion. Pressure, power, water, data and heat networks need sectional isolation and cross-connections. The city becomes robust when it can lose one module, one utility branch or one workshop and continue operating at reduced capacity.

A Mars habitat is not a house: it is a pressure machine inside a fortification

Science-fiction images favor transparent domes because they make a city instantly visible. Engineering starts somewhere else. Outside is a low-pressure, cold, dusty and irradiated environment. Inside must remain breathable, thermally controlled, cleanable, maintainable and evacuable. The enclosing shell therefore behaves first as a pressure vessel.

That separation changes architecture. The wall holding atmosphere does not have to be the same element that carries radiation shielding, stops dust, supports equipment or gives the settlement its external shape. An early base may use imported pressure modules covered by locally moved regolith. Later, massive local structures may support berms, galleries, workshops and landing infrastructure while certified liners or pressure shells remain separate.

NASA’s human-system standard makes the same functional point from the crew side. NASA-STD-3001 Volume 2, Architecture requires habitable volume and layout that actually support living and operations. It does not give one magical volume-per-person value; required space depends on crew size, mission duration, contingency days and the tasks that must be performed.

This book therefore separates six functions before recombining them: hold pressure, protect life, control dust, carry loads, permit maintenance and exploit local materials. Mixing them into one spectacular wall usually hides the real engineering.

This separation of functions also protects the construction schedule. A locally printed berm can be useful even if it cracks cosmetically; a pressure boundary cannot. A road can tolerate roughness that would be unacceptable in a sealing surface. By assigning requirements to the function actually performed, the settlement avoids over-qualifying simple civil works and under-qualifying life-critical hardware. That is how local construction can grow from bulk earthmoving toward certified inhabited structures without pretending every printed object is already a spacecraft pressure vessel.

The city as a permanent construction site

On Earth a building can be “finished”. On Mars the settlement remains an evolving system: excavating, inspecting, repairing, reinforcing, adding airlocks, moving workshops and opening new volumes. The best architecture is not the most spectacular one but the one that can change without forcing a general reconstruction every time population grows.

Regolith: resource, shielding mass and feedstock

Before it becomes a brick, regolith is valuable local mass

The earliest useful application may be simpler than advanced material conversion: move the material. Excavators can create berms, fill containment systems, cover pressure modules and build protective embankments. Local mass can reduce radiation, thermal excursions, dust transport and ejecta exposure without requiring every kilogram to be processed into a structural product.

If a machine averages 10 tonnes per productive hour after travel and handling, 1,125 tonnes represent 112.5 productive hours. At 50% operational availability, the calendar allocation becomes about 225 hours. Suddenly a bucket tooth, bearing or drive motor is part of habitat risk.

The pressure module also has to tolerate the construction method. A flexible habitat may require an external load-supporting shell so regolith does not create dangerous local loads. A robust architecture often separates the functions: certified pressure boundary inside, massive local shielding outside.

The simplest high-value use of local material may be as mass. Suppose a habitat needs an average protective cover corresponding to 1,000 kilograms per square metre over 1,000 square metres of exposed area. That is one million kilograms, or 1,000 tonnes, of material to excavate, transport and place. No chemical transformation is required, yet the civil-engineering campaign is already substantial.

If one autonomous machine achieves a genuine average of 10 tonnes per operating hour, 1,000 tonnes represents 100 productive machine-hours. At 50 percent operational availability, the calendar allocation becomes roughly 200 hours before contingency. Bucket teeth, bearings, traction, dust seals, power supply and recovery from immobilization now become habitat-enabling functions. The calculation explains why excavation logistics belongs beside pressure-vessel design in a Mars architecture.

Regolith is not free concrete

“Print a house with Martian dirt” compresses geology, handling, chemistry, heat treatment and quality assurance into one sentence. Regolith is granular material whose composition, particle-size distribution and salts vary with location. Earth simulants are essential for development, but no simulant reproduces every property of the material a machine will encounter on Mars.

Several processing routes are plausible. Sintering heats grains until they bond. Some approaches use imported or locally produced binders. Others melt or vitrify material. Additive construction places material layer by layer. Sulfur-based binders, basalt fibers, glass and ceramics have all been explored conceptually or experimentally. Each route has different energy, temperature, defect sensitivity, low-temperature behavior and qualification requirements.

NASA’s MMPACT project develops robotic construction and processing with regolith simulants. NASA discusses landing pads, shelters, habitats, roads, berms and blast protection as candidate applications. The program demonstrates a technology direction; it does not mean that a printer is already qualified to construct a human-rated Mars city.

Quality assurance is the hidden half of construction. A robotic plant has to characterize feedstock, prepare it, measure process variables, detect blocked nozzles or poor layers and verify mechanical properties. “Printing” is only the visible step inside a much longer industrial chain.

Before regolith can become a certified structural product it has to become a characterized feedstock. Grain-size distribution, mineralogy, salts, volatile content and electrostatic dust behavior can vary by site. Crushing, screening, drying, blending and quality control may be needed before sintering or additive construction even begins. Every preprocessing step consumes energy, machinery time and wear parts.

This is why terrestrial simulants are valuable but insufficient. A process that prints a strong coupon from one simulant has demonstrated a manufacturing principle, not a Mars building system. Engineering maturity requires repeatability, larger elements, joints, thermal cycling, pressure-related load cases, repair methods and a way to verify the finished product with the instruments actually available on Mars.

Regolith as mass before it becomes a material

The simplest use of regolith is mass: berms, gallery cover, bags or gabions. That requires less chemistry than structural manufacturing and is therefore a plausible early step. Only later come sintered blocks, composites, ceramics or local concretes, all of which must demonstrate strength under thermal cycling, dust exposure and pressure loads.

What internal pressure imposes on every Martian architecture

Pressure never sleeps

An ordinary building on Earth mainly carries gravity loads, wind, snow and occupancy. A pressurized Mars module continuously pushes outward against a near-vacuum exterior. A thin cylindrical shell provides a useful first model:

That result is intentionally dangerous if mistaken for a design thickness. It excludes safety factors, joints, windows, pressure cycles, fatigue, impacts, penetrations, launch loads, handling, corrosion, damage tolerance and manufacturing defects. Its value is conceptual: internal pressure produces a large, permanent tensile load and geometry directly changes structural demand.

Doubling radius doubles ideal thin-wall thickness when pressure and allowable stress are held fixed. Lower Martian gravity therefore does not make a giant dome structurally free. Gravity reduces some dead-load effects; it does not reduce the pressure difference between a breathable interior and the exterior atmosphere.

Atmosphere choice also couples to EVA. NASA notes that reducing the pressure difference between habitat and suit can help frequent EVA operations, while oxygen-rich lower-pressure atmospheres increase fire concerns. NASA-STD-3001, Natural and Induced Environments.

A thin-wall cylinder makes the scaling visible. In the simplest idealization, hoop stress is proportional to internal pressure and radius and inversely proportional to wall thickness. Double the radius while keeping pressure, material and allowable stress unchanged and the idealized required thickness doubles. The equation is useful because it exposes the direction of the trade; it is not a construction drawing.

Real hardware then adds everything the thin-wall equation ignores: safety factors, openings, joints, pressure cycles, welds or bonded seams, local loads, launch and landing loads, defects, fatigue, inspection access and damage tolerance. A large transparent dome therefore does not become structurally easy merely because Martian gravity is lower. Gravity changes weight loads; it does not remove the pressure difference that is trying to pull the habitat apart.

Building pressure systems, not just shapes

A Martian city is not a collection of volumes. Every gallery must distribute air, water, power and communications, handle condensate and allow rapid isolation of a damaged sector. Technical corridors must be designed before walls are closed. An inaccessible pipe can become a lethal maintenance liability. Arcadia’s design logic therefore favours visitable, oversized service arteries that cost more initially but can support decades of expansion.

Choose among imported modules, buried structures and local construction

Why burial beats giant glass domes

Internal pressure pushes outward on every lightweight habitat, while radiation and micrometeoroid risk encourage heavy shielding. Semi-buried or underground architecture uses the terrain itself as protection, stabilizes temperature and reduces direct exposure to dust storms. The trade-off is a more complex construction campaign: excavation, thermal management, airlocks, access routes and structural inspection.

Excavate, move and prepare thousands of tonnes of ground

The settlement may begin with bulldozers that have no driver

A prudent architecture can pre-position power, communications, excavation equipment and habitat elements before people arrive. Printing an autonomous city in one mission is not the objective. It is to assign machines the most massive and repetitive tasks: survey soil, grade terrain, move regolith, prepare roads, build berms, test material processes and prove that hardware survives the real environment.

NASA’s MMPACT work is exploring planetary autonomous construction and large-scale fabrication with regolith simulants. NASA TechPort now marks MMPACT as a ‘Completed Technology Project’ in its 2026 record. That means this particular development project has concluded; it does not mean autonomous planetary construction is operational or human-rated for Mars. The processes, equipment and architectures still require maturation, integration and qualification before real Martian use.

Autonomy matters because Earth cannot joystick a loader in real time across interplanetary delay. A machine must detect obstacles, slope, slip, overload, thermal limits and wear locally, and reach a safe state without immediate human approval.

Availability often matters more than peak rate. A machine that moves 20 tonnes per hour but is down half the time may produce less useful work than a slower, simpler, repairable design. The construction architecture therefore includes a workshop from day one: lifting gear, diagnostics, wear parts, compatible lubricants and documentation.

Transform regolith: sintering, melting, binders and fibres

Sulfur and basalt fibre as prospective options

Sulfur occurs in Martian materials and can act as a binder without the water required by ordinary Portland cement. Sulfur-concrete concepts have been studied for extraterrestrial construction. Basalt fibre is another route: basalt can be melted and drawn into reinforcement fibres. These processes are not yet industrial systems on Mars, so they belong in the category of prospective engineering rather than demonstrated infrastructure.

Airlocks, dust, compartmentation and pressure boundaries

Martian dust gets everywhere: the habitat must behave like a contamination boundary

Martian dust is a mechanical, optical, electrical and health-control problem. It can abrade seals, coat radiators and solar panels, contaminate suits, enter mechanisms and carry undesirable chemical compounds. The best strategy is not to clean an entire city perfectly; it is to stop dust at defined boundaries.

An airlock can therefore have zones: dirty exterior, suit-maintenance area, intermediate decontamination space and clean interior. Suitport concepts can keep much of a suit outside but add their own mechanical and pressure interfaces. No architecture removes the need for cleaning procedures and monitoring.

At city scale, dust affects circulation planning. Excavation shops, rover garages and cargo interfaces belong near the dirty perimeter, while clinics, food production and biological laboratories require much stricter separation. Air systems must not turn a local workshop contamination into a city-wide distribution mechanism.

The same discipline protects science. Poorly controlled transfer between surface operations and laboratories can contaminate samples and make it harder to distinguish Martian material from terrestrial contamination. Cleanliness is therefore simultaneously a health, maintenance and scientific-integrity function.

Fire and decompression: architecture must provide somewhere to escape when outside is not an option

A Mars habitat combines two constraints rarely found together in terrestrial buildings: an artificial internal atmosphere and an uninhabitable exterior. During a fire, occupants cannot simply open a door and leave. During decompression, pressurized volume itself becomes the resource that must be preserved. District geometry therefore needs refuge spaces, pressure doors, isolatable networks and pressurized retreat paths.

The detailed physics and procedures belong in the dedicated fire, depressurization and refuge book so this construction chapter does not cannibalize it. For the architect, however, the consequence is simple: compartmentation must be drawn before interior decoration. Cable penetrations, ventilation ducts and water lines cannot bypass a pressure boundary so thoroughly that isolation is only theoretical.

Urban growth must preserve the same logic. Adding a module should never remove the only escape route from another. Important districts should have two independent paths toward safe volumes, while critical utilities need isolation and alternate supply routes. The settlement becomes less like a sequence of rooms and more like a graph of pressure volumes that can separate and reconnect.

This principle changes civil works: tunnel diameter, door placement, emergency breathing supplies, muster spaces and backup power. Floor area apparently “lost” to resilience functions is actually insurance against events that could otherwise disable an entire district.

The airlock is a sanitary, mechanical and operational boundary

Martian dust is fine, abrasive and chemically relevant. Carrying it into every living volume increases cleaning demand, seal wear and possible exposure to undesirable compounds. The airlock should therefore function as a transition zone for suit servicing, particle capture, tool storage and leak checks rather than simply a door to the outside.

More mature designs may use suitports, dirty/clean zones, local extraction and easily cleanable surfaces. Internal zoning matters too. Food preparation, microbiology, machine shops, medical rooms and crop chambers should not casually share contamination pathways.

Compartmentation also limits fire and decompression. A city cannot rely on one pressure volume with no isolation. Doors and pressure barriers become analogous to watertight and fire boundaries on ships. See Fire, depressurization and safe haven.

Maintainability must remain physical, not just schematic. NASA-STD-3001 emphasizes access, stowage and the ability to perform tasks. A filter hidden behind three panels can be “replaceable” on a diagram and still be effectively inaccessible during an emergency.

An airlock should be treated as a contamination boundary, not just a door to vacuum-like conditions. Dust carried by suits and tools can abrade seals, foul mechanisms and move chemically reactive material into living spaces. Suitports, localized extraction, dirty-to-clean zoning and dedicated tool storage can reduce the amount of material that enters the habitat.

Internal zoning matters for the same reason. Food production, medical care, laboratories, machine shops and waste processing do not benefit from sharing every airflow path. Compartmentation allows a contamination event to be isolated instead of becoming a settlement-wide cleaning campaign. The airlock is therefore where EVA operations, maintainability, industrial hygiene and life support meet.

Roads, landing pads and external infrastructure

The first major Mars structure may be a road or landing pad, not a palace

A base expecting repeated cargo arrivals must survive its own landings. Engines near the ground accelerate gas, dust and fragments. The detailed terminal descent and plume-surface interaction chapter explains why an untreated surface can become an erosion and ejecta source.

A prepared landing pad can therefore be high priority: protect vehicles, reduce dust thrown toward solar arrays and habitats, and provide a known geometry for operations. Roads serve a similar role. Every kilometre of stabilized route can reduce wheel wear, dust, travel time and rollover risk.

MMPACT explicitly treats pads, roads, berms and protective structures as part of the same construction technology family. This produces a more credible picture of settlement growth: geotechnical survey, excavation, grading, material processing and quality control before impressive architecture.

As traffic increases, a Martian spaceport becomes a system: multiple pads, stand-off distance from living areas, logistics corridors and redundancy so one damaged surface does not isolate the settlement.

Landing pads and roads are productivity infrastructure. A prepared pad can reduce engine-driven erosion and the quantity of high-speed ejecta thrown toward nearby hardware. A stable route can lower wheel damage, energy consumption, travel time and dust generation. These benefits accumulate on every cargo delivery and every maintenance trip, so an apparently unglamorous civil-engineering project can create more system value than an early architectural showpiece.

The sequence also matters. A robotic campaign can survey, grade and prepare infrastructure before people arrive. If that work succeeds, the first crew inherits known routes, safer landing zones and stockpiled shielding material. If it fails, the crew learns about the construction system before depending on it for a pressure boundary.

Roads, landing pads and dust

Outdoor surfaces must carry heavy rovers and limit erosion from descent-vehicle plumes. Landing zones need separation from inhabited areas and robust cargo links. Local soil stabilization or vitrification could reduce dust ejection, but every method consumes energy. Martian construction therefore has to treat external civil infrastructure as seriously as pressurized habitats.

Build robotically before crews arrive

The ideal construction campaign starts before the crew arrives

Every EVA hour consumes crew time and exposes people to operational and environmental risk. Heavy work therefore belongs as much as possible to autonomous robotics: survey, marking, excavation, regolith transport, grading, pad preparation and inspection. Mars radio delay makes continuous fine teleoperation from Earth impractical; machines need local obstacle handling and safe-stop logic.

Autonomy converts construction into a software and maintenance problem. An excavator is useful only if actuators can be diagnosed, wear parts replaced, sensors cleaned and failures recovered. Construction therefore connects directly to surface robotics, maintenance and spares and metrology.

NASA TechPort’s 2026 record marks MMPACT as a ‘Completed Technology Project’. Completion of that project is not equivalent to operational readiness for Mars. The boundary remains important: success with simulants is not equivalent to human-rated construction on Mars. Real feedstock, machine endurance and an off-Earth qualification chain still have to be demonstrated.

Autonomy is not optional when communications are delayed and construction runs for hours without a human standing beside the machine. Excavators and haulers need localization, hazard detection, task planning, safe stops and recovery logic. More importantly, the site needs a philosophy for what happens when autonomy is uncertain: stop and wait, retreat to a safe point, or continue at reduced capability.

Construction also requires verification. A robot can deposit material accurately and still create a weak structure if feedstock, temperature or bonding quality drifts. Metrology, sample testing, dimensional inspection and process records are therefore part of autonomous construction. The settlement needs to know not only that a wall was built, but under what conditions and with which quality evidence.

Inspect, test, repair and maintain for decades

A pressurized wall ages every sol: fatigue, seals and inspection

Internal pressure applies a continuous load to the shell. Cycles are superimposed on it: initial pressurization, airlock operation, thermal changes, maintenance, machinery vibration and abnormal events. Even when average stress is far below static failure, welds, attachments, penetrations and windows can concentrate stress and become crack-initiation sites.

Design therefore needs damage tolerance: assume defects will exist, detect them before they become critical, and retain enough margin to repair. On Mars that means accurate pressure sensing, acoustic or ultrasonic inspection, robotic visual surveys, witness coupons, repair records and spare parts compatible with more than one generation of habitat.

Burial under regolith complicates inspection. A protective layer cannot simply become a tomb around the structure. Inspection galleries, removable overburden zones, distributed sensors and machines capable of excavating and replacing fill become part of the architecture.

A city expected to last a century cannot depend on one terrestrial supplier still existing. Interfaces therefore need standardization: pipe sizes, flanges, seals, connectors, data protocols, fasteners and qualified material families. Standardization is less dramatic than 3-D printing, but it is one of the foundations of long-term maintainability.

Maintenance data should accumulate with the habitat’s age. Pressure cycles, seal replacements, leak-rate trends, strain measurements, dust exposure and thermal excursions form a health record for the structure. Inspection intervals can then be adjusted to observed degradation rather than relying forever on assumptions made on Earth.

Local repair capability is equally important. A settlement that can manufacture a bracket but cannot qualify a pressure seal is not autonomous where it matters. Repair procedures need compatible materials, surface preparation, cure conditions, leak checks and acceptance criteria. Structural maintainability is therefore a metrology and configuration-management problem as much as a fabrication problem.

The habitat is never finished: inspection and repair become permanent occupations

Pressure seals age. Dust abrades mechanisms. Thermal cycles move joints. Pumps vibrate. Filters clog. Sensors drift. A Mars building therefore cannot be designed like a monument expected to stand quietly for decades. It is closer to an aircraft, submarine and industrial plant that happens to contain homes.

Design for maintainability means access panels where people can actually reach them, standardized fasteners, replaceable seals, inspection ports, leak localization and configuration records that tell the crew exactly which component is installed. Local manufacturing can reduce spare-part mass only when metrology and qualification can prove that the replacement is fit for service.

At settlement scale, structural health monitoring becomes valuable: pressure trends, acoustic leak detection, strain sensors and periodic imaging can identify deterioration before it becomes an emergency. The city’s most important construction crew may eventually be the team that never builds anything new because it keeps everything old alive.

What is demonstrated, what remains experimental and what still requires qualification

Building on Mars: separate six functions that are too often confused

The word “habitat” hides several engineering problems. A wall may need to retain breathing pressure, provide radiation mass, carry loads, survive local impact, reduce heat flow, keep dust out and remain repairable. No construction technology should be judged without stating which function it performs.

1. Pressure structure contains atmosphere.
2. Radiation shielding supplies protective mass.
3. Thermal protection reduces heat flow and gradients.
4. Mechanical protection carries loads, impacts and ground movement.
5. Exterior infrastructure includes roads, pads, berms, platforms and supports.
6. Habitable finish provides cleanable surfaces, partitions and technical interfaces.

This distinction matters because raw regolith may be excellent shielding while being a poor pressure shell. A 3D-printed outer layer may protect an imported pressure vessel without becoming the airtight structure itself.

Geotechnical knowledge comes before architecture

Bearing capacity, settlement, slope, boulders, buried ice, particle size, excavation stability, rover traffic and plume erosion all influence design. A site attractive for water resources may impose different foundation constraints from a dry rocky site. Early construction therefore needs probing, sampling, compaction tests, radar mapping where appropriate and long-term settlement monitoring.

Regolith is valuable first because it is already there

Before sophisticated chemistry, the strongest advantage is logistical: local mass does not have to be launched from Earth. Regolith can serve as fill, berms, cover, ballast, confined shielding and terrain shaping. This naturally favors partly buried habitats, utility trenches and protected galleries.

When regolith becomes material: several routes, no universal recipe

1 — Compaction and confinement

Material may be compacted or retained in cells, bags or other containment. It requires little chemical processing but substantial handling and stability control.

2 — Sintering and vitrification

Heating can bond grains or produce vitrified surfaces. Potential products include tiles and traffic surfaces. Energy, cracking and repair become key design issues.

3 — Imported or local binders

Polymers, sulfur-bearing systems or geopolymer-like binders may create composites where chemistry permits. Binder availability, durability, toxicity, thermal behavior and recyclability must be included in the architecture.

4 — Large-scale additive construction

3D printing is a forming method, not a material. The system still needs feed preparation, dosing, deposition or consolidation, geometry control, defect detection and repair. NASA MMPACT and CHAPEA demonstrate why automated construction is attractive, but CHAPEA is a terrestrial analog built with a dedicated material; it is not proof that raw Martian regolith can simply be printed into a pressure habitat.

Roads and landing zones may be more urgent than the next building

Unprepared surfaces generate dust and ejecta. Landing plumes can erode soil and threaten nearby hardware. A growing base therefore needs separated landing areas, berms, stabilized surfaces and logistics roads early in its development.

Roads reduce suspension wear, vibration, energy use, travel time, immobilization risk and dust generation. Infrastructure mass placed once can save thousands of maintenance cycles later.

A pressurized building is fundamentally a leak-control system

Joints, penetrations, doors, airlocks, cable passes and pipe connections are as important as walls. Damage to an outer shielding layer may be tolerable; damage to a pressure boundary can become an emergency. Layered architecture — inspectable pressure shell, insulation, then external mechanical and radiation protection — often improves repairability.

Construction should minimize human EVA hours

Every EVA consumes crew time and spacesuit life while increasing exposure. Heavy work should therefore move toward robotic excavation, grading, hauling, inspection, printing or module placement. A practical rule is: people define goals, supervise and handle exceptions; machines move tonnes.

A city also requires buried networks

Water lines, power cables, data fibres, gas pipes, drains, tanks and isolation valves turn buildings into a city. Trenches or technical galleries can protect these services and keep them accessible. Networks must be sectionalized so a single failure does not drain an entire water stock or remove all power.

A sensible growth sequence

1. map terrain, resources and hazards;
2. prepare landing and logistics corridors;
3. install power, communications and maintenance capability;
4. deploy first pressure volumes;
5. add local regolith shielding rapidly;
6. establish excavation and material storage;
7. produce non-critical local materials;
8. build roads, berms, platforms and warehouses;
9. industrialize utilities and maintenance;
10. only then increase locally made safety-critical structures after qualification.

Acceptance testing: construction is not finished when printing stops

Pressure habitats may require geometric inspection, leak testing, staged pressurization, deformation monitoring, thermal inspection and emergency drills. Roads and pads need settlement, abrasion, dust generation and repeated-load testing. Material provenance, recipe, process parameters, defects and repairs should remain in the digital record.

What NASA demonstrators actually tell us

NASA supports autonomous construction and ISRU to reduce imported mass. MMPACT explores large-scale robotic construction, while CHAPEA shows that a 3D-printed Mars analog habitat can support long simulations on Earth. The correct conclusion is not “NASA has already printed a Mars city from Martian soil.” It is that additive construction, automation and ISRU are credible development directions, while each material-process-environment combination still requires qualification.

Martian construction checklist

• What exact function must the element perform?
• Which mass is local and which remains imported?
• What energy and maintenance does the process require?
• How are cracks and defects detected?
• Can it be repaired without dismantling the entire system?
• What happens if the main construction robot fails?
• Does construction dust threaten arrays, radiators or airlocks?
• Can the material be reused or recycled after demolition?

The six functions also mature at different rates. Excavation and berm building may become useful early because they mainly require moving mass. Certified pressure boundaries demand tighter process control, leak testing and long-term material data. Landing-pad construction must survive plume loading and repeated operations. Roads optimize mobility rather than habitation. The mistake is to assume that success in one function proves readiness in all the others.

A phased settlement can exploit that difference. Imported pressure modules provide an early known life-support boundary; local machines create shielding and civil works around them; local structural products first appear in non-pressure-critical roles; and only after manufacturing, inspection and repair mature does the colony consider local pressure structures. This sequence converts ISRU from a slogan into an increasing ladder of responsibilities.