Space Habitats

A space habitat is a pressurised volume that people live in off Earth: a hull, an atmosphere, a thermal system, a life-support loop, a way in and out, and an operating organisation behind all of it. The framing under test is that habitat engineering is a solved problem waiting on launch cost. This is the most directly falsifiable framing in the settlement cluster, because both of its terms are measured. Launch cost has a published series. Habitat engineering has a twenty-five-year operational record with a documented, unrepaired, seven-year-old air leak that the operator's own auditor rates at the top of its risk scale.

The brief should be generous before it is hard, because the record genuinely is strong and the popular treatment under-sells the achievement while over-claiming the conclusion. Continuous human habitation of space is twenty-five years old, across two independent programmes. Expandable habitats have been flown, instrumented and certified. A full-scale inflatable has been burst-tested past its requirement. Pressure vessels, atmosphere management, thermal control, extravehicular activity, docking and long-duration crew operations are solved in the engineering sense. The framing is not foolish. It is wrong for reasons that are countable.

The cut against O'Neill Cylinders is regime, not size: this brief owns habitats that exist or are funded — the ISS, Tiangong, BEAM, Sierra Space's LIFE, the Commercial LEO Destinations programme and its six companies — and the practical engineering of leaks, debris, life support and crew operations. FR-II-06 owns habitats designed to be self-contained at settlement scale and the comparative quantities that decide them. Artificial Gravity owns all rotation physiology and this brief carries one sentence. And Historical Space Colonization Concepts owns the 1970s station and settlement concepts; Skylab, Salyut and Mir appear here only as one clause of operational heritage.

Established What is solved is solved, and the numbers deserve to be stated properly. The ISS has been continuously crewed since November 2000. Tiangong has been permanently crewed since November 2022, when Shenzhou 15 arrived: three modules of about 22 tonnes each in a T configuration, 340 m³ pressurised and 122 m³ habitable — about a third of the ISS — with standard rotations of three crew for about 180 days, six during handover, a 10-year design life extensible to 15, and an announced expansion to six modules and 180 tonnes by 2027. Two independent national programmes have now demonstrated long-duration habitation, which is a stronger statement than one programme's twenty-five years alone.

Established Expandable habitats are not an open question; they are certified. NASA's own two-year report on the Bigelow Expandable Activity Module, attached to the ISS since 2016, closes it: about 1,400 kg packed, expanding from 3.6 m³ to 16 m³ — a 4.4-fold expansion; full pressurisation on 28 May 2016, with deployment instrumented by acoustic sensors that recorded “thousands of impulses... from the Rip-Stitch Strap stitches popping” at peaks of 0.5 g on the first inflation attempt and 0.3 g on the final one; stress requirements met at a fully loaded 109-CTBE configuration; and the conclusion, in the report's own words, that “BEAM has advanced human rated expandable modules to TRL 9”. Established One confirmed micrometeoroid impact on 28 February 2017 registered 1–3 g at the sensors with an estimated 260 g at the restraint layer, and analysis indicated no penetration of all structural layers; BEAM “meets MMOD penetration requirements through 12/31/2028”. Established And the report reports against itself on radiation: galactic cosmic ray dose rate in BEAM is similar to other ISS modules, but trapped-radiation exposure during South Atlantic Anomaly passages is higher “in BEAM than in other ISS modules due to thinner shell and lack of equipment racks”, with 3D-printed shielding trials at 1.1, 3.3 and 10 mm returning inconclusive results. Interest running against the finding, from the programme reporting its own module's worse performance.

Established The largest inflatable has been burst-tested and the margin is real. Sierra Space's full-scale LIFE module reached 77 psi ultimate burst pressure against NASA's recommended 60.8 psi — itself four times maximum operating pressure — a margin of about 27%. The module is 300 m³ in the five-metre-fairing version with a proposed 1,400 m³ seven-metre variant, using Vectran restraint straps developed with ILC Dover; an earlier subscale test with a metallic window simulator achieved a 33% margin. Remaining work at that point was atmospheric barrier and micrometeoroid layer testing, with flight hardware development to begin “in 24 to 36 months”. A vendor test reported by independent trade press, and the honest reading is that the structure is genuinely demonstrated while the integrated module is not.

Established Now the framing's own test, which it fails on its own terms. Cost per kilogram to low Earth orbit, on a stated like-for-like basis of list price divided by maximum stated payload in 2025 dollars: the Space Shuttle at about $54,000/kg; expendable Falcon 9 near $3,000; Falcon 9 with a reused booster at about $2,700; rideshare slots around $6,000; and Starship's quoted $100–200/kg, which the compiler explicitly labels “a design target, not an achieved price”. Established That is a factor of about twenty, achieved, between the Shuttle era and today. Over the same period: no new large habitat has been launched by anyone except China; the ISS has not been expanded; the first commercial station has slipped from 2025 to no earlier than the first quarter of 2027; and NASA's own budget projection is, in its own March 2026 clarification, “only sufficient to support one commercial space station”. Frontier State the inference carefully, because it is the brief's central move. If habitat engineering were solved and only launch cost were binding, a twentyfold cost reduction should have produced habitats. What it produced was one 14-tonne single module still awaiting launch, two multi-module plans slipping to 2028 and 2029, and a government programme that can afford one of them. Launch cost was not the binding constraint. Demand and operating cost were.

Established The operational counter-example is a leak that will not close. The ISS has been losing air continuously since 2019 through cracks in the Zvezda service module's PrK transfer tunnel. The record: 2.4 lb/day in February 2024, rising to 3.7 lb/day in April 2024 — a 50% increase and the peak — then reduced by roughly a third from that peak. As of August 2024 NASA rated the tunnel a 5 on its 1-to-5 risk scale — the top of the scale — making it the highest-rated safety risk on the station, per the OIG report released 26 September 2024, with both agencies having “narrowed their focus to internal and external welds” without identifying a root cause. Established And it got worse. NASA's own station blog of 5 June 2026 records the leak rate increasing to two pounds per day during cargo operations; Roscosmos planned a structural repair involving cutting a bracket for access; NASA objected on safety grounds; the repair was paused in favour of further measurement; and during the assessment NASA placed five crew into “a heightened safety posture, known as a safe haven, inside the SpaceX Dragon spacecraft”. Frontier Derived, from two published figures: two pounds per day is about 0.9 kg/day, or roughly 330 kg of atmosphere a year that must be resupplied from Earth; at reused-Falcon-9 pricing of about $2,700/kg that is on the order of $0.9 million a year in makeup gas alone, before tankage and before the flight. The arithmetic is ours and inherits the weaker flag. For a habitat intended to be independent, makeup mass is a permanent tax, and this is the only real-world number for it. Established The sharpest finding available against the framing follows directly: the most mature habitat ever built has an unrepaired structural air leak of unknown root cause, entering its seventh year, rated at the top of the operator's risk scale, which recently forced the crew into a lifeboat. Pressure-vessel engineering is solved; pressure-vessel ageing and repair in orbit is not, and there is no other data set on it. Established The second data point comes from the other station: Tiangong's Shenzhou 20 suffered cracks in a spacecraft porthole from an external debris impact, delaying the crew's return, who came home on Shenzhou 21 instead. Two stations, two independent pressure-boundary incidents, in the only two long-duration habitats humanity operates.

Established Closure is 98% of the water and none of the food. The genuinely good number: the ISS achieved 98% water recovery in June 2023 through the Water Processor Assembly, the Urine Processor Assembly with its Distillation Assembly, the Brine Processor Assembly and advanced dehumidifiers capturing crew breath and sweat — against NASA's own statement that missions beyond low Earth orbit need “close to 98%”. For water, the requirement is met. Established The ladder that is not, cited from Mars Colonization rather than re-derived here: food closure on the ISS is zero and all food is resupplied; the best food closure ever achieved with humans is 60%, in Lunar Palace 1, four crew over a year, with waste recycling not closed; the longest human closure of any kind is six months, in BIOS-3, one to three people at 95.4%; “no nation has demonstrated a completely closed BLiSS system” including regenerative operation, feedstock and nutrient recycling and human waste processing; and full readiness for a fully bioregenerative habitat is estimated to need four to eight years of operational experience that nobody is currently accumulating. Frontier What that does to the framing is precise. A station resupplied every few months does not need closure, which is why the ISS's is partial and why that does not matter there. But the framing says habitat engineering is solved, and the thing that makes a habitat a place to live rather than a place to visit — food, and closing the waste loop — is neither solved nor close. The solved part is the part that assumes a supply line.

Established The environment those habitats sit in is getting measurably worse. ESA's Annual Space Environment Report, Issue 10.0 of 1 May 2026, counts about 54,000 objects larger than 10 cm, about 1.2 million between 1 and 10 cm, and about 130 million between 1 mm and 1 cm, with roughly 16,000 tonnes of catalogued and asserted objects in geocentric orbit, and 9.8 non-deliberate fragmentations per year on a two-decade average — 2024 alone adding more than 3,000 catalogued fragments, mostly from propulsion-related breakups. The assessment: “in the business as usual extrapolation scenario, the level of risk associated with the environment is predicted to be 4 times higher than this acceptable first target threshold towards long-term sustainability”, and current compliance with debris mitigation guidelines is “insufficient”. Established The 1–10 cm population is the one that matters for habitats: too small to track reliably, too large to stop with a Whipple bumper. 1.2 million objects. BEAM's certification is worth quoting in full for exactly this reason — “meets MMOD penetration requirements through 12/31/2028” — because it makes visible that micrometeoroid survivability is a time-limited claim about a changing environment rather than a property of the hardware.

Established And the most current and most decisive source in the subject is an audit. GAO-26-107805, 17 June 2026: the ISS deorbit is targeted for the end of 2030 or early 2031, with a deorbit vehicle contract to SpaceX worth up to $843 million; current ISS annual operating cost is about $3 billion; Phase 1 of the Commercial LEO Destinations programme was funded at nearly $530 million across one contract and three funded Space Act Agreements; Phases 2 and 3 are projected at $1–1.5 billion across FY2026–2031, which NASA clarified in March 2026 is “only sufficient to support one commercial space station”; six companies are in development — Axiom Space, Blue Origin, Starlab Space, Sierra Space, SpaceX and Vast Space; and GAO found that NASA “has not yet assessed the likelihood or duration of a gap” in US low-Earth-orbit presence despite identifying it as a critical concern. On demand, GAO records uncertainty over “whether commercial companies will be able to secure other customers to complete their business case”, citing industry concerns about market viability. GAO makes no statement that commercial stations will be ready by 2030. Established The schedule record behind that: Vast's Haven-1 — a single module, four crew, missions to two weeks, relying on a docked Dragon for life support and power — has slipped from 2026 to no earlier than the first quarter of 2027; Axiom restructured its assembly plan in December 2024, abandoning the 2020 concept of assembling four modules at the ISS before detaching, so that the Payload Power Thermal Module now launches first no earlier than 2027 and docks to the ISS, Habitat One follows no earlier than 2028 and the PPTM undocks to join it, with the Airlock, a second habitat and a Research and Manufacturing Facility following into the early 2030s; and Axiom and Starlab are described as planning two-module stations in 2028 and 2029. Frontier Three years from ISS deorbit, the successor programme has six competitors, funding for one, no assessed gap risk, and a first article that is a single 14-tonne module dependent on a visiting spacecraft for its life support. That is not a capability problem. It is a demand and funding problem, and it is the direct refutation of “waiting on launch cost”.

Established Everything above was solved inside Earth's magnetosphere, and the moment a habitat leaves it the numbers change category. Deep-space transit dose is 1.84 ± 0.3 mSv/day, measured. Shielding to a 20 mSv/yr general-population limit in deep space requires 6–7 tonnes per square metre of polyethylene or 10–11 of lunar regolith. Galactic cosmic radiation is halved at 8 g/cm² and quartered at 50 g/cm², giving 13 and 78 tonnes per person respectively at 100 m³ of living space. And aluminium dose-equivalent reaches a minimum at 20–30 g/cm², above which secondaries increase the dose. Frontier Derived, to make the ratio concrete: the ISS masses about 420 tonnes with a habitable volume of order 400 m³ for a crew of seven — on the order of 60 tonnes per crew member, none of it dedicated galactic-cosmic-ray shielding. A deep-space equivalent at 50 g/cm² would add 78 tonnes per person on top of that. Going beyond the magnetosphere roughly doubles the per-person mass of an already-$150-billion object, and that mass buys only a fourfold dose reduction. Frontier So the precise claim is this: habitat engineering is solved for low Earth orbit, with resupply, inside the magnetosphere, for crews of single digits, on missions of months. Every one of those five qualifiers is doing work, and the framing drops all five.

Established One sentence on rotation, and then the handoff. No space habitat has ever rotated, no funded programme includes a rotating module before the 2030s, and every operational habitat to date — ISS, Mir, Tiangong, BEAM — has been designed for microgravity. Frontier That is a deliberate commercial choice rather than a technical failure, because microgravity is the product a low-Earth-orbit station sells. Whether long-duration microgravity is survivable, what rotation rate is tolerable and what the physiological dose–response looks like all belong to Artificial Gravity; what radius a settlement would need belongs to O'Neill Cylinders. This brief re-derives neither.

Established The successor programme has a certification problem that is separate from its funding problem, and it is the one with a directly transferable base rate. NASA's Phase 1 arrangements with the commercial destination companies are Space Act Agreements: milestone-funded development instruments rather than contracts for a delivered service. Established Before the agency can put its own astronauts aboard a station it does not own, it has to certify that station against human-rating requirements it writes and the operator meets — the same two-step used for Commercial Crew, where a development phase was followed by a certification phase and then a services contract. Established The precedent is not encouraging about duration: Commercial Crew certification ran years beyond its original dates for both providers, and the causes were requirement interpretation, verification evidence and flight-test anomalies rather than anything a larger development budget would have bought. Frontier A station is a harder certification article than a capsule in one specific respect: it is occupied continuously, so its safety case is a statement about years of operation rather than about a mission, and there is no flight-test analogue for that.

Frontier Schedule overlap, not schedule, is the requirement that actually binds. The programme's logic requires a commercial destination to be certified and operating before the ISS deorbits, because what the agency is buying is continuity of crewed presence rather than the existence of a station. Established “Operating” in that sentence means certified for NASA crew, which is later — sometimes much later — than the date a company first flies hardware. Frontier Every commercial date recorded in this brief is a first-element date, and none of them is a certification date. Speculative The interval between those two quantities is the entire gap question, and no source in this pack states it for any provider. That is a specific and answerable request an auditor could put to the agency: publish, per provider, the interval assumed between first element on orbit and certification for NASA crew.

Frontier Anchor tenancy is the model, and the model has a known failure mode that this programme inherits. Commercial cargo and commercial crew worked because a government customer with predictable multi-year demand let private vehicles amortise development, and because the destination those vehicles flew to was a government asset. Established Commercial destinations invert the second half: the private party now owns the thing the agency needs, so a provider's commercial failure removes a capability rather than a supplier. Frontier An anchor tenant that is the only tenant is the whole rent roll, and a programme funded at a level its own buyer describes as sufficient for one station is not creating a market — it is selecting a monopolist by tournament.

Established A station's value is conditional on vehicles its operator does not build. Crew access to any American commercial destination currently runs through a single vehicle in routine service, with a second that has not yet entered rotation; cargo runs through a slightly wider set. Frontier A destination company's business case therefore depends on the availability, price and manifest of a transport provider that is in some cases a direct competitor for the same agency funding. Established The dependence is sometimes literal rather than commercial: the first single-module commercial station is designed to draw power and life-support services from a docked crew vehicle, which makes the vehicle part of the habitat rather than a visitor to it. Frontier The correct unit of analysis is the transport-and-destination system, and every schedule published in this subject is a module schedule.

Frontier The demand side is the weakest evidence in the whole subject, and it is weak in a specific and measurable way. Microgravity research has real results and a real user community: protein and pharmaceutical crystallisation, tissue and organoid culture, fibre drawing in fluoride glasses, and materials processes that convection spoils on the ground. Established The demonstrations that have actually returned saleable product did so at the scale of kilograms, in free-flying capsules, without crew. Speculative That is the uncomfortable fact underneath the crewed-station business case: most of the demonstrated in-space manufacturing demand does not need people, and an uncrewed free-flyer is far cheaper than a habitat. Frontier What a crewed station uniquely sells is human research — mostly a government purchase — together with sovereign and private astronaut missions, which are a small number of seats at high prices. Handwave No source in this pack offers a bottom-up estimate of non-government demand for crew-tended pressurised volume in tonnes, seats or dollars per year, and until one exists every statement about market viability, optimistic or pessimistic, is an assertion wearing a spreadsheet.

Frontier Position one is the framing: habitat engineering is solved and only launch cost binds. Held by much commercial-space commentary. For it: BEAM at TRL 9, LIFE burst at 77 psi against a 60.8 psi requirement, and twenty-five years of continuous habitation across two independent programmes. Against it: launch cost fell about twentyfold with no new habitat, a seven-year unrepaired leak at the top of NASA's risk scale, zero food closure, and GAO reporting funding for one station and unassessed gap risk. frontier on the engineering half; speculative on the causal claim about launch cost.

Frontier Position two is the most under-covered live position in the subject: the binding constraint is demand, not capability. GAO holds it in substance, recording uncertainty over “whether commercial companies will be able to secure other customers to complete their business case”. The evidence is arithmetical — six competitors, $1–1.5 billion described by the customer as sufficient for one, and an incumbent station costing about $3 billion a year to operate. If this position is right, then every technical roadmap in the subject is answering the wrong question, and the thing to watch is not a burst test but a signed contract with a non-government buyer.

Established Position three: expandable habitats are the answer to volume per launched kilogram. The structure is established: 4.4-fold volume expansion at 1,400 kg, TRL 9 by the operator's own assessment, a 77 psi full-scale burst, and a 1,400 m³ variant proposed. Frontier The trade is frontier: BEAM's own report finds higher trapped-radiation dose through the South Atlantic Anomaly “due to thinner shell and lack of equipment racks”, and the micrometeoroid certification is explicitly dated — through 2028 — against an environment that is worsening. Light structure buys volume and costs shielding, and the report that established the capability is the same one that says so.

Frontier Position four: a commercial station will be flying before the ISS deorbits. Held by NASA's strategy and the six companies. For it: Haven-1 no earlier than Q1 2027 as a single module, Axiom's PPTM no earlier than 2027 and Hab-1 no earlier than 2028, Starlab in 2029. Against it: GAO says NASA has not assessed gap likelihood or duration and makes no readiness statement, and Axiom restructured its entire assembly sequence in December 2024. Frontier Position five: there will be a gap in US low-Earth-orbit presence after 2030. Implied by GAO's recommendation that NASA analyse it. The evidence is the funding, the schedule and the unassessed risk — and the honest observation that the auditor asked for the analysis rather than making the prediction, which is why this sits at frontier rather than higher.

Speculative Position six: pressure-boundary ageing is a solved problem. Implied by the framing and contradicted by the only long-run data anyone has. Seven years, two agencies, focus narrowed to welds, no root cause, risk rated 5 of 5, rate rising to two pounds per day in June 2026, a structural repair paused as too dangerous, and a crew placed in a Dragon safe haven — plus a debris-cracked porthole on the other station. This is the operational counter-example the page should lead with, and it is one data set of size two. Speculative Position seven: closure will be solved by the time it is needed. Implied by every station roadmap. Against it: 98% water and 0% food, no closed waste loop anywhere, and four to eight years of operational experience estimated for full readiness with no facility accumulating it.

Frontier Position eight: debris will not constrain habitats. Implicit in commercial low-Earth-orbit planning, and trending negative: 1.2 million objects between 1 and 10 cm that are neither trackable nor shieldable, 9.8 fragmentations a year, business-as-usual risk projected at four times the sustainability threshold, and a real porthole cracked on a real station. Frontier Position nine: Starship changes the calculation. For it: twelve flights, seven successes, Block 3 flying, and a fairing large enough for single-launch station modules, which is what Starlab's design assumes. Against it: $100–200/kg is a design target rather than a price, and ship-to-ship propellant transfer has never been attempted.

Speculative Position ten: microgravity stations are a stepping stone to rotating habitats. Held by the settlement narrative. Against it: nothing in any funded programme rotates before the 2030s, and microgravity research is the low-Earth-orbit business case — the two designs point away from each other rather than toward each other. Route the physiology to Artificial Gravity. Frontier Position eleven: low-Earth-orbit habitat experience transfers to deep space. Implied whenever the ISS is called a first step. Against it: everything solved was solved inside the magnetosphere with resupply, and deep space adds 1.84 mSv/day and 13 to 78 tonnes per person of shielding. The transferable part is operations and crew procedure; the untransferable part is the mass.

Frontier And a seam worth naming as a gap rather than resolving. The one orbital regime in which O'Neill Cylinders reports a settlement needing essentially no radiation shielding — equatorial low Earth orbit below about 500 km — is inside the region where debris flux is highest. The two hazards trade against each other and no source in this pack analyses them jointly. That is a real hole in the literature and both briefs say so in the same words.

Frontier And a twelfth position, arriving after the audits rather than in them: the acquisition strategy itself is unstable. The successor procurement was designed as a competitive certification-and-services contract following the development agreements, and the approach has since been revisited in the direction of further milestone agreements, reduced initial capability requirements and a demonstration article rather than a full station. Speculative If that is what has happened, it is the rational response to a budget that funds one station and a schedule that will not hold — and it also converts a firm requirement into an option, which is precisely the move that makes a gap more likely rather than less. Speculative The evidence that would confirm or refute it is a published solicitation with a stated minimum capability: crew size, mission duration and the certification milestone that triggers payment. This brief does not have that document, and nobody should describe the programme's current shape with confidence without it.

Frontier The first bottleneck is a customer. GAO records $1–1.5 billion for FY2026–31 that NASA itself says is sufficient for one station, six companies competing for it, an incumbent costing about $3 billion a year to operate, and explicit uncertainty about whether anyone other than NASA will buy. Every technical bottleneck below is downstream of that one, because a capability nobody funds does not mature regardless of its readiness level.

Established The second is pressure-boundary ageing, and it is the bottleneck with a live example. A leak since 2019, root cause unidentified after two agencies narrowed their focus to welds, rated 5 of 5, worsened to two pounds a day in June 2026, with the structural repair paused because cutting a bracket for access was judged too dangerous. There is no established method for locating, characterising or repairing a structural leak in an orbiting pressure vessel, and the sample size for the whole problem is two stations.

Established The third is closure, and specifically food and waste. Water is done at 98%. Food closure on any flown system is zero. The best ever achieved with humans is 60% over a year with waste uncontained, and no nation has demonstrated a completely closed system. The estimated path is four to eight years of operational experience in a facility that does not exist. Until that changes, every habitat is an endpoint of a supply chain rather than a place.

Frontier The fourth is a debris environment that is worsening faster than mitigation compliance is improving. 1.2 million objects between 1 and 10 cm, untrackable and unshieldable; 9.8 unintentional fragmentations a year; over 3,000 new catalogued fragments in 2024 alone; and an agency assessment that compliance is “insufficient” with business-as-usual risk at four times the sustainability threshold. The certification language is the tell: BEAM meets requirements “through 12/31/2028”.

Frontier The fifth is mass, the moment a habitat leaves the magnetosphere. 13 to 78 tonnes per person of shielding at settlement dose targets, against an ISS that carries about 60 tonnes per crew member of everything else and no dedicated shielding at all. Established And the sixth is operating cost, which is the one the framing never mentions. About $3 billion a year to run the ISS, against a successor programme funded at $1–1.5 billion over six years. Launch price is a small term in that budget, which is precisely why a twentyfold reduction in it produced no habitats.

Frontier A seventh bottleneck sits underneath the first and is not a customer: it is a certification route. Nothing in the programme's funding addresses the problem that the agency must human-rate a privately owned, continuously occupied facility against requirements not yet in final form, using verification evidence that a commercial operator has a commercial interest in limiting. Established The Commercial Crew precedent says that step takes years and that its duration is not predictable from the development schedule. Speculative Add it to a first-element launch date and the overlap with an end-2030 deorbit becomes the tightest number in the subject — and the one nobody publishes.

Frontier The adjudication carries two edges and both are narrow on purpose. Artificial Gravity because the single largest open question about long-duration habitation is whether microgravity is survivable indefinitely and what, if anything, replaces it — and because no habitat has ever rotated. FR-I-02 owns the physiology, the tolerable rotation rate, the countermeasure record and the programme history; this brief owns only the fact that microgravity is what low-Earth-orbit stations sell.

Frontier Space-Based Manufacturing because the assembly record is the ceiling on habitat scale. The ISS is about 420 tonnes across dozens of flights over more than a decade at roughly $150 billion — about $0.36 billion per tonne, the only empirical figure for assembled orbital mass. Whether that falls, and whether structures can be fabricated on orbit rather than launched whole, is FR-I-24's question. The commercial designs currently answer it by avoiding it: a single large module launched whole, sized to a fairing.

Established What this brief does not depend on is a technology at all. The pressure vessel is solved, the expandable is at TRL 9, the burst margin is 27% above a requirement that is itself four times operating pressure, and atmosphere, thermal, extravehicular and docking operations have twenty-five years of record. The dependency that would change the subject is a paying customer who is not a government, and no brief on this map produces one.

Frontier And two handoffs. O'Neill Cylinders supplies the shielding areal densities that show how orbit-specific “solved” is, and this brief supplies FR-II-06 with the operational record of what a real pressure vessel costs to keep. Mars Colonization owns the closure ladder and the 900-day requirement; this brief cites it rather than restating the analysis.

Established The most valuable experiment available is one already running badly: find the root cause of the ISS leak. Seven years, two agencies, a focus narrowed to internal and external welds, and no identification. It is the only long-duration pressure-boundary ageing dataset that exists, and whatever is learned from it is the entire empirical basis for designing a hull meant to last decades. The repair attempt was paused rather than abandoned, which means the diagnostic question is still open and still answerable.

Frontier Second: run a closed life-support facility with humans for years rather than months. The published estimate of what full readiness needs is four to eight years of operational experience; the record is six months; and the specific gap is waste processing, which no nation has closed. This is a facility programme with a schedule rather than a research question, and the same experiment appears in Mars Colonization and O'Neill Cylinders because all three subjects wait on it.

Frontier Third: characterise and demonstrate protection against the 1–10 cm debris population. It is the population that matters and the one nothing addresses — 1.2 million objects, too small to track and too large to stop with a bumper. BEAM's dated certification is the honest current state: a modelled claim with an expiry. What is missing is an experimental basis for extending it against an environment ESA projects at four times the sustainable risk level.

Frontier Fourth: fly the integrated inflatable, not just the burst article. LIFE has been burst-tested at 77 psi; the remaining work named at the time was atmospheric barrier and micrometeoroid layer testing, with flight hardware “in 24 to 36 months”. The gap between a certified restraint layer and a certified habitable module is precisely the layers that keep the air in and the debris out, and BEAM's radiation finding says the trade is not free.

Speculative And fifth, the experiment this brief hands off but which would change it most: rotate a habitat with people in it. Nothing has ever rotated and nothing funded rotates before the 2030s. Artificial Gravity owns the physiological question; what this brief would gain from an answer is a design constraint it currently does not have, because every station it describes was designed for the opposite condition on purpose.

Established The pressure-vessel requirements are met with stated margins, and this is the part of the subject that deserves to be called solved. NASA's recommended ultimate burst pressure for a habitable module is 60.8 psi, itself four times maximum operating pressure; the full-scale LIFE article reached 77 psi, a 27% margin, with Vectran restraint straps. BEAM met stress requirements at a fully loaded 109-CTBE configuration and was rated TRL 9 by the operator. Twenty-five years of continuous habitation across two programmes covers atmosphere management, thermal control, extravehicular activity, docking and crew operations.

Established The volume-per-kilogram requirement has an answer with a measured cost. BEAM packs 1,400 kg and expands 3.6 m³ to 16 m³, a 4.4-fold expansion; a 300 m³ module fits a five-metre fairing and a 1,400 m³ variant is proposed for seven metres. Frontier And the cost, stated by the programme that established the capability: higher trapped-radiation dose through the South Atlantic Anomaly “due to thinner shell and lack of equipment racks”, with printed-shield trials at 1.1, 3.3 and 10 mm inconclusive. A habitat's structure is also its shield, and making the structure lighter makes the shield thinner.

Established Micrometeoroid protection is specified as a dated claim, and the phrasing is the requirement. “Meets MMOD penetration requirements through 12/31/2028.” That is a modelled survivability against a projected environment with an expiry date, not a property of the hardware — and the environment it is projected against is adding roughly ten unintentional fragmentations and thousands of catalogued fragments a year.

Frontier The life-support requirement is where the specification stops being met. Water recovery is at 98% and sufficient. There is no flown system that closes food or waste at all, and the requirement for anything beyond a resupplied station is both. The engineering specification for a bioregenerative loop exists in the literature; the operating experience does not, and the published estimate of how much would be needed is four to eight years.

Speculative And a requirement that no design document states: how to inspect and repair a pressure boundary in orbit. The ISS case is the whole of the evidence — seven years, focus narrowed to welds, no root cause, a repair proposal involving cutting a bracket, and a decision that attempting it was more dangerous than continuing to measure. A habitat with a multi-decade design life needs an inspection method, a leak-localisation method and a repair concept, and this brief could not find any of the three specified anywhere.

Established O'Neill Cylinders is the nearest neighbour and the cut is regime rather than size. This brief owns habitats that exist or are funded and the practical engineering of leaks, micrometeoroid protection, environmental control and crew operations. FR-II-06 owns settlement-scale designs and the three comparative quantities that decide them — shielding mass per unit area, gravity and its cost, and where the mass comes from. The two overlap at exactly two points and each is cited once: this brief cites FR-II-06's shielding table to show that “solved” is magnetosphere-specific, and FR-II-06 cites this brief's operational record to show what a real pressure vessel costs to keep. Neither carries the other's source table.

Established Historical Space Colonization Concepts owns the 1970s station and settlement concepts, the Ames summer study, the solar-power-satellite economics and Biosphere 2. Skylab, Salyut and Mir sit on the seam and are treated here as operational heritage in one clause — long-duration habitation has a fifty-year record — with all programme history routed. In the other direction, FR-X-08 should not carry BEAM's TRL 9, the Commercial LEO Destinations programme, the leak, the debris statistics or the ISS closure figures.

Established Artificial Gravity owns rotation physiology, the centrifuge programme history and any published artificial-gravity roadmap commitment. This brief owns one sentence: no habitat has ever rotated, and microgravity is the product low-Earth-orbit stations sell rather than a failure to solve rotation. Frontier Space-Based Manufacturing owns in-space assembly and whether large structures can be built rather than launched; this brief owns the assembled-mass record of 420 tonnes at about $150 billion. Mars Colonization owns surface habitats and the 900-day closure requirement. And Space Resource Economies owns the market analysis behind GAO's demand finding, which this brief states as a fact rather than analysing.

Established The institutional picture is unusually well documented because a government audit office published it three months ago. ISS deorbit at end-2030 or early 2031, with an $843 million deorbit vehicle contract already let; about $3 billion a year to operate the incumbent; nearly $530 million spent on Phase 1 of the successor programme; $1–1.5 billion projected for Phases 2 and 3 over six years, which the agency itself says supports one station; six companies competing; and no assessment by NASA of the likelihood or duration of a gap in US presence, which is what GAO recommends it produce. Interest against the programme being audited, and it is the most decisive source in the subject.

Established The schedule record is the second institutional exhibit. Vast's Haven-1 slipped from 2026 to no earlier than Q1 2027 and depends on a docked Dragon for life support and power. Axiom restructured its entire assembly sequence in December 2024, inverting the order in which its modules launch and abandoning the plan to assemble at the ISS before detaching. Starlab is described as a 2029 article. None of those changes was driven by launch availability, which is the observation the framing has to answer.

Frontier The demand question is where the institutions and the technology part company. GAO flags uncertainty over whether the companies can “secure other customers to complete their business case”, citing industry concerns about market viability. A capability with one customer is a procurement, not a market, and a procurement funded at a level its own buyer describes as sufficient for one supplier will produce one supplier. Whether there is a private market for orbital volume is Space Resource Economies' question; that there is not yet one is this brief's finding.

Established The leak is also an institutional story and it is instructive. Two space agencies, jointly operating one vehicle, agreed on the observation and disagreed on the remedy: Roscosmos proposed cutting a bracket for access and NASA objected on safety grounds, so the repair was paused. A shared asset with divided technical authority and no agreed repair procedure is a governance arrangement as much as an engineering one, and it is the only precedent anyone has for maintaining a large habitat over decades.

Frontier And a structural asymmetry worth naming. The two long-duration habitats humanity operates are run by a multinational partnership approaching its end and by a single national programme in the middle of an announced expansion to 180 tonnes by 2027. The programme that is expanding is the one that is not commercially funded, which is at least suggestive about which institutional model currently produces habitats.

Frontier One further institutional exhibit, and it is an auditor's rather than a company's. NASA's Office of Inspector General examined the station transition in September 2024, and the recurring themes in that class of report are the ones this brief has reached from the other direction: schedule realism, the adequacy of the demand assumptions underpinning the commercial business cases, and the agency's own readiness to certify and hand over. Speculative The specific findings are summarised here from memory rather than quoted, and the report should be read before any figure from it is used anywhere. Established What is not in question is the institutional shape: two independent audit bodies examined the same programme inside roughly a year, and both asked the agency to state something it has not stated — whether there will be a gap in crewed presence, and how long it will last.

Established The clearest ethical fact is that crew are currently flying on a vehicle whose top-rated safety risk is unresolved. A leak rated 5 of 5 by the operator, seven years old, root cause unidentified, worsening in June 2026, with a repair judged too dangerous to attempt and five crew placed in a docked Dragon as a safe haven. Every part of that sentence is from the operator's own reporting, which is to its credit and does not make the situation better.

Frontier Second, the standard question that the commercial transition raises and nobody has answered publicly. The ISS operates under a government human-rating regime with an auditor. Six commercial stations are in development with one funded customer, a first article that borrows life support from a visiting spacecraft, and micrometeoroid certifications expressed as dated modelled claims. Who sets the standard for a privately operated habitat, and who audits it, is not established in any source in this pack.

Frontier Third, the debris externality. ESA reports compliance with mitigation guidelines as “insufficient” and projects business-as-usual risk at four times the acceptable sustainability threshold, driven substantially by fragmentation events that impose costs on every other operator. A habitat is the most exposed possible asset in that environment — large, permanently occupied and unable to manoeuvre quickly — and one has already had a porthole cracked.

Speculative Fourth, the gap. GAO's finding that NASA has not assessed the likelihood or duration of a break in US low-Earth-orbit presence is a risk-management observation, but the risk falls on people: crews whose careers depend on flight opportunities, researchers whose experiments have no platform, and international partners whose participation is structured around a station scheduled for disposal. An unassessed risk is one whose distribution has not been examined either.

Frontier And fifth, a smaller point that the record makes visible. The BEAM report published a finding unfavourable to its own module — higher trapped dose due to a thinner shell — and the ISS blog published a worsening leak on its own vehicle. Both are examples of an operator reporting against interest, and both are the reason this brief can be written at all. The ethical infrastructure that matters most in this subject is the habit of publishing the bad number.

Established Start with what has actually been achieved, because it is a civilizational fact and the framing's error is to draw the wrong conclusion from it. Human beings have lived continuously off Earth for twenty-five years. Two independent national programmes have sustained it. A pressure vessel made of fabric has been flown, struck by a micrometeoroid, instrumented and certified to the highest readiness level its operator awards. That is a permanent change in what the species has done, and it happened without anyone declaring a milestone.

Frontier The conclusion it does not support is that the rest follows from cheaper launch. The price fell about twentyfold and the habitat count did not move. What binds is the annual cost of operating a station, which is about $3 billion for the incumbent, and the absence of a buyer other than a government — and neither of those is a launch problem. The civilizational lesson is that access is necessary and nowhere near sufficient, and it generalises: the same pattern appears in Lunar Industry, where the delivered cost to the lunar surface rose 20% over the life of the contract designed to lower it.

Frontier The five qualifiers are the honest boundary of the achievement. Habitat engineering is solved in low Earth orbit, with resupply, inside the magnetosphere, for crews of single digits, on missions of months. Removing any one of them opens an unsolved problem: leaving the magnetosphere adds 13 to 78 tonnes per person of shielding; removing resupply requires closing food and waste, which nobody has done; extending to decades runs into a pressure boundary whose ageing behaviour is one unexplained leak; and scaling past single-digit crews has never been attempted anywhere.

Speculative And the checklist that a page like this can honestly close on. Pressure vessel and expandable structure: solved. Atmosphere, thermal, extravehicular activity, docking and crew operations: solved, across twenty-five years and two programmes. Pressure-boundary ageing and in-orbit repair: not solved — a seven-year leak of unknown root cause and a cracked porthole. Closure sufficient to remove the supply line: not solved — 98% of the water and none of the food. Radiation protection outside the magnetosphere: not solved as a mass problem. A customer other than a government: not solved. Two of six, and the four that are open are the four that would turn a station into a place to live.

These horizons track funded programmes, an audited budget and one unrepaired leak, because those are the three things that actually determine what exists in orbit:

• 10 yr: Established The ISS is scheduled for deorbit at the end of 2030 or early 2031, with the deorbit vehicle already under contract at up to $843 million. That is the most certain date in this brief. Frontier Expect one commercial station rather than six: NASA's own clarification is that the projected $1–1.5 billion for FY2026–31 supports one. Haven-1 is no earlier than Q1 2027 as a single module dependent on a docked Dragon; Axiom's first element is no earlier than 2027 and its habitat no earlier than 2028; Starlab is described as 2029. Frontier Expect a gap in US low-Earth-orbit presence to be a live possibility that has still not been formally assessed unless GAO's recommendation is acted on. Speculative Expect nothing to rotate. Frontier And expect the binding date in this window to be a certification date rather than a launch date: the interval between a first element reaching orbit and that element being certified for NASA crew is the quantity that decides whether there is a gap, and it is currently unpublished for every provider.
• 25 yr: Frontier If a commercial market appears, this is the horizon on which multi-module stations with non-government customers are plausible; if it does not, the pattern of one government-anchored station continues, and the evidence today points at the second. Speculative Closure is the capability most likely to change the character of a habitat on this horizon, and only if somebody funds a multi-year closed facility — four to eight years of operating experience is the published estimate and no facility is accumulating it. Frontier The debris environment is the constraint most likely to have got materially worse, on ESA's own business-as-usual projection.
• 50 yr: Speculative A habitat outside the magnetosphere is the qualitative step, and its cost is arithmetic rather than invention: 13 to 78 tonnes per person of shielding on top of the roughly 60 tonnes per crew member that the ISS already represents. Route the settlement-scale version to O'Neill Cylinders. Speculative A rotating habitat is plausible at this horizon only if the physiological case is made first, which is Artificial Gravity's question, and if microgravity stops being the commercial product, which is a market change rather than a technical one.
• 100 / 250+ yr: Handwave Beyond useful forecasting. The defensible statement is that the two things this brief finds unsolved are of different kinds: closure and pressure-boundary maintenance are engineering problems with no known obstacle, while a customer is a fact about the world that no amount of engineering produces. Speculative A subject in which the enabling technology got twenty times cheaper and the deployed capability did not change is a subject whose base rate is set by demand, and nothing in this pack forecasts demand at this horizon.

• Depends on Two edges, both narrow on purpose. Artificial Gravity, because the largest open question about long-duration habitation is whether microgravity is survivable indefinitely and what replaces it if not — and because no habitat has ever rotated and nothing funded rotates before the 2030s. FR-I-02 owns the physiology, the tolerable rotation rate and the countermeasure record; this brief owns only the observation that microgravity is the product low-Earth-orbit stations sell, which is a commercial choice rather than a technical failure. And Space-Based Manufacturing, because the assembly record is the ceiling on habitat scale: the ISS is about 420 tonnes across dozens of flights at roughly $150 billion — $0.36 billion per tonne, the only empirical figure for assembled orbital mass, and the current commercial designs avoid the problem by launching a single module sized to a fairing rather than solving it.
• Requires (not on this map) The first is the binding constraint and it is not a technology: GAO reports six companies developing commercial stations against NASA funding of $1–1.5 billion for FY2026–31 that NASA itself says is “only sufficient to support one”, with explicit uncertainty over whether any of them can secure customers other than NASA to complete their business case. The second is a method nobody has: the ISS has been losing air since 2019 through cracks whose root cause two agencies have not identified after narrowing their focus to welds, and the proposed structural repair was paused because it was judged more dangerous than continuing to measure. The third is a rule rather than a technology and it gates the whole transition: a route by which the agency certifies a privately owned, continuously occupied facility for its own crew. Commercial Crew shows the step exists and shows it takes years whose duration does not follow from the development schedule; nothing on this map produces the requirements, the verification regime or the agency capacity to apply them, and the interval it consumes is the difference between a transition and a gap.
• Enables No typed enabling edge is claimed, and the record is the reason. Launch cost fell about twentyfold and produced no new large habitat; the successor programme has six competitors and funding its own customer describes as sufficient for one; and GAO has not been given an assessment of whether there will be a gap in US low-Earth-orbit presence at all. O'Neill Cylinders and Mars Colonization both draw on this brief's operational record — the leak, the closure figures, the debris counts, the burst margins — but they declare those edges from their own side, where the depending brief can say what it is waiting for.
• Adjacent O'Neill Cylinders across a regime boundary: that brief owns settlement-scale design and the shielding areal densities, this one owns habitats that exist or are funded. Historical Space Colonization Concepts owns Skylab, Salyut, Mir and the 1970s concepts as programme history. Mars Colonization owns surface habitats and the 900-day closure requirement and supplies the closure ladder cited here. And Space Resource Economies owns the market analysis behind the demand finding that this brief states as an audited fact.

Established “Habitats are just waiting on cheap launch.” Launch cost to low Earth orbit fell from about $54,000/kg in the Shuttle era to about $2,700/kg on a reused Falcon 9 — a factor of about twenty, achieved. Established In the same period no new large habitat was launched by anyone except China, the ISS was not expanded, the first commercial station slipped to no earlier than Q1 2027 as a single 14-tonne module that borrows life support and power from a docked Dragon, and GAO reported that NASA's projected funding is “only sufficient to support one commercial space station” out of six competitors. The constraint was the customer and the roughly $3 billion a year it costs to operate a station, not the ride.

Speculative “We know how to keep a pressure vessel airtight; that part's routine.” The most mature habitat ever built has been losing air since 2019 through cracks in the Zvezda transfer tunnel; the rate went from 2.4 to 3.7 lb/day in two months of 2024, was rated 5 on NASA's 1-to-5 risk scale, rose again to two pounds per day in June 2026, and the structural repair was paused after NASA objected that cutting a bracket for access was unsafe — with five crew placed in a Dragon safe haven during the assessment. Established Two agencies narrowed their focus to internal and external welds and have not identified a root cause. And the other station humanity operates had a porthole cracked by a debris impact. Building a pressure vessel is solved; ageing and repairing one in orbit is not, and the entire dataset is two vehicles.

Established “The ISS recycles everything.” It recycles 98% of the water, which is genuinely excellent and meets NASA's own stated beyond-LEO requirement. It recycles none of the food. Established The best food closure ever achieved with humans anywhere is 60%, over a year, with waste recycling not closed; the longest human closure of any kind is six months; and the 2025 review statement is that no nation has demonstrated a completely closed system including waste processing, with four to eight years of operational experience estimated as the path to full readiness.

Established “Inflatable habitats are unproven.” They are proven and the proving is documented. BEAM packed at 1,400 kg, expanded 3.6 to 16 m³, survived a micrometeoroid impact estimated at 260 g at the restraint layer with no full penetration, met stress requirements fully loaded, and was assessed by NASA as having “advanced human rated expandable modules to TRL 9”; a full-scale successor burst at 77 psi against a 60.8 psi requirement. Frontier What the same report also says, against its own interest, is that trapped-radiation dose through the South Atlantic Anomaly is higher in BEAM than in metal modules “due to thinner shell and lack of equipment racks”, with printed-shield trials inconclusive. Light structure buys volume and costs shielding.

Established “Micrometeoroid protection is a solved property of the hardware.” The certification language is the correction: BEAM “meets MMOD penetration requirements through 12/31/2028”. Established That is a dated modelled claim against a changing environment — and the environment now holds about 1.2 million objects between 1 and 10 cm, too small to track and too large to stop with a bumper, growing by roughly 9.8 unintentional fragmentations a year, with ESA calling mitigation compliance “insufficient” and projecting business-as-usual risk at four times the acceptable sustainability threshold.

Frontier “The ISS proves we can live in deep space.” Everything solved was solved inside the magnetosphere. Outside it the measured transit dose is 1.84 mSv/day, a settlement dose target needs 6–7 t/m² of polyethylene or 10–11 of regolith, and galactic cosmic radiation halved or quartered costs 13 or 78 tonnes per person. Frontier Derived: the ISS is about 420 tonnes for a crew of seven, roughly 60 tonnes per person with no dedicated shielding at all, so a deep-space equivalent at 50 g/cm² would more than double the per-person mass of a $150-billion object to buy a fourfold dose reduction.

Speculative “Today's stations are stepping stones to rotating habitats.” Nothing has ever rotated, nothing funded rotates before the 2030s, and microgravity is the product a low-Earth-orbit station sells — which makes the two designs point away from each other rather than toward each other. Frontier That is a commercial fact rather than a technical failure, and whether long-duration microgravity is survivable belongs entirely to Artificial Gravity.

Frontier “Starship will change all of this.” Twelve flights, seven successes, Block 3 flying and a fairing large enough for single-launch station modules is a real capability and Starlab's design assumes it. Established But the price that would change the arithmetic — $100–200/kg — is labelled by the compiler who published it as “a design target, not an achieved price”, and the vehicle has never attempted the ship-to-ship propellant transfer its wider architecture needs. And the deeper problem is that the last twentyfold price reduction did not produce habitats either.

Frontier “There will obviously be a commercial station before the ISS is deorbited.” GAO makes no such statement. What it reports is six competitors, funding for one, Haven-1 at NET Q1 2027, Axiom's assembly sequence restructured in December 2024 with its first element NET 2027 and habitat NET 2028, Starlab in 2029 — and that NASA “has not yet assessed the likelihood or duration of a gap” despite identifying it as a critical concern.

Established And the framing itself. “Habitat engineering is a solved problem waiting on launch cost” has two halves and both fail, in different ways. Established The launch-cost half is falsified by the launch-cost data. The solved half is true only inside five silent qualifiers — in low Earth orbit, with resupply, inside the magnetosphere, for crews of single digits, on missions of months. Established The honest version deserves stating warmly, because it is a real achievement: pressure vessels, expandable structures, atmosphere and thermal management, extravehicular activity and long-duration crew operations are flight-proven and certified, with BEAM at TRL 9 and LIFE bursting 27% above a requirement that is itself four times operating pressure. What is unsolved is everything that would let a habitat stop being resupplied, stop being repaired from the ground, and stop being inside Earth's magnetic field.