Artificial Gravity

Artificial gravity means providing a persistent apparent weight where there is none, chiefly for humans living in space. Two very different things share the name and sit at opposite ends of the honesty scale: (a) simulating gravity with acceleration — a spinning habitat, a short-arm centrifuge, or a ship under continuous thrust — and (b) generating a real gravitational field on demand. The second has no mechanism and belongs to Gravity Modification. This brief is about the first.

The framing under test is that rotating habitats solve the physiology problem, and this slot is the outlier in its cluster. The other exotic-physics briefs in Category I are about the absence of evidence. This one has sixty years of it: rotating rooms, short-arm centrifuges, a sixty-day randomised bed-rest trial, an orbital mouse centrifuge that has actually run at lunar gravity, and a bone-loss literature with meta-analysed numbers. Almost every claim below is established, and that is what makes the brief unusual.

It is also why the blocker here is different in kind from the rest of the cluster. Inertial Manipulation and Gravitational Wave Engineering are blocked by the absence of physics. This slot is blocked by a facility, a budget and a decade — which is the clearest illustration anywhere on this map of what the difference between a scientific and an industrial dependency actually looks like.

The result of taking the evidence seriously is that the framing survives in a narrower and more useful form. Rotation is a solved way to produce apparent weight. Producing weight is not the same as producing health, and the mapping between them is measured only in mice.

Established The physics is finished, and this is the one slot in the cluster where that sentence is literally true. Spin a structure and its occupants feel a centripetal acceleration locally indistinguishable from gravity; the equivalence principle is what guarantees the indistinguishability, and it is measured to 2.7 × 10⁻¹⁵ by MICROSCOPE on macroscopic test masses and to about 1.6 × 10⁻¹² by atom interferometry on quantum superpositions. Established Nothing in this brief is contingent on a discovery. Continuous linear acceleration works too — a ship thrusting at 1 g would feel like Earth — but sustaining it needs propellant and power belonging to Advanced Nuclear Propulsion and Fusion Spacecraft, not here.

Established The problem statement has numbers, and they are worse than “astronauts get weaker”. Stavnichuk and colleagues pooled 25 articles covering roughly 189 astronauts, 148 of them with bone-density measurements. Loss is site-dependent: lower limbs −0.8% per month [−1.1, −0.5], lumbar spine and pelvis −6.2% total [−6.7, −5.6], upper limbs and thorax −0.7%, and the skull +2.2% [+1.1, +3.3]. The authors state that exercise regimes “only partially protected against bone loss”. Established And a substantial part of it does not come back: Gabel and colleagues followed 17 astronauts through missions averaging 170 days and found that twelve months after return, distal tibia failure load and trabecular bone volume fraction and thickness remained 0.9–2.1% below preflight, with nine of seventeen showing incomplete recovery of total bone mineral density — a deficit the authors compare to “a decade or more” of terrestrial age-related loss.

Established The rotation-tolerance number has moved, and by a factor of sixty. Clark and Hardy proposed a maximum of 0.1 rpm in 1960, implying a radius of about 90,000 metres for 1 g. Stone proposed 6 rpm in 1973 — three times the nausea threshold — implying about 25 metres, and that became the consensus comfort zone alongside a minimum 12-metre radius, 0.3 to 1.0 g, and a gravity gradient under about 15%. Frontier The modern adaptation literature says the conservative end was far too conservative: Young and colleagues demonstrated adaptation to head movements at 23 rpm in 2001, and MIT's Human Systems Laboratory works with short-radius centrifuges under 2.5 metres in diameter requiring about 30 rpm for 1 g at the rim, reporting that inappropriate vestibulo-ocular reflexes, motion sickness and perceptual illusions are all reduced after several adaptation periods.

Established The strongest evidence that the 1973 limit is not binding is operational rather than theoretical. The AGBRESA study ran daily centrifugation at a mean of 30.5 ± 1.0 rpm for sixty days on bed-rest subjects and terminated only 10 of 960 runs, about 1%, mostly for presyncopal symptoms. Established What has not moved is the gravity gradient. For radii under ten metres, NASA's own evidence report gives a head-to-foot gradient of 20 to 100%, “which may be perceived as a bent posture”. A short-radius centrifuge is not a small habitat; it is a fundamentally different stimulus, and conflating the two is the commonest error in the popular treatment.

Established And then the central result: the only randomised controlled trial of artificial gravity as a countermeasure missed its primary endpoint. AGBRESA — the Artificial Gravity Bed Rest study, run at DLR's :envihab with NASA and ESA — put 24 participants through 60 days of 6° head-down bed rest in three arms of eight: intermittent artificial gravity (six 5-minute exposures with 3-minute breaks), continuous artificial gravity (one 30-minute session), and control. Loading was +1 Gz at the centre of mass and about +2 Gz at the feet. Established The outcomes, from Kramer and colleagues: VO2max fell 20 ± 5%, 23 ± 5% and 24 ± 8% across intermittent, continuous and control, with no significant difference between arms. Jump power fell 25%, 26% and 33%; plantar flexion strength 35%, 31% and 48%; rate of force development 28%, 12% and 40%. Established The authors' conclusion is unambiguous: “The AG protocols were not suitable to maintain aerobic exercise capacity … AG would have to be combined with adequate exercise protocols before potentially being considered as an integral countermeasure.”

Established Thirty minutes a day of 1 g at the heart did essentially nothing for cardiovascular deconditioning and produced a real but partial benefit for muscle. This does not falsify rotating habitats, which propose continuous gravity rather than a daily dose. It demolishes the intermittent-centrifuge shortcut — the option the field has been exploring precisely because it is cheaper than a rotating habitat. Frontier A separate dose-response study on a short-arm human centrifuge puts the median effective dose for cardiovascular response near 1.08 g for mean arterial pressure, 1.13 g for cardiac output and 0.91 g for heart rate, and finds that 1.7 g most closely approximates standing. Read carefully, that is an argument that the effective dose is above 1 g at the relevant body location, not below it.

Frontier Partial-gravity physiology is usually described as a number nobody has measured. That is right for humans and wrong for animals, and the distinction is a finding rather than a correction. JAXA's Multiple Artificial-gravity Research System on the ISS has run mice at lunar gravity across three missions: MHU-1 in 2016 at microgravity and artificial 1 g for 35 days, MHU-4 in 2019 at one-sixth g for 25 days, and MHU-5 in 2020 at one-sixth g for 26 days. Established Bone, from Okamura and colleagues: trabecular bone volume fell 70% in microgravity against 28 to 30% at lunar gravity, and mostly recovered at artificial 1 g; cortical cross-sectional area fell 17% against 5 to 7%; cortical thickness 16% against 5 to 7%; thymus weight ran at 56.6%, 76 to 79% and 79.8% of ground control across the three conditions.

Established Muscle tells a different and more interesting story. Hayashi and colleagues found soleus mass 17% lower in microgravity than in ground control and similar between ground control and lunar gravity; myofibre cross-sectional area fell 40% in microgravity and not at all in partial or artificial gravity. Established But fibre type shifted anyway — type IIa down 43% and type IIb up roughly tenfold in microgravity — and lunar gravity only partially reverted the shift where 1 g prevented it entirely. Frontier The shape of the answer is therefore already visible: partial gravity is not proportionally protective, and it protects some systems much better than others. Lunar gravity recovered most of the muscle mass, less than half of the bone, and did not normalise fibre type.

Speculative The one quantitative human claim is a model rather than a measurement, and it points the other way. Harry Jones at NASA Ames concludes that “partial gravity exposure below 0.4 g seems to be insufficient to maintain musculoskeletal and cardiopulmonary properties in the long-term”, with modelled bone mineral density losses of 0.39% per week at lunar gravity and 0.22% per week at Mars gravity — rates similar to or above microgravity calcium loss — and bone strength “seriously reduced in two or three years”. Frontier That floor is above Mars gravity. It is also a model, the mouse data cut against it, and the honest position is to declare the tie rather than average two incommensurable kinds of evidence.

Established NASA names five gaps and has not closed any of them since at least 2015. The Human Research Program evidence report on Artificial Gravity lists them: the minimum effective gravity level is unknown; the effect of the radius-dependent gravity gradient is unclear; the sensorimotor consequences across the range of rotation rates are undefined; continuous versus intermittent exposure duration and frequency is unresolved; and the interaction with exercise countermeasures is unstudied. Established The report states flatly that “the physiological responses to continuous Mars g exposure are also unknown”. Established One calibration datum from the same report belongs on the page because it shows how coarse the human instrument is: astronauts needed somewhere between 0.22 and 0.5 g merely to perceive artificial gravity, against a theoretical sensory threshold of 0.007 g.

Established The in-space demonstration record is thinner than almost anyone assumes. On 14 September 1966, Gemini XI was tethered to an Agena target vehicle and spun with side thrusters, producing about 0.00015 g — pilot Dick Gordon: “There is an artificial gravity field. It makes the camera move back very rapidly.” Established That is the entire human record: one demonstration, sixty years ago, at one part in seven thousand of Earth gravity. Every subsequent orbital artificial gravity has been in a centrifuge for rodents, plants or cells, and no human has ever experienced sustained artificial gravity in orbit. Established Meanwhile the design canon is equally old: O'Neill's Island Three cylinder, published in Physics Today in September 1974 and in The High Frontier in 1976, specifies 6.4 to 8.0 km in diameter, 32 km long, rotating about 28 times an hour — roughly 0.47 rpm — for 1 g at the inner wall. It is fifty years old, dimensionally consistent, and nothing resembling it has been built or costed by anyone with the money.

Frontier The live questions here are practical rather than fundamental, which is what distinguishes this slot from the rest of the cluster — but “practical” does not mean “nearly answered”. Each of NASA's five gaps is a measurement that requires a facility that does not exist.

Frontier Position one: continuous rotational artificial gravity at 1 g fully solves the physiology problem. This is the O'Neill tradition and most habitat design work, and it is better supported than the older version of this page implied — the MARS mouse data show artificial 1 g in orbit restoring bone, preserving muscle mass, largely normalising fibre type and bringing thymus weight to within a point of ground control. Frontier What would settle it is a human at continuous 1 g in orbit for months, which has never happened and for which no vehicle exists.

Frontier Position two: intermittent short-arm centrifugation is a sufficient countermeasure. The only randomised trial found no aerobic benefit and partial muscular benefit at thirty minutes a day over sixty days. The trend is negative and the authors themselves name the remedy — a longer or higher-dose protocol, or combination with exercise. Frontier The dose tested was insufficient; whether a sufficient dose exists inside the tolerable envelope is open.

Frontier Position three: partial gravity at Moon or Mars level is sufficient for long-term health. This is implicitly assumed by every settlement plan and the evidence genuinely cuts both ways. Against: modelled human losses of 0.39% per week at lunar gravity and 0.22% at Mars, and a proposed 0.4 g floor. For: lunar gravity in mice preserved soleus mass and myofibre size outright and cut trabecular bone loss from 70% to under 30%. Frontier Declare the tie. A model and a rodent are different kinds of evidence and neither is a human at partial gravity for a year.

Frontier Position four is the most under-covered live position in the subject: partial gravity protects some systems and not others. Read the MARS results literally — muscle mass preserved at one-sixth g, fibre type not, bone partly, thymus partly. Frontier If that generalises to humans, the design question stops being “what is the minimum g” and becomes “which system are you protecting”, and a settlement architecture might need different answers for skeleton, muscle, immune function and vision.

Frontier Position five: the classical 6 rpm comfort limit is far too conservative and small habitats are viable. Adaptation is demonstrated at 23 rpm and 30.5 rpm was sustained daily for sixty days with a 1% termination rate. Established Position six, which competes with it: the gravity gradient rather than the rotation rate is the binding human-factors constraint — 20 to 100% head-to-foot below ten metres of radius is a physical fact, and whether humans tolerate it chronically is not known, because the trials dosed rather than housed their subjects.

Frontier Position seven, and the one the operational record actually endorses: exercise countermeasures alone will suffice and artificial gravity is unnecessary. This is the ISS programme's revealed preference — the advanced resistive exercise device and the exercise protocol are what actually flies. Established The evidence against it is direct: exercise “only partially protected” in meta-analysis, and nine of seventeen astronauts had incomplete bone mineral density recovery a year after return with the full protocol in place. Frontier The revealed preference is evidence about agency priorities rather than about efficacy.

Speculative Position eight: artificial gravity would fix Spaceflight Associated Neuro-ocular Syndrome. This is frequently assumed and has no evidential support at all. Roughly 15% of astronauts show clinically significant optic disc oedema after long-duration flight, optical coherence tomography detects some degree of it in nearly all of them, and choroidal folds appear in over 20%. Speculative The leading hypotheses are cephalad fluid shift and compartmentalised cerebrospinal fluid, which a rotating habitat would plausibly address — but the field's principal review of the syndrome does not discuss artificial gravity as a countermeasure at all, and the risk that most threatens long missions therefore has zero artificial-gravity evidence in either direction.

Speculative Position nine: tethered spin is the cheap path. Mission-architecture studies have proposed it since Gemini; the flight record is one demonstration at 0.00015 g in 1966 and nothing since, and no tether system has been flown at habitat scale. Speculative Position ten: commercial stations will deliver artificial gravity before agencies do. One company has published a 2035 commitment; its flown record is a single module. Established Position eleven, continuous-thrust artificial gravity, is trivial physics and out of reach propulsively, and belongs to the nuclear and fusion propulsion briefs. Handwave Position twelve is field-generated gravity plating, which is fiction and routes to Gravity Modification.

Established The bottlenecks here are cost, scale and a missing facility — not physics, and not even, primarily, engineering difficulty. This is the load-bearing distinction between this slot and its neighbours, and the evidence for it is that the hard object has already been built once.

Established The launch-cost trend is real and has already delivered most of what it can. Published list prices divided by maximum stated payload to low Earth orbit, in 2025 dollars: the Space Shuttle at about $54,000/kg, reused Falcon 9 at about $3,000/kg, rideshare slots near $6,000/kg. Starship's target of under $100–200/kg is labelled by the compiler as a design goal rather than an achieved price. Frontier An eighteen-fold reduction has already happened and it has not produced a rotating habitat, which is the point: the binding constraint is not only price per kilogram but the number of kilograms. A structure at O'Neill scale is not a launch problem at any price; a structure at the scale currently being marketed is a launch problem in the hundreds of tonnes. Cheap heavy launch is necessary and demonstrably not sufficient.

Established The second bottleneck is a centrifuge, and its history is the argument. The ISS Centrifuge Accommodations Module was built by NASDA and owned by NASA under a barter agreement in which Japan invested approximately $700 million in exchange for NASA launching the Kibo module. The centrifuge was 2.5 metres in diameter, designed to expose specimens to 0.01 g to 2 g and to run two different gravity levels simultaneously — a purpose-built partial-gravity instrument. Established NASA issued a stop-work order to the contractor in the week of 1 August to redirect funds to the Crew Exploration Vehicle, affecting 42 contractor employees, 15 NASA civil servants and 60 support contractors at Ames, and the module was cancelled in 2005 on ISS cost overruns and Shuttle scheduling, alongside the Habitation Module and Crew Return Vehicle. The flight model was manufactured, never flew, and is now on outdoor display at the Tsukuba Space Center.

Established And it happened twice. ESA's AGREE — Artificial Gravity with Ergometric Exercise — was cancelled in 2013 “following a stress analysis showing that the ISS nodes would have been structurally compromised”. Established Two space agencies have separately funded, designed and abandoned an orbital centrifuge, once for budget reallocation and once for structural incompatibility with the host station, and NASA's own evidence report states the consequence flatly: “a human-rated centrifuge is currently not available on board the ISS”. The largest radius ever flown is 15 centimetres, against the 5 metres a proposed deep-space habitat centrifuge would need.

Established The third bottleneck is a decade of data collection that nobody has started. Five named gaps, unclosed since at least 2015; one randomised trial, sixty days, primary endpoint missed; and the human partial-gravity response described as unknown in the responsible agency's own words. Established None of these is a discovery. All three are a facility, a budget and a decade, which is exactly what the three requires rows say.

Frontier A fourth, genuinely engineering bottleneck deserves naming even though it is not the binding one: spin-up, balance control and mass-property management on a large rotating structure with moving crew and consumables, plus the structural design of a kilometre-scale spinning body. These are demanding and unsolved in practice, and no vehicle has ever had to solve them — but they sit downstream of the mass and the facility, not upstream.

Established This brief waits on no result another brief produces, and that is the finding rather than a gap. Nothing here requires new physics; the equivalence principle is measured to fifteen decimal places, rotation is textbook mechanics, and the physiological problem is characterised in humans with pooled numbers. What the slot waits on is three things that a legislature, an agency and a decade could supply.

Established Mass to orbit is the first, and it is shared. Space-Based Manufacturing and Deep Space Infrastructure own how habitat-scale mass gets there and gets assembled — mass drivers, lunar regolith, in-space construction. This brief owns why you would want it and at what spin rate; the industrial argument is theirs.

Established The physiology is shared too, and the division is by question rather than by dataset. The bone, muscle, ocular and immune data belong equally to the space-physiology literature. This brief's specific claim on them is the dose–response question: how much g, for how long, at what radius, for which system. General microgravity deconditioning and its countermeasures are somebody else's; the numbers are cited here because the dose-response question cannot be posed without them.

Frontier There is one methodological sibling worth naming explicitly, because it makes the generic finding legible. Large ground and space observatories have the same shape of blocker: an instrument that exists on paper, costed, designed, and repeatedly deferred. Established In both cases the missing thing is an appropriations decision. Saying that plainly is more honest than describing the obstacle as technical.

Established The needed experiments are concrete, fundable and unglamorous, and each one closes a named gap. None requires an invention.

Established First: an orbital centrifuge sized for animals at multiple partial-gravity levels, run for longer than a month. The MARS system has demonstrated the method and produced the only in-orbit partial-gravity data that exist, at 25 to 35 days per mission and a 15-centimetre radius. Extending it to several gravity levels — Mars gravity at 0.38 g has never been flown for animals — and to multi-month durations would convert the current mouse snapshot into a dose-response curve. Frontier This is the cheapest high-value experiment in the slot and nothing prevents it but scheduling and money.

Established Second: a human-rated centrifuge in orbit. This is the experiment that answers the question the field is actually asking, and it is the one that has been cancelled twice. A 2.5-metre instrument spanning 0.01 to 2 g was built and never flown; a station-integrated ergometric version was cancelled on structural grounds. Established The design work is done and the hardware has been manufactured once.

Frontier Third: a higher-dose or combined-modality bed-rest trial. AGBRESA's own authors specify the follow-up — artificial gravity combined with adequate exercise protocols, or a longer or higher-dose exposure. The dose-response work on a short-arm centrifuge suggests the effective cardiovascular dose may sit above 1 g at the relevant body location, which would mean AGBRESA tested below threshold rather than testing a failed principle. Frontier That is a discriminating experiment and it is a bed-rest study, not a spaceflight.

Frontier Fourth: long-duration living at high rotation rates rather than dosing at them. Everything known about tolerance above 6 rpm comes from exposures of minutes to half an hour. Nobody has slept, eaten, worked or turned their head casually at 20 to 30 rpm for weeks. Established That is a ground experiment in a rotating room, the technique is sixty years old, and the answer would decide whether small habitats are viable at all.

Speculative Fifth, and the gap with no experiment attached: whether artificial gravity does anything for the ocular syndrome. No study has looked. The mechanism hypotheses are about fluid distribution, which is exactly what a gravity gradient changes, and the absence of even a pilot study on the risk that most threatens long missions is the strangest hole in the literature.

Established Requirements can be stated here, which is not true of most of this cluster, and they are demanding rather than impossible. Large tethered or rigid rotating structures; reliable spin-up, de-spin and balance control against moving mass; docking and crew transfer to a rotating body; radiation shielding; and life support at habitat scale. The challenge is integration and cost, not feasibility.

Established The radius-versus-rate trade is the design's spine and it is fully specified. One g at 0.47 rpm needs a radius of kilometres; at 3.5 rpm it needs roughly 70 metres; at 6 rpm about 25 metres; at 30 rpm under 1.3 metres. Established The gradient scales the other way — 20 to 100% head-to-foot below ten metres — so the small end of the trade buys a structure that is cheap to launch and delivers a stimulus no human has lived in.

Frontier The commercial claim currently on the table sits between the two historical design points. Vast's published roadmap commits to an artificial-gravity station by 2035, “generating artificial gravity by rotating end over end at 3.5 RPM”, with 950 cubic metres habitable and 2,160 pressurised for 40 crew. Established What exists is much smaller: Haven-1 is a single module of 45 cubic metres habitable, 14,000 kg, four crew, missions to two weeks, now targeting the first quarter of 2027 after slipping from May 2026, and trade coverage records no spin demonstration planned for Haven-1 itself. Speculative The 3.5 rpm figure implies a radius of roughly 70 metres for 1 g — an order of magnitude larger than anything the company has flown or is building. Treat the 2035 date as a statement of intent, not a schedule.

Established The oldest specification remains the most complete. O'Neill's Island Three is dimensionally consistent at 6.4 to 8.0 km diameter and 32 km length rotating at about 0.47 rpm, and it assumes lunar material delivered by mass drivers rather than launched from Earth — which is the correct architectural response to the mass problem and the reason the design has outlived every proposal that ignored it.

Established Artificial gravity is less a standalone technology than a property of any serious long-duration habitat, which is why its adjacencies are the whole habitat stack: space stations and habitats, in-space manufacturing, orbital tethers, radiation shielding, closed-loop life support, and low-cost heavy launch.

Established Within this map the sharpest adjacency is a contrast rather than an overlap. Gravity Modification shares a name with this brief in fiction and nothing else: that one is a field you switch on and it has no mechanism; this one is acceleration you feel as weight and it is textbook physics. Inertial Manipulation is the exact complement — it would remove a felt force where this adds one — and the pairing is the clearest illustration on the map of the difference between a slot blocked by facilities and a slot blocked by the absence of physics.

Established Two propulsion briefs own the continuous-thrust variant. A ship under sustained 1 g is artificial gravity by acceleration and is a propellant and power problem: Advanced Nuclear Propulsion and Fusion Spacecraft carry the arithmetic. Established And Space-Based Manufacturing and Deep Space Infrastructure own the mass problem this brief's first requires row points at.

Established The institutional history here is the strongest evidence in the brief, because it is a record of decisions rather than of difficulty. A partial-gravity centrifuge was funded to roughly $700 million, built, and stopped by a stop-work order redirecting money to a different vehicle. A second was cancelled because it did not fit the station structurally. Neither cancellation was a scientific judgement about artificial gravity.

Established The revealed preference of the operational programme is exercise, and it has been for twenty-five years. Resistive exercise devices fly; centrifuges do not. That is a legible institutional choice with a legible cost: nine of seventeen astronauts with incomplete bone recovery a year after landing, under the full protocol. Frontier Whether the choice is right depends on data the same institutions have declined to collect, which makes it hard to evaluate and easy to perpetuate.

Established The programme this needs is decade-scale and nobody owns it. Five gaps named in a 2015 evidence report and restated in a 2017 international roadmap remain open in 2026. The roadmap surveys proposed protocols ranging from five minutes a day to ninety days of continuous exposure — a spread of four orders of magnitude in dose, which is what a field looks like before its dose-response study is done. Established Human research of this kind cannot be done quickly, because the endpoints are months to years long, and it therefore requires exactly the kind of stable multi-decade commitment that agency budget cycles are worst at.

Frontier The commercial entrants change the funding shape but not yet the evidence. A private roadmap with a 2035 artificial-gravity commitment is a real signal about capital availability and no signal at all about physiology, and no commercial operator has proposed the multi-year partial-gravity health study the field actually needs. Speculative A plausible division of labour has agencies buying the physiology and companies buying the structure; nothing in the current programme record indicates that either has been arranged.

Established The ethical questions here are unusually concrete, because the risk is measured. Informed consent for long-duration crews means telling them that roughly 15% will develop clinically significant optic disc oedema, that the majority will show some degree of it, that lower-limb bone loss runs at 0.8% a month, and that nine of seventeen astronauts in the best longitudinal study had not recovered bone mineral density a year after return. Those are numbers a participant can weigh.

Frontier The harder consent question is about partial gravity and it cannot currently be answered. Anyone signing up for a multi-year lunar or Martian stay is consenting to a dose whose long-term effect on humans is, in the responsible agency's own words, unknown — with one model saying the floor is above Mars gravity and one set of mouse data saying protection is real but uneven. Speculative Consent to an unquantified risk is a different act from consent to a quantified one, and settlement advocacy tends not to make the distinction.

Frontier There is also a research-ethics question about the animals, and it should be stated rather than skipped. The only in-orbit partial-gravity data that exist came from mice flown for 25 to 35 days in an orbital centrifuge and analysed post mortem. That is the experiment that produced the field's most useful recent finding, and expanding it — more gravity levels, longer durations — means more animals. Established The alternative is continuing to send humans on multi-year missions with no dose-response data at all.

Speculative And the distributional question, at the far end. If sustained health in space turns out to require continuous 1 g in a structure at the scale the design canon specifies, then living in space is an activity available to whoever can build kilometre-scale rotating bodies — which is a much smaller set than whoever can buy a launch. The physiology result would then be doing distributional work nobody has framed as such.

Established Reliable artificial gravity is what would make indefinite human presence in space healthy rather than merely survivable, and it underwrites everything from long transits to permanent orbital settlement. That claim is not in dispute; what is in dispute is how much gravity, delivered how, is enough.

Frontier The partial-gravity question is the civilizational fork, and it is genuinely undecided. If Mars gravity is sufficient, planetary settlement is a transport and life-support problem. If the floor is above 0.38 g — as the one modelling paper argues — then permanent settlement on Mars requires either rotating structures on the surface, intermittent centrifugation as a lifelong medical regime, or an accepted health cost, and the Moon is worse on every count. Frontier A single well-designed experiment decides which civilizational architecture is available, and it has not been run.

Frontier The mouse data hint at an outcome nobody has planned for: uneven protection. If partial gravity preserves muscle mass but not fibre type, most of bone but not all, and some immune function, then a partial-gravity population is not a slightly-deconditioned Earth population — it is a physiologically distinct one, with a different medical profile and a different relationship to returning. Speculative Nothing in the settlement literature is written for that case.

Established And the honest summary at civilizational scale is that this is a subject where the answer is purchasable. Unlike every neighbouring brief in this category, nothing here waits on a discovery. A centrifuge, a decade and a health programme would settle the questions that decide whether humans can live off Earth at all — and two agencies have separately started building the centrifuge and stopped.

The unusual feature of this topic is that the physics is done, so these horizons track appropriations, facilities and study durations rather than discovery:

• 10 yr: Established Two things are already scheduled and will resolve in this window: the first commercial station module flies or does not, on a date that has already slipped a year, and it carries no spin demonstration; and the ISS reaches end of life, taking with it the only platform on which the MARS mouse centrifuge operates. Frontier Expect further rodent partial-gravity results if a successor centrifuge is manifested, and expect no human partial-gravity data at all, because no facility capable of producing them is under construction. Frontier Expect at least one further bed-rest trial responding to AGBRESA's own recommendation to combine artificial gravity with exercise; that is a ground study and it is fundable now. Speculative The live hypothesis worth watching is whether any agency re-manifests a human-rated centrifuge after two cancellations.
• 25 yr: Frontier If cheap heavy launch delivers on its stated targets, the first crewed rotating structure is a plausible object in this window — at the small end of the radius trade, which means the gravity-gradient question becomes operational before it is answered scientifically. Speculative The commercial 2035 commitment sits inside this horizon and implies a roughly 70-metre radius against a company whose flown hardware is a 14-tonne module; treat any date attached to it as intent. Speculative The competing trajectory is that nothing rotates, exercise countermeasures continue to fly, and the five gaps are still open in 2051 — which is what the last twenty-five years would predict.
• 50 yr: Speculative O'Neill-class habitats become plausible only if in-space manufacturing from lunar or asteroidal material matures, which is the architecture the original design assumed and the reason it has outlived proposals that assumed Earth launch. Speculative By this horizon the partial-gravity threshold question should be answered one way or the other, because a settlement programme cannot run for decades without answering it — but it will be answered by exposure rather than by trial if nobody builds the facility. Frontier That distinction matters: an answer obtained from a settled population is an epidemiological finding with no control arm.
• 100 / 250+ yr: Speculative Routine large-scale artificial-gravity settlement is a question of will and economics rather than physics, and that has been true since 1974. Handwave Any specific claim about what humans at partial gravity look like after generations is assertion: there is no multi-generational data on any vertebrate at any gravity level other than one, and the mouse studies run for a month. Handwave The one defensible statement at this horizon is that the design canon is fifty years old, dimensionally consistent, and has never been costed by anyone with the money.

• Depends on Nothing on this map, and that is the finding. This brief waits on no result another brief produces: the equivalence principle is measured to 2.7 × 10⁻¹⁵, rotation is textbook mechanics, and the human physiological problem is characterised with pooled numbers from about 189 astronauts. No typed depends-on edge is claimed, because there is no discovery to wait for — which is exactly what distinguishes this slot from Gravity Modification, Inertial Manipulation and Gravitational Wave Engineering, its neighbours in the same category, every one of which waits on physics nobody is producing.
• Requires (not on this map) The physics here is finished; what is missing is mass in orbit, a facility, and a number nobody has measured — and all three rows are descriptions of the historical record rather than forecasts. First, launch cheap enough for habitat-scale mass. Cost per kilogram to low Earth orbit has fallen roughly eighteen-fold, from about $54,000 on the Shuttle to about $3,000 on a reused Falcon 9, with sub-$200 figures still labelled design goals rather than achieved prices — and that eighteen-fold reduction has produced no rotating habitat, because the binding quantity is not the price but the tonnage: a structure at the published commercial design point is a launch problem in the hundreds of tonnes and an O'Neill cylinder is not a launch problem at any price. Cheap heavy launch is necessary and demonstrably not sufficient, which is why this row does not stand alone. Second, an orbital centrifuge sized for partial-gravity physiology. This one has been built. The ISS Centrifuge Accommodations Module was a 2.5-metre instrument spanning 0.01 g to 2 g and able to run two gravity levels at once, built by NASDA under a barter agreement worth approximately $700 million, stopped by a NASA stop-work order that redirected funds to the Crew Exploration Vehicle, cancelled in 2005, and manufactured anyway — the flight model stands on outdoor display at the Tsukuba Space Center. ESA's AGREE followed it into cancellation in 2013 after a stress analysis showed the ISS nodes would have been structurally compromised. NASA's own evidence report records the consequence: no human-rated centrifuge is available on board the station, and the largest radius ever flown is fifteen centimetres. Third, a decade-scale partial-gravity health programme. The agency named five gaps in 2015 — gravity level, gradient, rotation rate, exposure duration and combination with exercise — and none has closed; the only randomised trial of artificial gravity as a countermeasure ran for sixty days and missed its primary endpoint, with VO2max falling 20%, 23% and 24% across intermittent, continuous and control arms; and the physiological response to continuous Mars gravity is, in that same report's words, unknown. None of the three is a discovery. All three are a facility, a budget and a decade, and that is what separates this brief from every other slot in its category.
• Enables Any indefinite human presence beyond Earth rests on this: long transits, permanent orbital settlement, and the surface architectures that Deep Space Infrastructure and Space-Based Manufacturing assume. No typed enabling edge is claimed, for a specific reason rather than a cautious one: what those briefs need from this one is a threshold number — the minimum gravity that maintains human health — and that number does not exist, so nothing measurable is currently being enabled.
• Adjacent Space habitats and stations, orbital tethers, radiation shielding, closed-loop life support, and low-cost heavy launch supply the structure; space physiology, bone and muscle biology, vestibular science and human factors supply the requirements. Within this map: Space-Based Manufacturing and Deep Space Infrastructure for habitat-scale mass, Advanced Nuclear Propulsion and Fusion Spacecraft for the continuous-thrust variant, and Gravity Modification for every claim about switching gravity on.

Established “Rotating habitats solve the physiology problem.” The framing's error is a category error rather than an overclaim about physics. Rotation demonstrably produces weight; nobody disputes that. What the framing smuggles in is that producing weight is the same as producing health, and the evidence base — which is substantial here, unlike everything else in this category — says the mapping is not simple. Established The honest version is narrower and still strong: rotation is a solved way to produce apparent weight, and the open questions are how much, how fast, at what radius, and for which physiological system — every one of which is a measurement nobody has made in a human.

Established “Artificial gravity has been shown to prevent spaceflight deconditioning.” One randomised controlled trial exists, it ran sixty days, and it missed its primary endpoint: VO2max fell 20 ± 5%, 23 ± 5% and 24 ± 8% across intermittent artificial gravity, continuous artificial gravity and control, with no significant difference. Established The muscular benefit was real and partial — plantar flexion strength fell 35% and 31% against 48% in control — and the trial's own authors say artificial gravity would have to be combined with exercise before it could be considered an integral countermeasure. That is a negative result on the dose that was tested, not a vindication.

Established “Partial gravity gives you partial protection.” The one dataset says otherwise, and the mismatch is the most interesting number in the brief. At lunar gravity — one-sixth g — mice lost 28 to 30% of trabecular bone volume against 70% in microgravity. That is not one-sixth of the loss; it is well under half of it. Established Muscle went further: soleus mass and myofibre cross-sectional area were preserved outright at one-sixth g. Established And fibre type shifted anyway, where 1 g prevented it entirely. Protection is system-specific, not proportional, and any settlement plan that scales a health projection linearly with g is using an assumption the only available data contradict.

Frontier “Mars gravity is enough” and “Mars gravity is not enough” are both beyond the evidence. Declare the tie. The only quantitative human claim is a NASA Ames model putting the floor at 0.4 g — above Mars — with modelled losses of 0.39% per week at lunar and 0.22% at Mars gravity. Established The only in-orbit data are in mice and are more encouraging than the model. Established And NASA's own evidence report says the response to continuous Mars gravity is unknown. Averaging a model against a rodent produces a number with no meaning.

Established “Humans cannot tolerate high rotation rates.” The 0.1 rpm and 6 rpm criteria date from 1960 and 1973. Adaptation to head movements at 23 rpm is demonstrated, and 30.5 rpm was sustained daily for sixty days with 10 of 960 runs terminated. Frontier What is untested is living at those rates rather than dosing at them — and the gravity gradient, at 20 to 100% head-to-foot below ten metres of radius, is a separate constraint that has not moved at all.

Established “Exercise countermeasures are adequate, so the question is moot.” Meta-analysis says exercise “only partially protected against bone loss”, and the longitudinal study says nine of seventeen astronauts had incomplete bone mineral density recovery twelve months after return, with the full flown protocol in place and missions averaging 170 days. Frontier Exercise is what flies. That is a revealed institutional preference, and it is evidence about priorities rather than about efficacy.

Established “An orbital centrifuge is a hard engineering problem.” One was built. It was 2.5 metres across, spanned 0.01 g to 2 g, could run two gravity levels simultaneously, cost about $700 million in Japanese investment under a barter agreement, and was cancelled in 2005 when NASA redirected funds to a different vehicle. Established The flight model was manufactured and now stands on outdoor display at the Tsukuba Space Center. A second, ESA's AGREE, was cancelled in 2013 because the station nodes would have been structurally compromised. The obstacle is programmatic.

Established “Cheap launch will deliver it.” Cost per kilogram has fallen roughly eighteen-fold since the Shuttle and no rotating habitat has been started. Frontier Launch cost is necessary and not sufficient, which is precisely why this brief's adjudication carries three rows rather than one.

Speculative “A company will fly artificial gravity in 2035.” That is a published roadmap statement — 3.5 rpm, 950 cubic metres habitable, 40 crew — against a flown record of one 14-tonne, 45-cubic-metre module with four crew berths, a launch date that has already moved by a year, and no spin demonstration planned on the vehicle now being built. The 3.5 rpm figure implies about a 70-metre radius. Treat it as intent.

Speculative “Artificial gravity would fix the vision problem.” It might, and there is no evidence. The principal review of Spaceflight Associated Neuro-ocular Syndrome — the risk that most threatens long missions, affecting roughly 15% of astronauts at clinically significant severity and over 20% with choroidal folds — does not discuss artificial gravity as a countermeasure at all. The leading hypotheses concern fluid shift and compartmentalised cerebrospinal fluid, which a gravity gradient would plausibly affect. Plausibly is the whole of the claim.

Established “Astronauts float because they are beyond Earth's gravity.” Gravity at ISS altitude is about 90% of its surface value; weightlessness is continuous free fall, not any modification of gravity. Established And rotational artificial gravity is not a gravitational field — it is acceleration you feel as weight, indistinguishable locally because the equivalence principle holds to fifteen decimal places, and it has side effects the field variety would not: turn your head quickly in a small centrifuge and Coriolis forces make the room appear to lurch. Handwave Gravity plating, which switches a real field on, has no mechanism and belongs to Gravity Modification.

Established “The physics is contested.” It is not, and this is the one slot in the category where that sentence is literally true. Established The page should say so without letting it imply that the problem is finished — because the physics being finished is exactly what makes the remaining gaps embarrassing rather than excusable. Five named gaps, unclosed since 2015. One randomised trial, endpoint missed. Two cancelled centrifuges. One human demonstration, in 1966, at 0.00015 g.