Artificial Satellites for Climate Control

An orbital sunshade is a structure or swarm placed between the Sun and the Earth to intercept a fraction of the incoming solar flux. The framing under test is that orbital sunshades are a climate intervention option — that they belong on the menu of things a society could choose to do about warming.

This brief is the orbital hardware question only, and the scope line is worth stating before the first number. Climate Engineering owns solar radiation modification as a class of action: the physics of sulphate and sea-salt forcing, the regional-signal dispute, termination shock, moral hazard, consent, detection and attribution, unilateral deployment, and every governance instrument proposed for any of it. That brief's own boundary statement assigns this slot its territory in as many words — a space-based sunshade or L1 reflector swarm is owned here, and its presence in the menu is what makes “SRM” an ambiguous word. So: VI-01 asks whether deliberate intervention in the climate is a usable option for anybody; this page asks only whether this particular piece of hardware can be built.

What that leaves here is arithmetic: areal density, total mass, launch capacity, the displaced-equilibrium problem at L1, attitude control at gram scale, cost per watt per square metre as an engineering comparison, and the flown record. The answer is that the architecture does not close, and it does not close by three to four orders of magnitude — a margin large enough that no reasonable handling of the assumptions rescues it. The most striking feature of the evidence is that every load-bearing number comes from the proposals' own authors.

Established Angel’s 2006 PNAS paper is the reference design, and its numbers must be read from the paper rather than from summaries of it. The design intercepts 1.8% of the solar flux using a sunshade area of 4.7 × 10⁶ km², built from flyers with an areal density of 4.2 g/m² including structure and controls. Each flyer is 0.6 m across and masses about 1.2 g. Total mass: 20 million tons. And the count, verbatim from the paper: “Each 1,000-kg payload will contain 800,000 flyers. To transport the total sunshade mass of 20 million tons, a total of 20 million launches will be needed.” Established That is 16 trillion flyers and 20 million launches, on purpose-built electromagnetic launchers accelerating payloads to escape velocity followed by ion propulsion, at a target price of $50/kg and a capital cost of about $30 billion per launcher.

Established Angel’s own cost and schedule claims, quoted as claims. “A project total including development and operations of <$5 trillion seems also possible. If the 50-year lifetime is achieved, the cost per year averages to $100 billion”; and “It seems feasible that it could be developed and deployed in ≈25 years.” Speculative These are the proposer's figures for his own concept and they assume the $50/kg launch industry that the same paper requires be built. A widely circulated secondary account gives Angel's cost as in excess of US$130 billion over twenty years and the area as 3.8 million km²; both conflict with the paper, and the paper wins.

Established Now the comparison that settles the mass question. The ESA Annual Space Environment Report puts the total mass of human-made objects in Earth orbit as of 1 January 2025 at roughly 10,000 to 12,000 tonnes across 30,000–35,000 catalogued objects, and records 2025 as a record year with more than 300 launches, more than 4,000 new payloads and about 1,500–2,000 tonnes injected to low Earth orbit. Established Derived from those two published inputs — Angel's 20 million tonnes against ESA's measured totals, the division is this brief's — the sunshade is about 1,800 times the entire mass currently in Earth orbit, and at the record 2025 global rate it is about ten thousand years of world launch capacity, to a destination far harder to reach than LEO. Angel makes the same point differently by proposing 20 million dedicated launches; a secondary summary puts it as 550 years at 100 tonnes per day. The architecture does not require a bigger launch industry. It requires one that does not resemble the existing one in any respect.

Established An independent group redid the problem and got larger numbers. Sánchez and McInnes, in PLOS ONE in August 2015, target a 1.7% insolation reduction to offset a doubling of CO₂, which is 23.24 W/m² off a solar constant of 1,367 W/m². A classic single disc at L1 needs a 915 km radius, rising to 1,434 km once solar-radiation-pressure displacement is accounted for, at 2.44 × 10⁶ km from Earth. Mass depends entirely on the optical assumption: a perfect reflector needs about 240 million tonnes, a black non-reflective disc about 40 million tonnes, and Angel's near-transparent refracting design at about 4 g/m² about 14 million tonnes. Frontier An independent team reproduces Angel's order of magnitude and, under most optical assumptions, exceeds it by three to seventeen times.

Frontier The same paper contains a result more interesting than the mass figures: even a perfectly executed sunshade does not restore the previous climate. Their optimal two-disc configuration — radii of 1,200 km and 790 km, 6.5 × 10⁶ km² of shading area — puts about 40% of Earth's surface within ±0.1 °C of the pre-warming climate, against less than 10% for a static single disc. Established That is a physics result about a geometric mismatch: a point-source shade reduces insolation in a pattern that does not match the pattern of greenhouse forcing. What a residual regional error implies about who consents to it is Climate Engineering's question; the residual itself is this page's.

Established The dust alternative hits the same wall, and its own authors publish the number that shows it. Bromley, Khan and Kenyon, in PLOS Climate in February 2023, propose dust rather than structures and require about 10¹⁰ kg per year delivered near L1 for the same 1.8% attenuation — which they describe as “about 6 days per year of an obscured Sun” — and they state in their own paper that this is “700 times more mass than humans have launched into space.” Micron grains are optimal, and lunar regolith conveniently peaks near 0.2 micrometres, but radiation pressure disperses the cloud within about a week, so the entire 10¹⁰ kg must be resupplied continuously. Their own caveat is that the concept lacks controllability compared with rigid shades. Proponents publishing the fatal number about their own proposal — weight it heavily.

Frontier The design lives or dies on 4.2 g/m², and no hardware has ever come close. The comparison, with the divisions derived here from published masses and areas: NASA's ACS3, flown in 2024, is about 16 kg over roughly 80 m², which is about 200 g/m² — 48 times Angel's figure. Reflect Orbital's planned Eärendil-1 is 16 kg over an 18 m by 18 m sail, 324 m², about 49 g/m² — 12 times Angel's figure. Established The ACS3 inputs are from NASA's own engineering overview: about 16 kg total spacecraft, roughly 80 m² of sail, an aluminium / PEN / chromium laminate membrane, and four composite booms of 164 g each at 7.0 m, extensible to 14–16.5 m. The document gives no areal density; the division is this brief's.

Established The only orbital-mirror hardware anyone is actually building is being built to add sunlight, not to remove it. Reflect Orbital plans an 18 m by 18 m Mylar mirror at 16 kg, first launch mid-2026 on a Falcon 9 into a 600–650 km sun-synchronous orbit, delivering 0.1 lux over a 5 km spot — about full-moon brightness — with a planned constellation of 4,000 satellites. Funding totals $35.2 million: a $6.5 M seed in September 2024, a $20 M Series A in May 2025, and a $1.25 M AFWERX Phase II SBIR in June 2025. Its stated purpose is extending solar generating hours and after-dark lighting, and astronomers and dark-sky advocates already object to it on night-sky grounds. Frontier That is the real state of climate-control satellite hardware in 2026: the only constellation being funded has the opposite sign, its whole fleet is a rounding error against a sunshade, and it is already drawing scientific objection.

Established The historical precedent is smaller still, and it is thirty-three years old. Znamya 2, a 20 m mirror deployed on 4 February 1993, produced a 5 km bright spot of about full-moon luminosity tracking across Europe at 8 km/s and deorbited within hours. Znamya 2.5, at 25 m, deployed on 5 February 1999 and snagged on a Progress antenna and tore. Znamya 3, at 60–70 m, was never built and the programme was abandoned. The demonstrated capability of space mirrors is one full moon over one city, once.

Established Station-keeping at L1 is measured, not modelled, and the measurement does not transfer. NASA's study of the long-lived Sun–Earth L1 missions records that “LPO station-keeping delta-V costs grow exponentially with time elapsed from the last maneuver performed. For these missions, the doubling time constant is approximately 16 days.” SOHO spends about 1.0 m/s per year, averaging 0.09 m/s over 2004–2011 with manoeuvres roughly every 90 days; ACE about 1 m/s per year with burns every 103 days; WIND trends toward 1.0 m/s per year with an average burn of 0.17 m/s every 91 days. All three average four station-keeping manoeuvres a year, and remaining propellant as of mid-2011 was 129 kg, 58.1 kg and 57.8 kg respectively.

Frontier Apply that to a 4.2 g/m² object and the problem changes character entirely. Solar radiation pressure at 1 AU is about 4.56 micronewtons per square metre for full absorption, roughly double that for perfect reflection. Dividing the absorbed-case pressure by Angel's areal density gives an acceleration of about 1.1 × 10⁻³ m/s², which integrates to about 34 km/s per year of uncompensated velocity change — roughly 30,000 times SOHO's budget. This derivation is this brief's, from published pressure and areal-density inputs. Established It is not a station-keeping budget; it is a statement that a sail-density object cannot occupy the classical L1 point at all, and both primary designs say so and handle it: Angel displaces the cloud sunward to 1.85 Gm and holds it by modulating radiation pressure with tiltable mirrors over 2% of each flyer's area, while Sánchez and McInnes find a displaced equilibrium at 2.44 × 10⁶ km and fly forced orbits maintained by cone and clock angle changes with ±15,000 km of prescribed out-of-plane motion.

Frontier Which restates the real difficulty correctly: station-keeping is not solved by propellant, it is solved by making the shade itself the actuator. That means every one of 16 trillion elements needs working attitude control, at 1.2 grams, for fifty years, with no servicing. Speculative It is a reliability problem at trillion-unit scale, not a delta-v problem, and nobody has demonstrated attitude control at gram scale at all.

Established The cost comparison is where the concept loses, and it loses to the cheapest competitor rather than the best one. Smith and Wagner's reference estimate for stratospheric aerosol injection gives $36 billion over the first 15 years, about $2.25 billion a year, at roughly $1,400 per tonne of SO₂, reducing radiative forcing by 0.25 W/m² by the end of the first decade — about −0.2 K per decade, or −0.30 K by year 15 — using 0.1 Mt of sulphur in year one rising to 1.5 Mt by year 15, with a fleet growing from 8 aircraft to 95 and pre-deployment capital of about $3.6 billion.

Frontier The comparison, derived here with its steps shown. A 1.8% insolation reduction intercepts 24.6 W/m² at the disc, which as a global-mean radiative forcing is that figure divided by four and multiplied by one minus the albedo, giving about 4.3 W/m². Against that: aerosol injection at $36 bn for 0.25 W/m² is about $1.4 × 10¹¹ per W/m²; the sunshade at Angel's own $50/kg and under $5 trillion is about $1.2 × 10¹², roughly eight times worse; and at Falcon Heavy pricing of about $2,350/kg the launch bill alone is about $47 trillion, or $1.1 × 10¹³ per W/m², roughly 78 times worse. Established These are not like-for-like — the aerosol figure is an operating cost paid forever, the sunshade's is capital over a fifty-year life — and this page says so rather than pretending the ratio is clean. But the direction survives any reasonable handling: the sunshade is about an order of magnitude more expensive per watt at its own author's launch price, and about two orders more expensive at the launch price that exists.

Frontier An expert ranking of the same options reaches the same place by judgement rather than arithmetic. Pete Irvine's SRM technology tier list places space-based sunshades in D tier — theoretically unlimited cooling with no non-climatic side effects, defeated by needing “millions of tons of material into orbit” at “approximately $2,350 to get each Kg to Low Earth Orbit”; not F, because Starship-class lift might change the arithmetic, but “unlikely to be practical on a timescale relevant to the climate problem.” This is a solar-geoengineering researcher ranking his own field's options, so use it for relative ordering rather than as a computed result.

Frontier What is the true areal density floor? Angel's 4.2 g/m² is a design figure for an object that has never been built; ACS3 flew at about 200 g/m² and Reflect Orbital plans about 49 g/m². The open question is whether the gap of twelve to forty-eight times is a manufacturing problem, a functionality problem or a physics floor, and it is genuinely open because Angel's flyer includes attitude-control surfaces that no flown sail carries.

Speculative Can attitude control work at gram scale, at all? Every displaced-equilibrium architecture — Angel's tiltable mirrors over 2% of flyer area, Sánchez and McInnes's cone and clock angle control — makes the shade its own actuator. The demonstrated state of the art is a 16 kg spacecraft with four booms. Nothing in the retrieved literature demonstrates or costs attitude control on a 1.2 g object, and the reliability requirement is fifty years without servicing across a population of 16 trillion.

Frontier Is dust cheaper than structures? Bromley, Khan and Kenyon argue for lunar or cometary dust on cost grounds and then publish the requirement: 10¹⁰ kg per year, 700 times all mass ever launched, resupplied continuously because radiation pressure clears the cloud in about a week. Struck's earlier treatment of dust clouds at lunar Lagrange points exists and was only partially retrievable during research for this page. The live question is whether any dust architecture has a residence time long enough to matter, and its own proponents say the concept lacks controllability.

Frontier How much of the climate can any shade actually restore? The Sánchez and McInnes result — 40% of the surface within ±0.1 °C for an optimised two-disc system, under 10% for a single static disc — is a strong claim about a geometric limit and it has not been independently reproduced. If it holds, it means the best possible orbital shade leaves most of the planet outside the target, which is a physics finding rather than a policy one and belongs on this page.

Speculative Does cheap heavy lift change the verdict? This is the field's live conditional. Irvine declines to put sunshades in F tier for exactly this reason; a tertiary catalogue records Borgue and Hein's 2022 proposal at about 100,000 tonnes and 399–899 Starship launches a year. Handwave None of those figures was obtainable in primary form during research for this brief and none is asserted here. The arithmetic that can be checked is unforgiving: launch is 10 to 30% of mission cost, and even a fifty-fold reduction to Angel's $50/kg leaves the concept about eight times worse per watt than aerosols.

Frontier What is the residual-risk profile of a shade that fails partially? A swarm of 16 trillion independent elements does not fail all at once; it degrades. Nothing in the retrieved literature models partial deployment, partial attitude-control failure, or the climate consequence of a shade that is 60% functional. That is an unasked question rather than an answered one, and it is the kind of question that only gets asked once a concept is taken seriously enough to engineer.

Frontier And the question the assessment literature has quietly already answered. Parson and Keith's 2024 review of solar geoengineering cites McInnes once in its methods section and gives space-based approaches essentially no further treatment, naming stratospheric aerosol injection “the approach that is most prominent and promising.” Established The most consequential fact about space SRM in the assessment literature is how little of it there is.

Established Mass is the bottleneck and everything else is downstream of it. Twenty million tonnes against 10,000–12,000 tonnes currently in orbit and 1,500–2,000 tonnes a year of world launch capacity is not a scaling problem. It is a demand for an industry that does not exist, sized by its own proposer at 20 million launches on purpose-built electromagnetic launchers costing $30 billion each.

Frontier Areal density is the second bottleneck and it is a materials-and-function problem simultaneously. Getting to 4.2 g/m² is hard; getting there while carrying attitude control, thermal management and fifty-year survivability is the actual requirement, and the flown record sits at 200 g/m² for a sail that does none of those things. The membrane is not the hard part; the membrane plus the machine is.

Speculative Reliability at trillion-unit scale has no precedent in any industry. Sixteen trillion units at fifty years, unserviced, in a radiation environment, each individually responsible for its own station-keeping. Even a failure rate low enough to be unmeasurable in a terrestrial product line produces billions of dead flyers, and nothing in the retrieved literature states an acceptable failure rate or models the consequence of exceeding it.

Established The L1 instability is real and quantified, and it is not the bottleneck people expect. The 16-day doubling constant and roughly 1 m/s per year of operational spend at Sun–Earth L1 are measured facts from three long-lived missions. For a sail-density object the radiation-pressure acceleration alone is about 34 km/s per year, so the classical L1 point is unusable and both primary architectures abandon it. The residual bottleneck is actuator reliability, not propellant.

Frontier Cost per unit of forcing delivered. On the derivation set out above, the sunshade is roughly eight times worse per W/m² than stratospheric aerosols at its own author's launch price and roughly 78 times worse at real launch prices. A technology that is an order of magnitude more expensive than its cheapest competitor and several decades slower does not get selected, whatever its other merits.

Established And the institutional bottleneck, which is measurable. The National Academies' consensus study on solar geoengineering explicitly excluded space-based approaches from its scope; the field's standard review mentions them once; ARIA, the only funder to have put money into the question, allocated £400,000 out of £56.8 million — 0.7%. The concept is not being rejected on the evidence. It is not being assessed at all.

Established The launch dependency is the whole negative result restated. Solar Sail Systems owns the flight-proven hardware in the low-areal-density regime this concept requires, and it is the source of the only real datum in the comparison: ACS3 at about 200 g/m². The sunshade would be the largest single order any launch industry could ever receive, and it is 1,800 times the mass currently in orbit.

Established The materials and structures dependency runs to Planetary Scale Energy Systems — the 4.2 g/m² membrane, the twelve-to-forty-eight-fold gap to flown hardware, and gram-scale attitude control. Nothing about the shade is a climate-science problem; it is a thin-film and micro-actuator problem at a scale nobody has built.

Frontier The manufacturing dependency runs to Space-Based Manufacturing, and the numbers there are worth carrying across. That brief records that the one demand large enough to change its picture is assembly rather than production, citing a 5.9-million-kilogram space-solar satellite as the largest orbital-assembly job anyone has costed — belonging, in its words, to a customer that does not exist. A 20-million-tonne sunshade is about 3,400 times that job.

Frontier An independent reproduction of the residual-pattern result would matter more than another mass estimate. If the best achievable shade leaves 60% of the surface outside ±0.1 °C, that is a limit on what the technology could ever deliver even if built. It rests on one paper and one modelling framework.

Established Two dependencies on the literature itself, stated because they bound what this page can say. McInnes's 2010 engineering treatment of space-based geoengineering was obtained in abstract only, so its quantitative requirements are not quoted here. The concepts catalogued tertiarily — Early 1989, Benford 2004, Borgue and Hein 2022, the MIT space-bubbles proposal — were not obtained in primary form, so no figure from any of them appears on this page.

Established The experiment that has actually been run twice is a deployable mirror, and it is very small. Znamya 2 in 1993 produced a 5 km full-moon-brightness spot from a 20 m reflector and deorbited within hours; Znamya 2.5 in 1999 tore on a Progress antenna during deployment. Deployment is the failure mode with flight evidence behind it.

Established ACS3 is the modern datum and it is a sail rather than a shade. About 16 kg, roughly 80 m², an aluminium/PEN/chromium laminate on four 164 g composite booms extensible to 14–16.5 m. It establishes that a low-areal-density structure can be deployed and flown; it establishes nothing about holding an attitude against radiation pressure at 4.2 g/m².

Frontier Reflect Orbital's first launch in mid-2026 is the next real measurement, and it points the other way. An 18 m by 18 m mirror at 16 kg delivering 0.1 lux over a 5 km spot from a 600–650 km sun-synchronous orbit. It would be the first orbital mirror with a commercial purpose and the first hard measurement of large-membrane pointing accuracy from orbit — both directly relevant here, even though its product is added sunlight.

Speculative The decisive experiment nobody has proposed is a gram-scale free-flyer with working attitude control. One flyer, 1.2 g, 0.6 m across, holding a commanded orientation against solar radiation pressure for a year. It is small, cheap and would test the single assumption on which both primary architectures rest. Its absence from the literature is more informative than any of the mass estimates.

Frontier A displaced-equilibrium station-keeping demonstration is the second one. Holding position sunward of L1 by modulating radiation pressure rather than by burning propellant has never been flown. The operational L1 record — ACE, SOHO and WIND, with a measured 16-day doubling constant and four manoeuvres a year — is about propellant-controlled spacecraft and does not transfer.

Frontier And an experiment that requires no hardware: reproducing the Sánchez and McInnes residual-pattern computation in an independent climate model. If the 40%-within-±0.1 °C result is robust, it constrains every orbital-shade proposal ever made, and it can be checked with existing codes on existing computers.

Established Requirement one: 4.2 g/m² including structure and controls. Best flown: about 200 g/m². Best funded: about 49 g/m². The gap is a factor of twelve to forty-eight, and the flown articles carry no attitude-control surfaces, so the true gap is larger than the ratio suggests.

Established Requirement two: 20 million tonnes delivered beyond LEO. Angel's own answer is 20 million launches on electromagnetic launchers at $30 billion each and $50/kg. This is not a payload requirement; it is a specification for a new industrial civilisation, written into the paper as an assumption.

Speculative Requirement three: gram-scale attitude control, 16 trillion units, fifty years, unserviced. Angel allocates 2% of each flyer's area to tiltable mirrors. No mechanism at that mass scale has been built, characterised for radiation tolerance, or life-tested, and the whole station-keeping solution depends on it working almost everywhere almost always.

Frontier Requirement four: a displaced equilibrium that stays put. Angel places the cloud 1.85 Gm sunward of Earth; Sánchez and McInnes find their own displaced point at 2.44 × 10⁶ km and prescribe ±15,000 km of controlled out-of-plane motion. Both are solutions to the same fact: at 4.2 g/m² radiation pressure produces about 34 km/s per year of acceleration, so classical L1 is not available.

Frontier Requirement five: manufacturing throughput. Sixteen trillion units in twenty-five years is about 2 × 10¹⁰ flyers per year, or roughly 600 per second, continuously, for a quarter of a century — a derivation from Angel's own count and schedule. No production line in any industry runs at that rate for any product.

Established Requirement six, which the concept passes: the radiative physics. Reducing insolation by 1.7–1.8% does offset a doubling of CO₂ in the global mean, and Angel, Sánchez and McInnes, and Bromley et al. agree on it. The physics of the intervention is the one part of this subject that is settled, which is precisely why the engineering deserves this much scrutiny.

Established Climate Engineering owns the whole terrestrial question and this page must not re-run it. Termination shock, moral hazard, unilateral deployment, ocean acidification, the non-use argument, governance instruments and the regional-precipitation dispute are all VI-01's, and each gets a clause and a pointer here at most. Frontier The one place the two genuinely touch is the residual-pattern result — that even an optimal shade leaves most of the surface outside ±0.1 °C. That is physics and it belongs here; what it implies about consent belongs there. Established And one reciprocal correction: VI-01 should not be where a reader learns the sunshade's mass, and the sailcraft count carried in both briefs was wrong by four orders of magnitude and is corrected to 16 trillion in both.

Established Solar Sail Systems is the adjacent field with real hardware and the only source of measured areal densities in the relevant regime. Every architecture here is a solar sail with a different job: instead of using radiation pressure to travel, it uses radiation pressure to stay put while blocking light. The dynamics are the same dynamics.

Frontier Planetary Scale Energy Systems owns the materials-and-structures side — thin films, deployable membranes, and the question of what a structure of continental area is made of. The sunshade is 4.7 million square kilometres of engineered surface, which is about the area of the European Union, assembled from parts weighing 1.2 grams.

Frontier Space-Based Manufacturing owns whether any of it could be produced or assembled off Earth, which is the only route that avoids 20 million launches. Its own finding is that the largest orbital-assembly job anyone has costed is 5.9 million kilograms, for a customer that does not exist. The sunshade is 3,400 times that job, so the manufacturing route does not currently rescue the launch route.

Speculative Terraforming shares the toolkit at a different planet. One of the routes surveyed in the 2026 Martian systems analysis requires 10¹³–10¹⁴ m² of orbital reflector area at Mars, which is two to twenty times the area of Angel's terrestrial sunshade. The mass arithmetic is not repeated there; it is cited across from here.

Established The single most informative institutional fact is a budget line. ARIA's Exploring Climate Cooling programme committed £56.8 million in May 2025, plus £11 M from NERC in April 2025, and announced £45 million across 21 projects on 8 May 2025 in four categories: governance and ethics (5), modelling (7), outdoor monitoring (4), and controlled small-scale outdoor experiments (5). The five outdoor experiments are Arctic sea-ice re-thickening, marine cloud brightening, controlled electric discharge, milligram-quantity mineral-dust injection, and cirrus monitoring. Established The only space project in the programme is a £400,000, 15-month Space Reflector Baseline Survey led by Morgan Goodwin of the Planetary Sunshade Foundation.

Established Both halves of that sentence matter and both have been mis-stated in circulation. It is a baseline survey, not an engineering feasibility study; and its lead is an advocacy organisation that exists to promote planetary sunshades — an interested party surveying the technology it advocates. And £400,000 out of £56.8 million is 0.7% of the programme. Frontier The UK's flagship climate-cooling research programme allocated less than one per cent to the space option and gave that fraction to an advocate for a scoping exercise. That is the most informative number in this section, and it is an institutional judgement expressed as an allocation.

Established The National Academies report cannot be cited in support of this technology, because it excluded it. Reflecting Sunlight (2021) acknowledges in Box 1.1 that other strategies exist, including reflecting sunlight above the atmosphere, and then states: “This study was designed to focus specifically on the set of atmospheric-based strategies described below” — stratospheric aerosol injection, marine cloud brightening and cirrus cloud thinning. It offers no cost estimate, no timeline, no feasibility assessment and no research recommendation for orbital sunshades. Established Any brief that cites it as governance framing for climate-control satellites is mis-stating what the source covers, and this page previously did exactly that.

Frontier The same absence runs through the field's standard review. Parson and Keith's 2024 Annual Review chapter cites McInnes once in methods and otherwise leaves space-based approaches alone, while naming stratospheric aerosol injection the most prominent and promising approach. Three independent institutional signals — a consensus study that excluded it, a review that skips it, a funder that gave it 0.7% — all point the same way.

Frontier What an institution would need in order to take this seriously. Not governance machinery, which VI-01 covers, but an engineering assessment: an independently reviewed areal-density roadmap, a gram-scale attitude-control demonstration, and a production-rate analysis. None of the three has been commissioned by anyone, and the only money spent to date bought a survey from a foundation that already knows what it thinks.

Frontier Meanwhile the actual regulatory question in orbit is arriving from a different direction. Reflect Orbital's 4,000-satellite plan for delivering sunlight after dark is already drawing objection from astronomers and dark-sky advocates. The first real institutional fight about orbital mirrors will be about night-sky brightness and licensing, not about climate, and it will set precedents that any later climate proposal inherits.

Established Most of the ethics of solar radiation modification belongs to Climate Engineering and is treated there at length. Termination shock, moral hazard, consent among affected populations, unilateral deployment and the non-use argument are all VI-01's, and repeating them here would be duplication rather than coverage. What remains genuinely specific to this page is narrow, and it is worth stating precisely because it is narrow.

Frontier The residual-pattern result is an ethical fact as well as a physical one. If the best available shade leaves about 60% of the surface outside ±0.1 °C of the pre-warming climate, then even a perfectly executed intervention distributes its benefit unevenly by geometry. The physics generates the distributional problem before any political choice is made, which is a different situation from an intervention whose unevenness is a matter of implementation.

Frontier Interest disclosure is the live ethical issue in this literature and it is unusually tractable. The reference design comes from its originator, the dust proposal from its proponents, the tier ranking from a researcher ranking his own field, and the only funded study from a foundation constituted to advocate the technology. Established In each case the authors have published the numbers that hurt their own case — 20 million launches, 700 times all mass ever launched, D tier — which is the strongest evidence available anywhere on this page and is why it is relied on.

Speculative Opportunity cost is a real argument at this scale. Under $5 trillion of capital over twenty-five years, on the proposer's own optimistic figure, is a fraction of world output committed to a single irreversible artefact. This brief does not adjudicate that trade-off; it notes that the comparison at the same forcing is roughly an order of magnitude cheaper by aerosol injection, and that the comparison is what an institution actually decides on.

Frontier And an ethical consequence specific to hardware in orbit. Sixteen trillion objects near L1 is not a debris problem in the usual sense — it is far from Earth — but 4,000 mirrors in low orbit already is. The astronomical objection to Reflect Orbital is the first real instance of the trade-off between using orbital space for a terrestrial benefit and preserving the sky as a common scientific resource, and it is arriving years before any climate proposal.

Established The civilisational reading of this concept is that it is the largest artefact ever seriously proposed, and its scale is best understood by comparison. Twenty million tonnes is 1,800 times everything in Earth orbit; 4.7 million square kilometres is roughly the land area of the European Union; 16 trillion units is about two thousand objects for every human alive. Nothing else in the engineering literature is this large.

Frontier Which means the concept is really a claim about industrial capacity rather than about climate. Angel's architecture requires electromagnetic launchers at $30 billion each and a $50/kg price, twenty million launches, and a manufacturing line producing roughly 600 flyers a second for twenty-five years. A civilisation that could do that would have solved problems that make the climate problem look small, which is the structural objection to the whole family and it does not depend on any particular number being right.

Speculative The reversibility argument is the strongest civilisational point in the concept's favour. Unlike an atmospheric intervention, an orbital shade can in principle be switched off by tilting it, and it puts nothing into the biosphere. It is the least atmosphere-invasive option on the menu and the least timely, and those two properties are causally connected: what makes it clean is that it is outside the system, and what makes it slow is the same thing.

Frontier The fifty-year lifetime is a governance claim disguised as an engineering parameter. Angel's cost-per-year figure divides by a fifty-year life, and the whole comparison with aerosols turns on capital being amortised over half a century of continuous, maintained operation of a trillion-unit swarm. No institution has maintained a technical system at constant purpose for fifty years without interruption, and the assumption is never stated as one.

Frontier And the civilisational fact that this page exists to record. The idea that we could move the Sun's light is one of the oldest engineering fantasies and one of the few with a fully worked reference design in a peer-reviewed venue. The design was worked out honestly, its author published the numbers, and the numbers say no by three to four orders of magnitude. That is a good outcome for the literature and a poor one for the concept.

These horizons track flown hardware and funded programmes, because there is no sunshade programme anywhere to track:

• 10 yr: Established Reflect Orbital's first mirror flies in mid-2026 and its constellation, if built, produces the first real data on large-membrane pointing from orbit — for adding sunlight, not removing it. Established ARIA's £400k Space Reflector Baseline Survey runs fifteen months and reports inside this horizon; it is a survey, and its lead is an advocacy foundation. Frontier Expect the first serious regulatory fight about orbital mirrors to be about night-sky brightness rather than climate. Speculative No sunshade element, at any scale, is scheduled to fly.
• 25 yr: Speculative This is the horizon Angel assigned to full deployment in 2006, on the assumption of a $50/kg electromagnetic launch industry built from scratch. Handwave Nothing in the intervening twenty years has moved toward that assumption: launch is $1,400–2,700/kg, total orbital mass is 10,000–12,000 tonnes, and the best flown areal density is forty-eight times too heavy. Speculative The plausible milestone at this range is a gram-scale attitude-control demonstration, which nobody has proposed.
• 50 yr: Speculative If heavy lift reaches Starship-class cadence and price, the launch term in the cost comparison improves and the concept moves from about 78 times worse per watt than aerosols toward about 8 times worse — which is its own author's best case, not a breakthrough. Handwave Any date for deployment at this horizon is an assertion; the binding terms are areal density and trillion-unit reliability, and neither has a development programme.
• 100 / 250+ yr: Handwave Beyond forecasting. Speculative The defensible structural statement is that the concept's requirements are set by the ratio of solar radiation pressure to areal density and by the mass of a shade large enough to matter — both physical quantities that no technology alters — so its horizon depends on the emergence of an industrial capacity in space rather than on progress in sunshade design.

• Depends on Three edges, and all three are about building and lifting rather than about climate. Solar Sail Systems supplies the only flight-proven hardware in the low-areal-density regime and the only measured number in the comparison: ACS3 at about 200 g/m², forty-eight times Angel's 4.2. Planetary Scale Energy Systems owns the thin-film and structures question at continental area. Space-Based Manufacturing owns the only route that avoids twenty million launches — and its own finding is that the largest orbital-assembly job anyone has costed is 5.9 million kilograms, for a customer that does not exist, which makes a 20-million-tonne sunshade about 3,400 times that job. None of these is a climate dependency. The climate physics is settled; the hardware is what fails.
• Requires (not on this map) Three constraints that are not briefs on this map, and each is derived from the reference design's own published figures. The launch constraint is 20 million tonnes to a destination beyond LEO, against 10,000–12,000 tonnes currently in orbit and 1,500–2,000 tonnes a year of world capacity — about 1,800 times the standing total and roughly ten thousand years of production. The manufacturing constraint follows from Angel's own count and schedule: 16 trillion flyers in about twenty-five years is roughly 600 units per second, continuously. The scientific constraint is the smallest and the most damaging: both primary architectures abandon the classical L1 point because radiation pressure on a 4.2 g/m² object produces about 34 km/s per year of acceleration, and both solve it by making each flyer its own actuator — a mechanism nobody has demonstrated at gram scale even once.
• Enables No enabling edge is claimed. A working orbital shade would be an instrument of Climate Engineering — but there is no hardware, no programme, no funded engineering study, and the only money ever allocated to the question was 0.7% of one national programme, for a baseline survey, to an advocacy foundation. An edge from that to anything would record an aspiration rather than a dependency.
• Adjacent Climate Engineering is the neighbour that owns the entire terrestrial question and is cited here only for the seam and for the aerosol cost comparison. Terraforming shares the toolkit at Mars, where one surveyed warming route requires 10¹³–10¹⁴ m² of orbital reflector — two to twenty times the area proposed here — and cites this page's arithmetic rather than repeating it.

Established “About 1.5 billion sailcraft.” This page said that, and it was wrong by roughly four orders of magnitude. Angel's paper says 16 trillion flyers — 800,000 per 1,000 kg payload, 20 million tonnes, 20 million launches, stated in those words in the primary source. Frontier The error made the concept sound about ten thousand times more tractable than its own author claims, and it circulated in this corpus in two places. It is corrected here and in Climate Engineering. Recording the correction rather than quietly fixing it is the point: a number that wrong survived because everybody was quoting summaries instead of the paper.

Established “The National Academies assessed space sunshades and set out how they should be governed.” It did not. Reflecting Sunlight (2021) states that the study “was designed to focus specifically on the set of atmospheric-based strategies described below” — stratospheric aerosol injection, marine cloud brightening, cirrus cloud thinning — and Box 1.1 explicitly acknowledges reflecting sunlight above the atmosphere as something it is not covering. Established There is no cost estimate, no timeline, no feasibility assessment and no research recommendation for orbital sunshades anywhere in it. This page previously cited it in support of the technology; that citation described the source incorrectly, and the report cannot be used that way by anyone.

Established “ARIA funded a space-based solar-reflector feasibility project.” ARIA funded a £400,000, 15-month Space Reflector Baseline Survey — a survey, not an engineering feasibility study — led by Morgan Goodwin of the Planetary Sunshade Foundation, an advocacy organisation constituted to promote the technology being surveyed. Frontier And the number that matters is the fraction: £400,000 out of £56.8 million is 0.7% of the programme. Describing it as serious feasibility study inverts the signal — the allocation is the clearest institutional verdict available, and the verdict is close to no.

Frontier “Station-keeping at L1 is a modest propulsion budget — SOHO manages on 1 m/s a year.” SOHO does, and the figure does not transfer. Established A 4.2 g/m² object experiences about 1.1 × 10⁻³ m/s² from radiation pressure alone, integrating to roughly 34 km/s per year — about thirty thousand times SOHO's budget, derived here from published radiation pressure and the design's own areal density. Which is why both primary architectures abandon classical L1 for a displaced equilibrium and hold station by tilting. The residual difficulty is the reliability of trillions of actuators, not delta-v.

Frontier “Cheaper launch closes the gap.” At Falcon Heavy pricing the launch bill alone is about $47 trillion, and launch is only 10 to 30% of mission cost. Established Even a fifty-fold price reduction to Angel's own $50/kg leaves the concept about eight times more expensive per watt of forcing than stratospheric aerosols, on the derivation set out in section 2. Cheap launch improves the concept from hopeless to uncompetitive.

Speculative “Dust is the cheap version.” Its own authors put the requirement at 10¹⁰ kg per year — 700 times more mass than humans have launched into space — resupplied continuously because radiation pressure disperses the cloud in about a week, and they note the concept lacks controllability compared with rigid shades. Frontier The dust route replaces a manufacturing problem with a throughput problem and makes the total worse.

Handwave “There are lots of other designs and some of them are cheaper.” There are: Early's 1989 disc, Benford's 2004 Fresnel lens, Borgue and Hein's 2022 estimate, the MIT space-bubbles concept. Speculative Every one of them is known to this brief only through a tertiary catalogue, none was obtained in primary form, and therefore no figure from any of them appears on this page. A reader who encounters a low number attached to one of these names should ask which paper it comes from, because this brief could not get any of them.

Frontier “Reflect Orbital is building climate-control satellites.” Its stated purpose is the opposite sign: delivering more sunlight to the ground, for solar generation hours and after-dark lighting, at 0.1 lux over a 5 km spot. Established Its entire planned 4,000-satellite constellation, at 16 kg each, is 64 tonnes — about three parts in a million of a sunshade.

Frontier “A sunshade would restore the climate we had.” On the only optimisation study available, the best two-disc configuration puts about 40% of the surface within ±0.1 °C of the pre-warming state and a single static disc under 10%. Speculative The mismatch is geometric — a point-source shade does not reproduce the spatial pattern of greenhouse forcing — so it is not fixed by building the shade better. What the residual means politically is Climate Engineering's question.

Speculative “Space sunshades are the safe option because nothing enters the atmosphere.” That much is true and it is the concept's real merit: no aerosol chemistry, no ocean acidification interaction, in-principle reversibility by tilting. Frontier But the same property that makes it clean makes it slow, and the institutional record shows the field has already priced that trade-off: excluded from the National Academies' scope, one citation in the standard review, 0.7% of the only funder's budget.

Established And the framing itself. “An intervention option” describes something on a menu. Frontier The L1 architecture does not close by three to four orders of magnitude on mass, by twelve to forty-eight times on areal density, and by roughly one to two orders of magnitude on cost per watt — and every one of those numbers comes from the proposals' own authors. Established Meanwhile the institutions that assess climate intervention have already quietly reached the same verdict. This is not on anybody's menu.