Asteroid Mining

Asteroid mining means extracting material from a small solar-system body and using or selling it. The framing under test on this page is that asteroid resources are an economic proposition — that somewhere in the near-Earth population there is a rock whose contents are worth more than the cost of going to get them.

The spine of this brief is a distinction the phrase “asteroid mining mission” erases: sample return versus mining. Sample return is a solved capability, demonstrated twice, with masses and costs on the public record. Mining — extracting a resource for use or sale — has never been attempted by any spacecraft. Everything else here follows from keeping those two apart: the engineering record is genuinely impressive, the commercial record is two liquidations and two lost spacecraft, and conflating them is how a subject with 127 grams of returned material acquires a reputation for imminence.

The page divides its territory explicitly. This brief owns how you get the material — targets, accessibility, extraction technique, the firm-level engineering and capital record. Space Resource Economies owns whether anyone buys it, and the same failure record appears there as evidence about market structure rather than about engineering. Space Law and Governance owns whether a mined kilogram can lawfully be owned and sold; that question is named here in one clause and not re-argued. Space-Based Manufacturing and Deep Space Infrastructure own the consumer and the pipe.

Established Two asteroid samples have been returned to Earth, and their masses are measured rather than estimated. NASA's OSIRIS-REx delivered 121.6 g of asteroid Bennu, a figure announced on 15 February 2024 after the TAGSAM sampling head was fully disassembled that January — 70.3 g from the canister and mylar flap, 51.2 g from the final pour, against a mission requirement of 60 g. The capsule landed 24 September 2023. It is the largest asteroid sample ever collected in space. JAXA's Hayabusa2 returned 5.424 ± 0.217 g of Ryugu including gas, landing at Woomera on 6 December 2020, against a minimum requirement of 0.1 g — a factor of 54 over requirement. The Ryugu material came from chambers A, B and C, including sub-surface grains excavated by the SCI impactor on 5 April 2019, which produced a crater about ten metres across.

Established Now put the costs beside the masses. OSIRIS-REx's lifecycle cost, on the Planetary Society's independent accounting, was $1.02 billion unadjusted. Hayabusa2 cost 16.4 billion yen, about US$149 million on the 2010 estimate, for a spacecraft of 600 kg wet and 490 kg dry. Established Derived from those two pairs — the division is this brief's, both inputs are quoted above — the delivered cost of asteroid material is about $8.4 million per gram for OSIRIS-REx and about $27,500 per gram for Hayabusa2. Frontier The 56-fold spread between two successful missions is itself the finding. The cost of asteroid material on Earth is not a physical constant; it is an architecture choice. And both numbers sit five to nine orders of magnitude above any commodity price on this planet.

Established Nothing has ever been mined. No spacecraft has extracted a resource from an asteroid for use or for sale. Every kilogram figure in the commercial literature is a projection, and the distinction between 127 grams returned for science and a tonne extracted for a customer is the whole subject.

Established The industry's record is two funded failures and one firm still trying. Planetary Resources was founded in 2009 as Arkyd Astronautics and publicly launched on 24 April 2012. It raised a $21.1 million Series A in May 2016, a €25 million capital-and-grant package from Luxembourg in November 2016, an undisclosed 3D Systems investment in June 2013, and took Bechtel as investor and partner in April 2013; a 2013 Kickstarter for a public space telescope was terminated in May 2016 with refunds promised. Contemporary reporting puts the total at roughly $50 million by 2016, including money from Eric Schmidt and James Cameron. Established What it actually flew was three CubeSats and no prospector. Arkyd-3 was destroyed in the Antares launch failure of 28 October 2014. Arkyd-3 Reflight, a 3U CubeSat, launched on CRS-6 on 17 April 2015, deployed from the ISS on 16 July 2015 and reentered on 23 December 2015 — an avionics, attitude-control and propulsion testbed. Arkyd-6, a 6U CubeSat, launched 12 January 2018 on PSLV-C40 and decayed 28 February 2024. Arkyd-100, -200 and -300 were never built, and no prospecting mission ever left Earth orbit. An expected round from an unnamed mining company failed to close in early 2018; the human assets were bought by ConsenSys on 31 October 2018, the intellectual property was released to the public domain in May 2020 and the remaining hardware was auctioned that June.

Established Deep Space Industries raised roughly $3.5 million in private funding plus government contracts before being acquired by Bradford Space, which redirected it to green propulsion. Chad Anderson's contemporaneous summary is the shortest available: “They’re gone; they’re done. They don’t exist.” Established The diagnosis that carries most weight comes from inside the best-funded firm. Peter Marquez, Planetary Resources' policy director, said the business model “wasn’t working out the way we’d hoped” and that there is “no customer base for asteroid mining in the next 12 to 15 years.” That is an executive of the company conceding the market does not exist — interest running hard against the statement, and therefore weighted heavily here. He said it in 2019; the window closes in 2031–2034.

Established AstroForge, founded 10 January 2022, is the successor attempt, and it has raised more and flown worse. Its funding is a $13 million seed in 2022 plus a $40 million Series A in 2024, about $53 million — a figure that rests on a single secondary source and should be read as approximate. Two missions, two losses. Brokkr-1, a 6U CubeSat built by OrbAstro, launched 15 April 2023 on Transporter-7 to demonstrate in-space refining, and failed on communications when magnetic interference from the refining payload disrupted attitude and antenna pointing; the company had identified the magnetic-field risk before launch and flew anyway to avoid a nine-month delay. Odin, a roughly 100 kg bus also from OrbAstro costing about $3.5 million, launched 26–27 February 2025 as an IM-2 rideshare toward 2022 OB5 with a flyby planned about 301 days out; contact was never reliably established and it was declared lost on 6 March 2025.

Established AstroForge's own debrief is the most useful single document in this record, because it is a company publishing evidence against its own competence. It states a build time of under ten months and a cost of about $3.5 million against NASA's Lunar Trailblazer at roughly $95 million, and a self-assigned 30% probability of success. Verbatim: “We went into the mission aware of several specific issues” — a power-amplifier configuration and possible solar-panel deployment problems, classified as known risks — and “Aerospace and venture capital success are all predicated on managing risk, not eliminating it. We chose to iterate quickly to maximize our learning rate.” Ground-station failures were near-universal: wrong configurations, mispointing, amplifier losses, local interference. Interest strongly against the finding; weight accordingly. Frontier Its successor, DeepSpace-2, is built — about 200 kg, electric propulsion, 1.7 kW against Odin's 180–190 W, landing legs for surface measurement of platinum-group metals, manifested on Intuitive Machines' IM-3. It is a measurement mission, and the company does not claim otherwise.

Established One operator's stated preference is a fact about the market and belongs in the record. AstroForge says it is “not interested” in water ice or propellant depots, citing absence of demand. An operator declining the more defensible product because it cannot find a buyer is evidence about the customer, not about the chemistry.

Established Government programmes failed differently, and the difference matters. NASA's Asteroid Redirect Mission proposed to retrieve a roughly four-metre boulder from a near-Earth asteroid and move it to lunar orbit for about $2.6 billion, with $105 million spent in 2014 on concept maturation. Proposed in 2013, its launch slipped from 2017 to 2020 to December 2021; it was defunded in April 2017 and formally closed out on 13 June 2017 in favour of what became Artemis. It would have demonstrated high-power solar-electric propulsion, robotic capture with anchoring grippers, gravity-tractor deflection and laser communications, and it demonstrated none of them. Frontier ARM was not cancelled because asteroid retrieval failed a technical test; it was cancelled by a change of budget priority. The general pattern is worth stating plainly: capability programmes die of politics, commercial ones die of customers.

Frontier China's Tianwen-2 is the live sample-return attempt. It launched 28 May 2025 on a Long March 3B, entered orbit around the quasi-satellite 469219 Kamoaalewa on 7 June 2026, and began close approaches from 4 July 2026 at about 20 km, descending through 3 km, 600 m and 300 m observation altitudes, with sampling by anchor-and-attach and touch-and-go and explosives used to expose sub-surface volatiles. Earth return is scheduled for 2027, with an extended mission to the main-belt comet 311P/PanSTARRS arriving January 2035. An agency reporting on its own mission — treat success claims as interested until the capsule is on the ground.

Established The most-quoted number in the subject is $10,000 quadrillion, and it traces to the mission's own principal investigator. NASA's Psyche launched 13 October 2023 and arrives in August 2029 for roughly two years in orbit. NASA's own mission page describes it as “the first mission to explore an asteroid with a surface that contains substantial amounts of metal rather than rock or ice,” possibly “the partial core of a planetesimal,” and contains no mention of mining, resources or monetary value. The valuation figure comes from Lindy Elkins-Tanton, the mission's PI, as an iron-content thought experiment: if Psyche's metal were on Earth it would be worth more than the world economy. An interested party on her own mission, and the figure has propagated for a decade with its framing stripped off.

Established The named rebuttal is short. Philip Metzger of the University of Central Florida: “Even if we could mine Psyche and bring the materials back to Earth, the asteroid is so far away that the cost of doing so would negate their value.” Frontier And the geological premise has since weakened. Cantillo and Reddy at the University of Arizona put Psyche at about 82.5% metal, 7% low-iron pyroxene, 10.5% carbonaceous chondrite, with roughly 35% bulk porosity, against earlier estimates running to 95% metallic — a reading closer to a rubble pile than an intact planetary core. Cantillo: “Psyche as a rubble pile would be very unexpected, but our data continues to show low-density estimates despite its high metallic content.” Three moves defeat the headline: the number came from an interested party, it prices a stock rather than a flow, and the object is probably not what the number assumed.

Established The platinum argument defeats itself, and the mechanism is elasticity. Elvis's own worked case takes a 100 m metallic asteroid at a conservative 10 ppm total platinum-group metal content, giving about 23.6 tonnes of PGM, which he prices at US$1.18 billion at then-current prices. The step always omitted is what happens to “then-current prices” when 23.6 tonnes arrive at once into an industrial demand curve that is inelastic in the short run. The price used to compute the valuation is a price that exists only in the absence of the delivery. Frontier The direction of that argument is secure; its magnitude is not asserted here. USGS Mineral Commodity Summaries returned HTTP 403 on three attempts during the research for this brief, so world PGM production and reserve figures were not obtained, and this page therefore puts no percentage on the price collapse. A brief that asserted one from memory would be doing the thing it criticises.

Frontier How rare a genuinely mineable target would be, on the field's own formalism. Elvis decomposes the ore probability for platinum-group metals as the product of three terms: 4% of near-Earth objects are nickel-iron, about 50% of iron meteorites beat terrestrial iridium ore grades of 0.3–0.9 ppm, and 2.5% sit within 4.5 km/s of low Earth orbit. The product is 5 × 10⁻⁴, about one NEO in two thousand. Against roughly 20,000 known NEOs above 100 m, that implies of order ten PGM ore-bearing near-Earth objects in existence. Relaxing the accessibility criterion to 5.7 km/s raises that term to 25% and the count to about a hundred — the accessibility term dominates everything else in the calculation.

Frontier The same formalism is far kinder to water, and the asymmetry is the most useful structural fact in the subject. For water Elvis gives 10% of the population, 25–31% rich enough, and 3% accessible, for an ore probability of 9 × 10⁻⁴. At an assumed 20% water content — 10% hydrated minerals plus 10% ice — the minimum profitable diameter falls to 18 m, about 236 tonnes of water, which he prices at US$1.2 billion delivered to low Earth orbit. The population is about 9,000 such objects at 18 m or larger, of which only about 18 exceed 100 m and are findable with present surveys; his own caveat is that the small ones “are hard to find with present surveys,” needing detection at around 0.04 AU. Frontier So water needs a smaller rock and there are a thousand times more of them — and its buyer is in orbit and does not exist, while platinum has a buyer on Earth whose price collapses on delivery. Neither case closes, and they fail for opposite reasons.

Established Sonter wrote the conditions down in 1997 and they are still the sharpest sentence in the literature. His conclusion was that robotic near-Earth asteroid resource recovery “is technically feasible in the near term, and … the returned product can potentially be highly profitable, given an in-space market of some thousands of tonnes per year, in competition against Earth-launch costs of several hundred dollars per kilogram.” Both are stated as conditions, not predictions. Established Twenty-nine years on, neither holds. There is no in-space market of thousands of tonnes a year of anything; the confirmed in-space market for extracted propellant is zero tonnes. Launch is $1.4k–2.7k per kilogram for Falcon Heavy and Falcon 9 on the NASA Ames cost baseline — thousands of dollars, not several hundred. The most careful economic case for asteroid mining ever written named the two conditions under which it works, and neither has arrived in three decades.

Frontier Is the accessible population small or large? The literature contains two counts that look contradictory and are not. Elvis, McDowell, Hoffman and Binzel define ultra-low-delta-v objects as those under 4.5 km/s from LEO, and found that as of March 2010 only 65 of 6,699 known NEOs — 0.97% — qualified, that those which did were “small and hard to recover,” and that transit time, launch windows, abort options and proximity-operations safety cut the set further. Abell and colleagues, writing for EPSC-DPS in 2025, report 38,812 near-Earth asteroids catalogued as of 1 August 2025, of which 6,294 — about 16% — meet NHATS accessibility criteria. Established The two use different gates: Elvis's is 4.5 km/s one-way, NHATS uses round-trip thresholds of 4 to 12 km/s. The honest synthesis is that the accessible population is small under a strict criterion and large under a loose one, that which criterion applies is a mission-design question rather than an astronomy question, and that prospecting has never distinguished the two — which is precisely why nobody knows how many mineable objects there are.

Frontier What would a survey change? Abell et al. project that a dedicated survey finds 200,000–300,000 new NEOs, reaching better than 25% completeness in five years and better than 50% in ten for sub-100 m NHATS-type targets, and that within a simulated ten-year survey 52–61% of 30–100 m objects and 74–76% of objects above 100 m are accessible at 4 km/s or less round trip. This is a survey-advocacy document from an agency, so treat the completeness projections as advocacy and the catalogue counts as fact. The question it does not answer is whether a larger catalogue produces a customer.

Established The delta-v asymmetry is the whole basis of the mass-payback argument, and it is real. On Ross's comparative table: Earth surface to LEO is 8.5 km/s; LEO to a near-Earth asteroid is 4.0 km/s or more; the return transfer from an asteroid to Earth is about 1.0 km/s; lunar surface to LEO is 2.4 km/s. About 90 known near-Earth asteroids, roughly 6% of the then-catalogued population, are more accessible than the Moon, a figure Sonter puts nearer 10%. Cheap to come back, expensive to go — that is what makes returning mass from an asteroid arithmetically attractive and going there expensive.

Frontier Which product? The field's live disagreement is between water for use in space and platinum-group metals returned to Earth. Sonter, Ross, Sercel, Elvis and Crawford converge on water as the defensible near-term product; AstroForge's stated business plan and most popular coverage back the metals. Frontier The strongest evidence against the water case is not physical but commercial: the only operator currently flying hardware has publicly declined the water market for want of demand, and the policy director of the largest firm ever funded dated the customer base at twelve to fifteen years away in 2019.

Speculative Does optical mining work? Sercel's APIS concept, a NIAC Phase I study from 2015, claims up to 100 tonnes of water returned from a near-Earth asteroid on a single Falcon 9 v1.1 launch by concentrating sunlight to drill and disrupt the body inside a containment bag, vaporising water at very low pressure and capturing it cryogenically as ice. Its own rationale is comparative — it argues asteroids beat the Moon on round-trip delta-v and surface-operations logistics. Sercel is TransAstra's founder and an interested party on his own architecture, and no element of APIS has flown.

Frontier What actually drives the economics, on the only modern techno-economic treatment? Hein, Matheson and Fries identify the drivers as throughput rate, number of spacecraft per mission, and the cadence of successive missions — fleet scale and learning curve, not ore grade. Established The dollar figures in that paper are not quoted here: only the abstract was obtained during research for this brief and the full text was blocked, so the driver list is safe and the numbers are not.

Frontier Is the right metric the one the field uses? Ross argues that mass payback ratio is the wrong figure of merit because it “does not take into account development costs, difference in value between mass launched and mass returned, nor … the time-cost of money,” that net present value is the right one, and that missions exceeding three years need very good payback ratios to survive discounting; the stated aim is ratios above 100. He estimates that about 50% of near-Earth asteroids may be water-bearing, and declines to declare feasibility.

Speculative The long-run position that keeps the field alive. The standing advocacy claim is that asteroid material becomes the industrial base for large-scale activity in space — that the first tonne is worthless and the millionth is civilisational. Nothing in the retrieved record tests it, because the first tonne has not been extracted. It is a coherent hypothesis with no direct evidence, and it should be labelled as one rather than argued with.

Speculative And the strategy question the sector is currently running as an experiment. AstroForge's explicit position is that fast, cheap, high-failure-rate iteration is the right development method for deep space — ten-month builds, $3.5 million spacecraft, 30% assigned odds. Frontier Two attempts have produced two losses and zero asteroid data, which is consistent both with the strategy failing and with the strategy working exactly as designed. The third flight is the test, and the honest position until then is that nobody knows.

Established The binding bottleneck is a customer, and it is named as such by the people who ran out of money. Marquez put the customer base at twelve to fifteen years away in 2019; the only NASA cost analysis of in-space propellant models a government-only customer base because there is no other one to model; AstroForge declines the water market on demand grounds. No property regime, launch price or drill design creates a buyer.

Frontier Prospecting is the missing measurement. Every composition figure for a candidate mining target is spectroscopic and remote. The in-situ measurement that would distinguish an ore body from a rock has never been made at a near-Earth asteroid by a commercial mission: Arkyd never left Earth orbit, Brokkr-1 failed on communications, Odin was lost. DeepSpace-2 is built and unflown. The industry is bounded by an absence of data about its own resource.

Frontier Detection of the useful population is a survey problem at the edge of capability. Elvis's water case requires objects as small as 18 m, findable only at around 0.04 AU with present instruments; his own paper proposes a heliocentric survey from a Venus-like vantage to find the low-delta-v population at all. The rocks that make the water arithmetic work are the ones hardest to see.

Frontier Refining in flight is unproven and has already broken one spacecraft. Brokkr-1's refining payload generated magnetic interference that disrupted attitude determination and antenna pointing — a failure mode that is specific to carrying a processing plant on a small bus, and that was known before launch. The interaction between a refinery and its host vehicle is a real engineering bottleneck with exactly one flight data point, and that point is negative.

Established Launch price is not the bottleneck, and treating it as one is the field's most common error. On the NASA Ames analysis, launch is 10 to 30% of total mission cost. Halving launch price moves the total by 5 to 15%, against a shortfall measured in orders of magnitude. Established The same source gives the history in consistent units: about $1,000k/kg at the 1950s peak, $10–20k/kg from 1970 to 2010, $62k/kg for the Shuttle, $2.7k/kg for Falcon 9 and $1.4k/kg for Falcon Heavy. That is a NASA analyst deflating a favourite agency talking point — interest against the finding.

Frontier Power is the quiet constraint on anything past a flyby. Odin flew with 180–190 W; DeepSpace-2 carries 1.7 kW for electric propulsion and surface operations. Extraction — heating, comminution, volatile capture — is a thermal process, and the step from a hundred watts to the tens of kilowatts a real plant needs has no flown precedent on a small commercial bus.

Established Communications and ground segment defeated the most recent attempt outright. AstroForge's own debrief records near-universal ground-station problems: wrong configurations, mispointing, amplifier losses, local interference. A deep-space mission is a ground system with a spacecraft attached, and the sector's newest entrant discovered that the expensive way.

Speculative Anchoring and material handling at microgravity remain undemonstrated for extraction. Both sample returns used touch-and-go contacts of seconds. ARM would have demonstrated robotic capture with anchoring grippers and was cancelled before flying. Optical mining proposes to sidestep anchoring by bagging the object; nothing in that architecture has been tested off the ground.

Established The result this brief most needs is a demand measurement, and it cannot be produced by any experiment in space. Whether an in-space market of thousands of tonnes a year exists is a question about Deep Space Infrastructure, Space-Based Manufacturing and crewed programmes, not about asteroids. Every technical result in this brief is downstream of it.

Frontier An in-situ composition measurement at a near-Earth asteroid by a mission designed for resource assessment. Sample return has produced two exquisite datasets at two bodies chosen for science. The resource question needs measurements at the accessible population — grade, distribution, mechanical properties — and DeepSpace-2's landing legs and platinum-group-metal instrumentation are the only funded attempt at it.

Frontier A survey that resolves the accessibility disagreement. Elvis's strict gate and NHATS's loose one produce populations differing by more than an order of magnitude. A dedicated survey of the kind Abell et al. model would settle which population a mission architect is actually choosing from, and would do it with telescopes rather than spacecraft — the cheapest useful work available to the field.

Frontier Published economics that survive scrutiny at the level of numbers. The modern techno-economic treatment by Hein, Matheson and Fries gives its drivers in an abstract; its cost and net-present-value figures were not obtainable during research for this brief. The companion water-propellant economics paper in Acta Astronautica 176 could not be fetched at all. A field whose two most relevant economic analyses are paywalled to the point of non-citation has a dependency on open publication, not on physics.

Established Terrestrial commodity data. The elasticity argument against platinum return is qualitatively sound and quantitatively empty without world production and reserve figures. Those were refused on every attempt during this research. This is a dependency on a database, and it is the reason no percentage appears on this page.

Frontier Autonomy. Round-trip light time to a near-Earth asteroid runs to minutes, and every extraction step — contact, anchoring, comminution, capture, containment — would have to be executed without supervision. No mission has demonstrated an autonomous extraction sequence anywhere, and the ground-segment record of the most recent attempt suggests the human-in-the-loop half is not solved either.

Frontier The decisive near-term experiment is a commercial prospector that actually reaches an asteroid. DeepSpace-2 — about 200 kg, 1.7 kW, electric propulsion, landing legs, platinum-group-metal instrumentation, manifested on IM-3 — is built. It measures; it does not mine, and the company says so. Its result would be the first commercial composition data ever taken at a near-Earth asteroid, and its failure would be the sector's third loss in a row.

Established Tianwen-2's capsule in 2027 is the next hard measurement. Anchor-and-attach and touch-and-go sampling at a quasi-satellite, with explosives used to expose sub-surface volatiles, followed by an extended cruise to a main-belt comet arriving January 2035. It tests sampling technique at a new class of target, which is a capability result rather than an economic one.

Speculative A volatile-extraction demonstration at any scale would be a first. Nothing has extracted and retained a volatile from an asteroid. The APIS optical-mining architecture is the most specified proposal, claiming 100 tonnes of water on one Falcon 9; the experiment that would matter is far smaller — kilograms captured, retained through a cruise, and measured — and it has not been proposed as a funded mission by anyone.

Frontier A refining payload flown without breaking its host. Brokkr-1's magnetic interference is a specific, reproducible engineering question: how much electromagnetic and thermal disturbance a processing plant imposes on a small bus, and what isolation costs in mass. That is a ground-testable experiment that was skipped, and it destroyed a mission.

Frontier A dedicated low-delta-v survey. Elvis's proposed heliocentric vantage and the NHATS-driven survey modelling both point at the same experiment: find the small accessible objects. It requires no new physics, produces a publishable catalogue whatever it finds, and would resolve the field's largest quantified disagreement.

Speculative And the experiment nobody will run: a market test. The clean version would be a standing purchase commitment for water delivered in orbit at a stated price, which would immediately reveal whether the demand curve exists. The nearest real-world analogue is on the lunar side, where a government bought regolith for a legal reason — described in Space Resource Economies — and the total was five figures.

Established The proximity-operations problem is solved at demonstration scale. OSIRIS-REx executed a touch-and-go at Bennu and returned 121.6 g; Hayabusa2 executed two touchdowns plus an impactor experiment and returned 5.424 g. Rendezvous, station-keeping around a low-gravity body, sample acquisition, containment and Earth return are proven capabilities. The engineering gap is not in getting there.

Frontier The gap is in everything that happens after contact. Continuous operation on a surface at microgravity requires anchoring, reaction-force management and thermal control over months rather than seconds. ARM would have flown anchoring grippers at the four-metre-boulder scale and was cancelled; optical mining proposes to avoid anchoring entirely by enclosing the object in a bag and using concentrated sunlight to disrupt it, which trades a mechanical problem for a thermal and containment one. Neither route has hardware in space.

Frontier Power scaling. Any thermal extraction process is limited by delivered watts. The step from Odin's 180–190 W to DeepSpace-2's 1.7 kW is a factor of nine and represents the sector's entire progress in installed power; a plant processing tonnes needs another two orders of magnitude, which on current architectures means large solar arrays or nuclear power and a mass budget that dominates the spacecraft.

Established Electromagnetic compatibility between payload and bus is a demonstrated failure mode. The Brokkr-1 loss is the only in-flight data on flying a refinery, and it says the refinery interferes with the spacecraft. Shielding, separation or magnetic-field cancellation all cost mass on a vehicle whose economics depend on being small.

Established The ground segment is engineering, not overhead. The AstroForge debrief's list of ground-station failures — configuration errors, pointing errors, amplifier losses, local interference — is the direct cause of two mission losses. A firm that builds a $3.5 million spacecraft in ten months and cannot reliably talk to it has optimised the wrong half of the system.

Frontier Mass return architecture. Ross's delta-v table gives the asymmetry that makes return cheap — about 1.0 km/s from a near-Earth asteroid to an Earth transfer, against 8.5 km/s to reach LEO from the ground — and Ross's own framework insists that mass payback is the wrong metric and net present value the right one, with three-year mission durations already straining the discounting. The engineering requirement that follows is speed, not size: a long mission cannot be rescued by a large payload.

Speculative And the requirement nobody costs: doing it repeatedly. Hein, Matheson and Fries name mission cadence and fleet size as the dominant economic drivers. That converts asteroid mining from a spacecraft problem into a production-line problem — which is the point at which it stops being an aerospace question and becomes the manufacturing question owned by Space-Based Manufacturing.

Established Space-Based Manufacturing is the adjudicated customer edge, and the customer says no. That brief owns orbital manufacturing from Earth-launched feedstock — ZBLAN fibre, protein crystals, semiconductor composites, and the down-mass constraint — and states its own boundary as: the moment the feedstock stops coming up from Earth, it is another brief's topic. This brief is that feedstock question. The dependency is a demand dependency rather than a technical one, and it must be reported honestly: I-24's own finding is that its market may be structurally small precisely because the manufacturing works, and it currently buys nothing off-world.

Established Deep Space Infrastructure owns depots, transfer stages, refuelling and relays; this brief owns extraction. The seam is one sentence: a mined kilogram with nowhere to be stored and no stage to move it is not a product. Reciprocally, that brief should not claim asteroid supply as an input, because nothing has been extracted.

Established Space Resource Economies owns whether anyone buys it. The Planetary Resources and Deep Space Industries record belongs to both pages with different weights: here it is evidence about engineering and capital formation, there it is evidence about market structure. The same paragraph is deliberately not printed twice.

Established Space Law and Governance owns property rights in mined material entirely. Whether a mined kilogram can lawfully be owned and sold is unsettled and is taken up there. Frontier One correction is worth making from this side: no operator in the retrieved record names legal uncertainty as the reason they failed. They name the customer.

Frontier Lunar Industry and Moon-Based Manufacturing are the competing supply route, and Ross's table quantifies the competition: lunar surface to LEO is 2.4 km/s, against 4.0 km/s or more from LEO to a near-Earth asteroid outbound and about 1.0 km/s inbound. About 6% of catalogued near-Earth asteroids are more accessible than the Moon, which is the entire case for preferring them — and the Moon has a government customer and a landing cadence that asteroids do not.

Frontier Solar-electric propulsion is the shared enabling technology. ARM was designed to demonstrate it at high power and did not fly; DeepSpace-2 carries it at 1.7 kW; every optical-mining and mass-return architecture assumes it. Its maturity is the single technical variable that most changes this brief's arithmetic, and it is developed elsewhere for other reasons.

Established The capital-formation record is the institutional record, and it is complete enough to read. Roughly $50 million into Planetary Resources including a $21.1 million Series A and a €25 million Luxembourg package; $3.5 million into Deep Space Industries; about $53 million into AstroForge. Total private capital of order a hundred million dollars, three CubeSats and two lost deep-space spacecraft, and no prospector that has reached an asteroid. That is the sector's balance sheet.

Frontier National industrial policy has been tried and has not produced a mission. Luxembourg's capital-and-grant package to Planetary Resources is the clearest case: a state buying into the sector directly. The company was wound up two years later. Sovereign money changed the cap table and not the flight record.

Established Agency programmes fail on a different axis. ARM consumed $105 million in 2014 on concept maturation against a $2.6 billion estimate, slipped three times, and was closed out on 13 June 2017 when priorities moved to Artemis. Capability programmes die of politics, commercial ones die of customers — and an institution that wants asteroid capability has to survive an election cycle rather than a funding round.

Frontier NIAC is the field's main public funding instrument and it funds concepts, not hardware. APIS reached Phase I in 2015 with a 100-tonne water claim written by its proposer. A NIAC abstract is a proposal document by an interested party, and the institutional consequence is that the field's most-quoted architecture numbers have never been through adversarial review.

Established The evidence base is unusually dependent on what firms choose to publish. The single most informative document in this brief is AstroForge's own failure debrief, which exists because the company chose to publish it. Frontier Nothing requires a private operator to disclose a loss, its causes or its odds, and a sector whose record is assembled from voluntary post-mortems and tertiary encyclopedia entries has an institutional transparency problem that no regulator currently addresses.

Frontier Publication access is an institutional bottleneck with a measurable effect on this page. Three of the most relevant items — the Hein full text, the Acta Astronautica water-propellant economics, and the USGS commodity summaries — were unobtainable during research. Every number this brief declines to print traces to a paywall or a 403, and that is an institutional fact rather than a scientific one.

Frontier The elasticity argument has a distributional edge that is rarely stated. If a delivery of platinum-group metals large enough to matter did arrive, the price effect would fall on terrestrial producers and the economies built around them. This brief asserts no magnitude — the production data were not obtainable — but the direction is a transfer from existing mining regions to whoever lands the cargo, and that is a distributive question rather than a technical one.

Frontier The environmental argument runs the other way and deserves stating in its strongest form. Crawford's review makes it: if ecosystem services were properly priced into terrestrial extraction, off-world sources become comparatively attractive, and “obtaining raw materials from uninhabited ones may come to be seen as ethically preferable.” Speculative The argument is coherent and currently untestable, because the comparison requires a delivered off-world cost that does not exist.

Established Crawford also supplies the sentence that bounds the whole ethical debate: “one would not go into space for a source of metals for use on Earth.” The value of off-world material is positional — it is worth something because of where it is. That reframes the ethics from resource justice on Earth to access and priority in space.

Frontier Planetary protection applies weakly to asteroids and not trivially. Sample-return missions carry containment requirements; extraction at scale would disturb bodies that are scientific records of solar-system formation. Elvis's own ore-probability work implies that the ore-bearing objects are of order ten in number — which means the mineable set and the scientifically distinctive set may be nearly the same set.

Speculative Risk transfer is a real if small consideration. Manoeuvring a near-Earth object is the same capability as deflecting one, and ARM's stated objectives included gravity-tractor deflection. A technology that can move an asteroid toward a useful orbit can move one toward a less useful one, and the governance of that dual use belongs to Space Law and Governance.

Frontier And the honesty question about the sector's public numbers. A $10,000 quadrillion valuation originating with a mission's own principal investigator as a thought experiment, propagating for a decade stripped of its framing, is a communication failure with consequences: it has recruited capital, shaped public expectation and outlived the composition estimate it rested on. Naming interest at the point of citation is the only available remedy, and this brief applies it throughout.

Speculative The civilisational claim for asteroid mining is that it removes the mass constraint on activity in space. Every large structure — habitats, power satellites, shielding — is presently limited by what can be launched, and asteroid material would in principle lift that limit. The claim is coherent and entirely untested, because the first tonne has not been extracted and the machinery to use it does not exist.

Established The measured version of the same claim is small. The entire history of resource production off Earth is 122 grams of oxygen produced by MOXIE on Mars, and the entire history of asteroid material delivered is 127 grams across two missions. Those numbers are the same order of magnitude, and they are the base from which any civilisational projection has to start.

Frontier If the water case ever closes, the consequence is architectural rather than economic. Propellant produced in space changes mission design more than it changes cost: it decouples departure mass from launch mass. Sonter's thousands of tonnes a year is the threshold at which that becomes real. Against a current in-space extracted-propellant market of zero, the threshold is not near.

Speculative The asymmetry between the two products has a civilisational reading. Metals returned to Earth make asteroids a commodity source and change terrestrial markets; water used in space makes them infrastructure and changes nothing on the ground. The second is the more defensible case and the less transformative one, which is an uncomfortable pairing for a field that has been funded on the first.

Frontier The concentration result imported from the lunar case applies here too. If ore-bearing near-Earth objects number of order ten under a strict accessibility gate, then the resource is not abundant but rivalrous, and the relevant institutional model is allocation among competing claimants rather than a commodity market. Scarcity in space looks like scarcity everywhere else: access, energy and cost bind.

Handwave And the framing that does not survive contact with the numbers. “Space is infinite resources, so scarcity ends” is the popular civilisational claim attached to this subject. It is defeated by the same arithmetic that defeats every other claim on this page: the resource is not the constraint, the cost of reaching it and the absence of anyone to sell it to are.

These horizons track missions that are manifested and firms that are funded, because there is no extraction programme anywhere to track:

• 10 yr: Frontier Tianwen-2's capsule returns in 2027, and DeepSpace-2 flies on IM-3 if Intuitive Machines' manifest holds — the first commercial composition measurement at a near-Earth asteroid, or the sector's third consecutive loss. Frontier A dedicated low-delta-v survey is the cheapest publishable work in the field and could begin at any time. Speculative Marquez's twelve-to-fifteen-year window from 2019 expires inside this horizon; nothing in the record since contradicts him. Speculative No extraction of any mass from any asteroid is scheduled by anyone.
• 25 yr: Speculative Tianwen-2's extended mission reaches 311P/PanSTARRS in January 2035, which is a science result. Speculative A first demonstration of volatile capture at kilogram scale is the plausible technical milestone at this range, and it is not funded, proposed as a mission, or costed by anyone outside a NIAC abstract. Handwave Any date offered for first commercial delivery of asteroid propellant is an assertion, because the quantity that would have to be bought does not have a buyer.
• 50 yr: Speculative If an in-space propellant market of thousands of tonnes a year ever exists, this is the horizon at which Sonter's first condition could be met — and it depends on crewed and infrastructure programmes owned by other briefs, not on anything asteroid-specific. Handwave Platinum-group return to Earth has no defensible horizon at all: the valuation collapses on delivery, the ore-bearing population is of order ten objects under a strict gate, and no mechanism has been proposed for pricing a delivery that changes its own price.
• 100 / 250+ yr: Handwave Beyond forecasting. Speculative The defensible structural statement is that the two conditions Sonter named in 1997 are the same two conditions today, that neither has moved toward being met in twenty-nine years, and that both are properties of demand rather than of asteroid technology — so the horizon is set by what else humanity does in space, not by progress in mining.

• Depends on Two edges, and both are about the other end of the transaction. Space-Based Manufacturing is the only demonstrated in-orbit consumer of materials, which makes it this brief's adjudicated customer — and the edge has to be reported with its own bad news, because that brief's finding is that its market may be structurally small precisely because the manufacturing works, and it buys nothing off-world today. Deep Space Infrastructure owns depots, transfer stages and refuelling: a mined kilogram with nowhere to be stored and no stage to move it is not a product. Neither edge is technological. Asteroid extraction does not wait on a physics result or a materials result; it waits on somebody at the other end of the delta-v curve wanting the material.
• Requires (not on this map) Three constraints that are not briefs on this map, and the first two are the ones Sonter named in 1997. The market condition is a demand fact: he required an in-space market of some thousands of tonnes a year, and the confirmed in-space market for extracted material is zero tonnes. The launch condition is a price fact: he required several hundred dollars a kilogram, and Falcon Heavy and Falcon 9 deliver at $1,400 and $2,700 — while launch is only 10 to 30% of mission cost, so even meeting it moves the total by less than it appears. The scientific constraint is a measurement nobody is producing: every grade figure for a candidate target is remote and spectroscopic, and the one funded attempt at an in-situ measurement, DeepSpace-2, has not flown.
• Enables No enabling edge is claimed, and claiming one would be dishonest at this stage. A working extraction industry would supply Deep Space Infrastructure and change the mass budget for every large structure in orbit — but zero kilograms have been extracted, both firms that raised money to try are gone, and the third has lost two spacecraft. An edge from a capability with no hardware, no delivered mass and no customer would record a wish rather than a dependency.
• Adjacent Space Resource Economies is the closest neighbour and takes the same events as market-structure evidence rather than engineering evidence. Lunar Industry and Moon-Based Manufacturing are the competing supply route, separated by Ross's delta-v table — 2.4 km/s from the lunar surface to LEO against 4.0 km/s outbound to an asteroid and 1.0 km/s back. Space Law and Governance owns whether a mined kilogram can be owned, which this brief names once and does not re-argue.

Established “Asteroid mining has begun.” Two sample returns totalling 127 grams are not a mining record. OSIRIS-REx and Hayabusa2 were science missions that demonstrated rendezvous, contact and return; neither extracted a resource for use or sale, and no spacecraft ever has. Frontier The phrase “asteroid mining mission” applied to either of them is the single most common error in the subject, and it is what allows a field with zero extracted kilograms to be described as operational.

Handwave “Psyche is worth $10,000 quadrillion.” The figure originates with Lindy Elkins-Tanton, the mission's own principal investigator, as an iron-content thought experiment about what the metal would be worth if it were on Earth. Established It prices a stock at a price that assumes the stock is never delivered — the delivery is what destroys the price — and Metzger's rebuttal is that transport cost would negate the value regardless. Frontier And the composition premise has since been revised: Cantillo and Reddy estimate about 82.5% metal with roughly 35% bulk porosity, reading Psyche as closer to a rubble pile than an intact core. Three independent failures in one number, and NASA's own mission page mentions no monetary value at all.

Frontier “Platinum from asteroids will be enormously profitable.” Elvis's worked case gives a 100 m metallic body at 10 ppm about 23.6 tonnes of platinum-group metal, priced at $1.18 billion at prevailing prices. Established The valuation is self-defeating by construction: the prices used are prices that exist only because the delivery has not happened, and the demand is industrially inelastic in the short run. Speculative This brief states the direction and refuses the magnitude. The USGS commodity data needed to quantify the collapse returned 403 on every attempt during research, so no percentage appears here — and any page that gives you one without a fetched production figure is guessing.

Established “Commercial prospecting has begun.” Commercial prospecting has been attempted. Planetary Resources flew three CubeSats and never left Earth orbit; Brokkr-1 failed on communications with a known, pre-identified magnetic-interference risk it flew anyway; Odin was declared lost on 6 March 2025 with no asteroid data returned. Frontier DeepSpace-2 is built and unflown, and its own company describes it as a measurement mission rather than a mining one.

Established “Cheap launch fixes this.” Launch is 10 to 30% of total mission cost on the NASA Ames analysis, so halving it improves the total by 5 to 15%. Frontier The shortfall this has to close is measured in orders of magnitude, not in percentages. Falcon 9 at $2.7k/kg and Falcon Heavy at $1.4k/kg are already a factor of forty below the Shuttle's $62k/kg, and the industry that was supposed to appear at those prices did not.

Frontier “There are thousands of accessible targets” — and “there are almost none.” Both circulate, and they are both true of different gates. Elvis's 4.5 km/s one-way criterion admitted 65 of 6,699 known objects in 2010; NHATS's 4-to-12 km/s round-trip criterion admits 6,294 of 38,812 today. Established These are not competing measurements, they are different questions, and the honest statement is that no mission has yet distinguished which population a real architecture draws from.

Speculative “Water is the easy case.” It is the better case and it is not easy. Elvis's water arithmetic needs bodies as small as 18 m, which present surveys can only find at about 0.04 AU; the 236-tonne payload is priced at $1.2 billion in LEO, which requires a buyer in LEO; and the only firm currently flying hardware has publicly declined that market for want of demand. Frontier The water case fails on the buyer, and the platinum case fails on the price. Neither fails on the chemistry, which is why the chemistry gets all the attention.

Speculative “A single Falcon 9 can bring back a hundred tonnes of water.” That is the APIS optical-mining claim, made in a NIAC Phase I abstract written by the concept's own proposer, Joel Sercel of TransAstra. Handwave No element of the architecture has flown, been prototyped in a relevant environment, or been through adversarial review, and the claim is quoted throughout the popular literature without its provenance.

Frontier “The firms failed because of legal uncertainty about ownership.” No operator in the retrieved record says so. Planetary Resources' policy director named the customer base; AstroForge names demand when declining the water market; NASA's own propellant cost study models a government-only buyer because no other exists. Established Property rights belong to Space Law and Governance and should not be promoted to a cause here.

Speculative “Fast iteration is obviously the right strategy for deep space.” It is a live hypothesis with two data points, both negative. AstroForge assigned Odin 30% odds, flew known power-amplifier and deployment risks to avoid a nine-month delay, and lost the spacecraft. Frontier Whether that is a strategy failing or a strategy working as designed cannot be told from two flights, and the company's willingness to publish the debrief at all is the strongest evidence in its favour.

Frontier And the framing itself. “Asteroid resources are an economic proposition” has never been refuted — it has simply never had its premises met. Established Sonter specified them in 1997: an in-space market of thousands of tonnes a year, and Earth launch at several hundred dollars a kilogram. In 2026 the market is zero tonnes and launch is $1,400 to $2,700. Frontier Roughly a hundred million dollars of private capital has been spent testing whether those premises would arrive early, and the answer so far is three CubeSats, two lost spacecraft and one still-unflown prospector.