High Temperature Superconductors

High-temperature superconductors are materials that carry current with no resistance above the boiling point of liquid nitrogen, 77 K, rather than requiring liquid helium at 4 K. The distinction is economic before it is physical: liquid nitrogen costs roughly what milk costs and can be made on site from air, while liquid helium is a finite geological product with a volatile price and a supply chain that has failed three times since 2006. A material that works at 77 K is not merely a better superconductor. It is a superconductor that an ordinary industrial site can afford to keep cold.

In practice the field means one family. Rare-earth barium copper oxide — REBCO, where the rare earth is usually yttrium or gadolinium — is manufactured as a coated conductor: a superconducting layer one to five micrometres thick, deposited on a textured buffer stack, on a Hastelloy substrate, wrapped in copper. The result is a tape, typically four to twelve millimetres wide and about a tenth of a millimetre thick, sold by the kilometre. Nearly every superconducting object built or planned since about 2010 is made of this tape, and the small number of remaining exceptions are legacy low-temperature niobium–tin and niobium–titanium magnets in MRI scanners and particle accelerators.

Where this brief stops, and it is a boundary worth stating precisely because the two slots collided in scoping. This brief owns everything priced per kilo-ampere-metre: the tape itself, its critical current, its in-field performance, its piece length, its deposition yield, the manufacturing base that makes it, and the still-open physics of higher-temperature families that might one day replace it. Energy Systems slot 5, superconducting infrastructure, owns everything priced per installed metre or per delivered megawatt: the cables, fault current limiters, storage rings and transformers that have actually been energised, their cryogenic plant, and the base rate of deployment. The arbitration rule where they meet is the unit of account. If the question is how cheap and how uniform the tape can get, it belongs here. If the question is what anyone actually built and whether it paid, it belongs there.

One consequence of that cut is worth flagging at the outset, because it runs against the way this subject is usually organised. The fusion magnet programme sits here, in the materials brief, not in the infrastructure brief. Commonwealth Fusion Systems is not an application of superconducting infrastructure; it is the demand signal that currently sets tape price, yield targets and capacity expansion for the entire industry. Treating it as an application would bury the fact that it is the market.

Established The material is not the problem, and has not been for about fifteen years. Long-length production tape exceeds 300 amperes per centimetre of width at 77 K in self field, and the best laboratory conductors — thick REBCO layers of around five micrometres with artificial pinning centres — reach 1,300 to 1,500 A/cm-width. In the cold, high-field regime that fusion magnets work in, engineering current density exceeds 1,000 A/mm² at 20 K and 20 T, with critical current above 200 A per 4 mm of tape width at that temperature and field — and that figure is quoted for the field orientation in which REBCO performs worst, with the field perpendicular to the tape face.

Established Field headroom is not the constraint either, and the margin is not close. The irreversibility field — the field above which the material stops usefully pinning flux, which is the practical ceiling rather than the thermodynamic one — has been measured at 45 T at 45 K, with extrapolated bounds around 100 T at 20 K and 110 T at 4.2 K. Set that against the duty a grid cable asks of it: a 35 kV superconducting cable operates in its own self-field of order 0.1 T. The material has roughly a thousandfold field headroom over what grid hardware needs. This is the single most under-appreciated number in the subject, and it is why arguments about whether we need a better superconductor for the grid are answering a question nobody asked.

Established What has not moved is the price. On the standard 77 K self-field metric, REBCO tape was around $100 per kilo-ampere-metre in 2022, and 80 to 100 euros per kilo-ampere-metre in February 2026. That is flat, and in dollar terms it is worse than flat. It has stayed flat across precisely the years in which world output grew roughly tenfold, roughly $80 million of United States public money went into a single new production facility, and suppliers put something like $230 million of their own capital into capacity during 2024 alone.

Established Capacity, by contrast, grew sharply. Global annual REBCO capacity now exceeds 5,000 km at twelve-millimetre equivalent width. The four top-tier pulsed-laser-deposition producers — Faraday Factory Japan, Shanghai Superconductor Technology, Fujikura, and S-Innovation — account for more than 3,000 km a year between them, over half of world output. Roughly fifteen manufacturers exist worldwide, with a second tier including MetOx and SuperPower/Furukawa in the United States, SuperMag in China, Suprema in Italy and SuNAM in Korea transitioning toward the same deposition route.

Established The two facts above are the argument of this brief. Volume rose by an order of magnitude and price per kilo-ampere-metre did not fall. That is not what a maturing manufactured product does, and the reason is structural rather than commercial: the superconductor is only a few percent of the tape's cross-section. Most of what you buy is Hastelloy substrate, oxide buffer layers and copper stabiliser, all of which scale with length and none of which scales with current. Making more tape moves dollars per metre. It barely moves dollars per kilo-ampere-metre, because the denominator is set by how much current the thin layer carries, and that is set by deposition quality rather than by production volume.

Established Yield is where the money is, and yield collapses with length. The controlling published statement is blunt: manufacturing yield for long-length tape is low, “especially for the ones longer than 500 m.” A magnet or a cable needs continuous piece lengths; a splice is a resistive joint, a heat source and a reliability liability. So the industry's real problem is not making more square metres of REBCO, it is making the same square metre uniformly enough that a kilometre of it passes acceptance end to end.

Frontier Whether the projected volume increase bends the price curve is genuinely open. The February 2026 review projects roughly an eightfold volume increase by 2028, and one grid project has already contracted at 50 €/kA·m for 2027 delivery — a below-market forward price obtained at the peak of fusion demand. That is real evidence that the curve can bend. It is not yet evidence that it has.

Handwave Every claim that tape is on the verge of $10 to $20 per kilo-ampere-metre. That target has been restated in roughly its present form for two decades without being met, and it is currently restated by a national research agency as the level needed for competitiveness against a commercial price of about $300/kA·m at fusion operating conditions — a thirtyfold gap. A number that has been five years away for twenty years is a target, not a forecast, and this brief treats it as one.

Established And the demand has moved. The anchor customer for REBCO tape is not a utility. It is a fusion programme. A single tokamak's toroidal field set consumes roughly 10,000 km of tape — about two years of total world output — and adds $100 million or more to the machine. Whatever the grid eventually wants from this material, it will be buying into a market whose price, specifications and delivery schedules are being set by someone else.

The genuine research frontier in this subject splits cleanly in two, and the two halves have almost nothing to do with each other. One is a process-engineering frontier in a material we already have. The other is a condensed-matter physics frontier in materials we cannot build anything from. Conflating them is the characteristic error of writing on superconductivity, and it consistently makes the field look closer to a breakthrough than it is.

Frontier The first frontier: kilometre-scale uniformity at high yield. This is where the money and the industrial effort actually are. Pulsed laser deposition, the dominant route among top-tier producers, has specific and documented bottlenecks: laser target utilisation below 50%, so more than half the expensive rare-earth target material is wasted; long lead times on target raw materials; and narrow stable deposition windows that make process reproducibility hard across a reel. Getting past one kilometre of uniform piece length at high yield is the real frontier of this brief, and it is a metallurgy and process-control problem, not a physics problem. No new physics is required and none is expected.

Frontier A second industrial frontier, less discussed: quench protection intrinsic to the tape architecture. Coated conductors have low normal-zone propagation velocity — when a small region goes normal, the resistive zone spreads slowly along the tape. That sounds benign and is the opposite. A slowly spreading normal zone concentrates the dissipated energy in a small volume, leaving coated conductors, in the published phrasing, “excessively unprotected against hot spots.” This is a property of the tape, not of any application built from it, which is why it belongs in this brief rather than the infrastructure one. It is also why every serious magnet programme spends real engineering on detection and dump circuits.

Frontier The second frontier: nickelates, and this is the live one in physics. The strongest result to date is bulk and under pressure. Single crystals of La_1.57Sm_1.43Ni₂O_7−δ show a 96 K onset, and La₂SmNi₂O_7−δ shows a 92 K onset with zero resistance at 73 K under 21.6 GPa, with the Meissner effect confirmed at 20.6 GPa. The crystals are grown at ambient pressure and then measured under pressure. This is the first nickelate result above liquid-nitrogen temperature with bulk Meissner evidence, and it was published in December 2025.

Frontier At ambient pressure the nickelate results are real and very thin — three unit cells thin. Films of La_2.85Pr_0.15Ni₂O₇ grown on strontium lanthanum aluminate, three unit cells thick, show a 45 K onset with zero resistance near 9 K and Meissner diamagnetism at 8 K. By mid-2026 a review reports the ambient-pressure onset pushed to 63 K, and 68.5 K onset under a modest 2.0 GPa on compressively strained films. That review is a preprint and is flagged accordingly.

Frontier What is unresolved in the nickelates is unresolved at a deep level, and the field says so. Photoemission gives conflicting Fermi-surface topologies between groups, specifically over whether a particular electron pocket appears at all. Pairing symmetry is contested between s±, d-wave and d+is scenarios. And the relationship between lattice ratio and transition temperature in films contradicts the relationship measured in bulk. These are not details awaiting cleanup. They are the mechanism, and it is not known.

Handwave Any framing of nickelates as a near-term infrastructure material. The ambient-pressure result is three unit cells of film on one specific substrate under compressive strain, with an onset in the forties to sixties of kelvin and zero resistance in the single digits. That is a condensed-matter result of the first rank. It is not a conductor, and there is no published route from it to one.

Established The hydrides are established physics and are not engineering materials. H₃S superconducts at 203 K at megabar-class pressure; LaH₁₀ reaches 250 K at about 170 GPa, with zero resistance, a demonstrated isotope effect, and an upper critical field extrapolating to around 136 T. In December 2025 the superconducting gap in H₃S was measured directly for the first time — about 60 meV, against 44 meV in the deuterated compound — confirming a conventional electron–phonon pairing mechanism. That is a significant result and its own authors frame it correctly: it enables the search for materials that work at more practical conditions. It does not itself reduce the pressure required by a single gigapascal.

Handwave Any claim that hydrides are a path to practical superconducting infrastructure. These materials exist inside diamond anvil cells at pressures exceeding one million atmospheres. No source consulted for this brief proposes a route from a megabar sample to a wire, a magnet or a cable, and the honest expectation from inside the field is instructive: roughly fifty groups work on hydride superconductivity, and current predictions target about 100 K at reduced pressure, not room temperature at ambient pressure. A 100 K material that still needed gigapascals would be worse than REBCO for every application in this brief.

Established No significant ambient-pressure room-temperature superconductivity claim emerged in 2025 or 2026. This is stated as the result of a specific search rather than as proof of absence, and the distinction matters: it is a negative search result. What it does establish is that the field's own attention has moved from room temperature to the nickelates and to lower-pressure ternary hydrides, which is a more informative signal about expectations than any individual announcement.

Established The first bottleneck is yield above 500 m, and it is the one that gates everything else. Long-length manufacturing yield is explicitly low above that threshold in the published review literature. Because splices are unacceptable in the highest-value applications, low long-length yield does not merely raise cost proportionally — it truncates the addressable market to buyers who can tolerate shorter unit lengths, which excludes exactly the large magnet and long-cable programmes that would drive volume.

Established The second is that the cost denominator is set by deposition, not by scale. With the superconductor at a few percent of cross-section, the dominant levers on $/kA·m are layer thickness and in-field pinning, both process-physics variables. This is why a tenfold volume increase from roughly 2016 to 2026 produced no fall in the 77 K self-field price, and it is why capacity announcements should not be read as price forecasts.

Established The third is target utilisation and consumables in PLD. Losing more than half the target material is a direct, recurring input cost on the dominant production route, compounded by lead times on target supply. Improving utilisation is unglamorous and is one of the few levers that moves cost without touching the physics of the deposited layer.

Established The fourth is quench vulnerability written into the conductor. Low normal-zone propagation velocity means detection must be fast and dump circuits must be sized conservatively, because the tape will not spread a hot spot for you. Every magnet built from this tape inherits the problem, and mitigating it costs mass, instrumentation and design margin at the system level.

Frontier The fifth is supplier concentration, and the evidence is mixed but pointed. The industry's own 2025 supply-chain survey found 31% of fusion companies concerned about precision-manufacturing supplier availability for current needs and 63% concerned about future scaling, against $230 million of supplier capacity investment in 2024. That is a sector reporting that it is supply-tight today and expects to be supply-constrained tomorrow. The source is an industry association speaking about its members' interests and is flagged accordingly.

Frontier The sixth is that one customer class can absorb the entire world's output. One tokamak's toroidal field set at roughly 10,000 km is about two years of global capacity. The industry association's aggregate figure for fusion prototypes industry-wide is about 300,000 km — sixty years of current world output — which is recorded here as a planning number rather than a demand curve, since it aggregates machines that do not exist. What follows even from the conservative version is that any grid-scale superconducting programme would be competing for the same production lines, and that the current absence of crowding-out reflects the absence of grid demand rather than surplus supply.

Established What is not a bottleneck, stated plainly because the field's reputation says otherwise. Critical temperature is not a bottleneck for any application in this brief; 77 K is enough and 20 K is where the high-field work happens anyway. Critical field is not a bottleneck, with roughly a thousandfold headroom over grid duty. Critical current density is not a bottleneck. The material has been adequate since roughly 2010. The bottleneck is that a kilometre of it, uniform, is hard to make and has not become meaningfully cheaper per unit of current.

Established This brief depends on no unresolved physics. That is an unusual sentence in a frontier-research corpus and it is the correct one here. Every property REBCO needs in order to do the jobs this brief describes has been measured, is reproducible, and has margin. What the field waits on is industrial capability and capital.

Established It depends on thin-film deposition process engineering at kilometre scale. Yield above 500 m, reproducibility across a reel, target utilisation, and layer thickness without texture loss. These are manufacturing capabilities, and they are the binding dependency for every application downstream.

Established It depends on capital willing to build deposition capacity ahead of demand. The recent expansions were financed substantially by fusion demand and public money, not by grid procurement. That is a dependency on a particular financing environment continuing, and it is more fragile than a technical dependency because it can reverse without anything being learned.

Frontier It depends, weakly, on rare-earth and substrate supply. Hastelloy substrate, silver, and rare-earth target material are all real inputs with real supply chains, and target lead times are documented as a bottleneck. No source consulted identifies rare-earth availability as a binding constraint at current volumes, but the sixty-fold volume figure implied by industry-wide fusion aspirations has not been tested against those supply chains by anyone whose analysis this brief could verify.

Established What depends on it is the more important direction. High-field magnets are the immediate dependant, and through them the compact-tokamak approach to fusion, which is the reason this material's market exists in its current form. Superconducting infrastructure depends on it entirely and is priced by it. So does any proposal for lossless long-distance transmission, and any of the exotic-propulsion concepts elsewhere in this corpus that assume high-field magnets as an input.

Established The decisive experiments in this subject are now industrial rather than academic, and the readouts are procurement documents. The clearest single test of the position this brief takes is a price series: whether the 77 K self-field figure quoted in the peer-reviewed literature falls below roughly 50 €/kA·m in general availability, rather than in a single contracted forward delivery, before 2030. One grid project has contracted at that level for 2027 delivery. Whether that is a market price or a strategic price is exactly the open question.

Frontier The projected eightfold volume increase to 2028 is the natural experiment. If price per kilo-ampere-metre falls materially as that capacity comes online, the sticky-price argument in this brief is wrong and volume was the binding variable after all. If it stays flat, the structural explanation — that the denominator is set by deposition physics and not by scale — is confirmed under the strongest test the industry can run. This resolves within the decade without anyone building anything new for the purpose.

Established SPARC is the largest experiment ever performed on this material, whatever else it is. Eighteen toroidal field magnets, a 20 T large-bore magnet already demonstrated at 20 K in 2021, an operating field of 12.2 T, a prototype central solenoid demonstrated in November 2024, and the first of the eighteen production magnets completed and on its assembly jig as of January 2026 with the facility reported around 75% complete. Independent of whether it makes net fusion energy, it is a full-scale test of whether kilometre-class REBCO can be wound, joined, cooled, protected and operated at manufacturing volume.

Handwave The schedule attached to it. First plasma has already slipped once, from 2025 to 2026, with net energy targeted for 2027, and no first-plasma announcement was found in the sources consulted. Fusion schedules are the least reliable numbers in this brief and are recorded as intentions.

Frontier The nickelate programme has a well-defined and near-term decisive test: raise ambient-pressure zero-resistance temperature, not onset temperature. The current gap between the two — onset in the forties to sixties of kelvin, zero resistance in the single digits to low teens on ambient-pressure films — is the whole question. An ambient-pressure nickelate film with zero resistance above 77 K would be the most important result in applied superconductivity since REBCO. The intermediate results are being published at a pace of several per year, so this is a frontier with a short reporting cycle.

Frontier The thick-layer and artificial-pinning route has a measurable target. Production long-length tape sits above 300 A/cm-width at 77 K self field; the laboratory best is 1,300 to 1,500. The experiment that matters is not another laboratory record, it is a demonstration of laboratory-class current density held over a kilometre at acceptance yield. No source consulted reports one.

Established A negative result worth recording as an experiment in its own right. The LK-99 episode in 2023 ran an unplanned global replication trial in about six weeks, and it worked: the resistivity drop was traced to a structural transition in a copper sulphide impurity near 385 K, and the partial levitation to soft ferromagnetic impurity components, with failed replications from at least five independent institutions. The scientific process performed well under conditions of extreme publicity, and that is a finding about the field's error-correction, not merely about one material.

Established A REBCO coated conductor is a laminate, and understanding why it is a laminate explains most of its economics. The superconducting layer must be grown with its crystal axes aligned — textured — because current crosses grain boundaries badly in these materials; a misaligned boundary of a few degrees costs an order of magnitude in critical current. Since you cannot texture a metre of oxide directly on an untextured metal, the substrate carries a stack of oxide buffer layers whose job is to transmit a crystallographic template from a textured or ion-beam-aligned surface up to the REBCO. Above it goes silver, then electroplated copper for stabilisation and quench protection. The superconductor is one to five micrometres of a hundred-micrometre sandwich.

Established That geometry is why price per kilo-ampere-metre is sticky. Volume manufacturing improvements reduce cost per metre — faster reels, better throughput, cheaper substrate. They do not change the fact that a few percent of the cross-section carries all the current. To move the cost of current rather than the cost of tape, you must either make the REBCO layer thicker without losing texture, which is hard because texture degrades with thickness, or improve pinning so that a given thickness carries more current in field. Both are deposition-physics problems. This is the mechanism behind the flat price curve in the position section, and it is not a market failure.

Established Deposition routes differ and the industry has consolidated toward one. Pulsed laser deposition dominates the top tier by volume, with metal-organic chemical vapour deposition and reactive co-evaporation in use elsewhere. PLD's economics carry three named liabilities: target utilisation below 50%, raw-material lead times on the targets themselves, and narrow stable process windows that make reproducibility across a long reel the binding difficulty. A process that is reproducible over ten metres and marginal over a thousand is exactly the process that produces a low long-length yield.

Established Piece length is the specification that matters and the one least often quoted. Critical current is quoted per centimetre width because it is the number that flatters; piece length at that current is the number that constrains. Yield falls off above roughly 500 m. A tokamak toroidal field coil, a kilometre-class cable run, and a large research magnet all want continuous lengths, because every splice is a resistive joint that dissipates heat inside a cryostat where heat is the expensive thing. A supplier who can guarantee a kilometre at specification is selling a materially different product from one who can guarantee five hundred metres, at the same nominal price per kilo-ampere-metre.

Established Anisotropy is real and this brief cannot put a number on it. REBCO's critical current depends on the angle between the applied field and the tape's crystallographic planes, and the dependence is substantially weaker than in the older bismuth-based conductors, with artificial pinning centres weakening it further. That much is well established qualitatively. A numerical ratio between the two principal orientations at a stated temperature and field could not be verified from any source successfully consulted for this brief, and is recorded as an open gap rather than estimated. What can be said precisely is that the headline in-field figures — over 200 A per 4 mm at 20 K and 20 T — are quoted for the worst orientation, so they are a floor rather than a best case.

Frontier Thick-layer and artificial-pinning routes are the credible engineering path to better $/kA·m. Laboratory conductors combining roughly five-micrometre REBCO layers with engineered nanoscale pinning defects reach 1,300 to 1,500 A/cm-width at 77 K self field, four to five times the long-length production figure. Closing part of that gap in production would do more for the economics of every application in this brief than any of the higher-temperature materials families, and it requires no new physics. Whether it can be done at kilometre scale and at yield is the open question.

The nearest neighbour is superconducting infrastructure, which takes this material and asks what was built from it. The two briefs share sources deliberately and split them by question: the same February 2026 review supplies the price anchor here and the deployment base rate there. Read together they make an argument neither makes alone — that a material with a thousandfold performance margin has produced, in twenty-five years, on the order of ten kilometres of installed cable worldwide.

Commercial fusion is adjacent in the strong sense: it is not an application of this material so much as the institution that currently governs its economics. Anyone modelling REBCO price, capacity or piece-length availability is modelling fusion procurement whether they intend to or not.

Condensed-matter physics is adjacent in the weak sense that matters for reading the news. The nickelates and the hydrides are excellent physics and belong to a different problem than the one this brief describes; the tape industry's frontier and the high-temperature search share a vocabulary and almost nothing else.

Research integrity is adjacent because this field has become one of its standard case studies, and the ethical section below treats it as substance rather than as an aside. So is grid regulation, which appears here only as a boundary — the reason the grid does not buy this material is largely institutional and is the sibling brief's argument to make.

Established The defining institutional fact about this material is who buys it. The anchor customer for REBCO tape is a fusion programme, not a utility. That single fact propagates into every number in this brief: which piece lengths are worth qualifying, which operating points get characterised in the literature, which capacity gets built, and what price a non-fusion buyer faces. A material's specifications are shaped by its largest customer, and the largest customer here is building machines that have not yet produced net energy.

Established Public funding has followed the same signal. A national research agency runs a conductor programme scoped to fusion operating conditions, quoting about $300/kA·m commercially against roughly $10/kA·m needed for competitiveness at fields above 20 T. Eighty million dollars of United States public money went to a domestic production facility in October 2024, with construction from 2025 and over two hundred jobs — and with annual capacity in kilometres not disclosed, which is itself an institutional fact worth recording: the public financed capacity whose size it cannot audit.

Frontier Producer concentration is real and poorly documented. Four producers account for more than half of world output; one supplier reported cumulative deliveries above 5,000 km — roughly one giga-ampere-metre — to fusion customers by September 2024. Individual annual capacities are not published by any producer. So the sector's concentration can be established in aggregate and cannot be audited firm by firm, which is a weak position from which to make industrial policy and an unusually weak one for a material several governments now treat as strategic.

Established Geographic distribution is a strategic fact in its own right. Top-tier production sits in Japan, China and Russia, with the United States, Korea, Italy and elsewhere in a second tier and expanding. A material that is on the critical path for fusion, for high-field science and potentially for grid infrastructure has a supply base that does not map onto any single bloc, and the $80 million domestic-capacity award is best understood as a response to exactly that.

Frontier The institution that does not exist is a grid buyer. No utility, regulator or transmission operator is a volume purchaser of this material anywhere. The sibling brief documents why in detail — capital cost multiples, absent regulatory incentives for loss reduction, and a thin operational record — but the consequence belongs here: with no grid demand, nobody qualifies tape for grid duty cycles, nobody characterises the operating points a utility would care about, and the learning that would make grid deployment cheaper does not happen. The chicken-and-egg is not a metaphor in this sector; it is the observed state.

Established What would change the institutional picture, stated as a testable condition. A single sustained non-fusion buyer at volume — the most likely candidate on current evidence being data-centre power delivery rather than transmission — would diversify the demand signal, put a second set of specifications into the qualification literature, and give producers a reason to build capacity that does not depend on tokamak schedules. Whether that buyer appears is the institutional question this brief is actually about.

Research integrity is the live ethical question in this field, and it is not an aside. Established Superconductivity has produced two of the most consequential research-misconduct episodes of the last decade, and the field's handling of them is a legitimate object of study for anyone thinking about how frontier claims should be received.

Established The institutional record is specific and should be stated in full rather than gestured at. A ten-month investigation initiated at a federal funder's request, concluding April 2024, found numerous instances of research misconduct, characterised data manipulation as blatant and significant and as intentional fabrication, identified plagiarism in a grant application that clearly rose to the level of intentional, concluded the researcher could not be trusted, and recommended termination. Five papers were retracted, five more carry expressions of concern, and the co-authors of the final paper requested its retraction themselves. The correction worked, and it worked because named individuals — the critics who raised susceptibility-data concerns in 2021, and the co-authors who broke ranks — accepted professional cost to make it work.

Established It also worked slowly, and the delay is the ethical content. Three years elapsed between the first published critique and the institutional finding, and during those years the claims shaped funding conversations, press coverage and the expectations of people entering the field. The system corrected; it did not correct fast enough to prevent the harm that the claims themselves caused.

Frontier There is a second-order harm that is harder to see and worth naming. A field with a prominent fraud case acquires a defensive reflex, and that reflex is now applied to legitimate frontier work. The nickelate results are careful, replicated across groups and candid about their own unresolved mechanism, and they arrive into an audience trained by two retraction cycles to treat any high-temperature claim as probable fraud. Calibrated scepticism is the right response; blanket dismissal is a cost the misconduct imposed on innocent researchers, and it falls hardest on early-career people whose results are least likely to be given the benefit of the doubt.

Established A disclosure obligation this brief takes on itself. A substantial share of what is publicly known about tape capacity, delivery volumes and price trajectories comes from suppliers, fusion developers and an industry association — parties with a direct financial interest in the perception that this material is scaling and getting cheaper. Those sources are individually reliable on facts they would be embarrassed to get wrong, such as shipped kilometres, and individually unreliable on framing, such as market-leadership claims and cost trajectories. This brief marks interested parties in its reading list and relies on the peer-reviewed review literature for every price figure it quotes.

Frontier Finally, an allocation question that is not usually posed as one. Public money in this sector has followed fusion. A national research agency runs a conductor programme scoped explicitly to fusion conditions; $80 million of public funding went to a production facility whose anchor demand is fusion. That may well be correct allocation. It is nonetheless an allocation decision made largely without public deliberation, and its effect is that the specifications, price points and delivery schedules of a strategically important material are being set by the requirements of machines that do not yet work.

Established The civilisational significance of this material is not the electricity it would save. It is that a well-understood, adequately performing, commercially available material can sit for fifteen years without being deployed at scale, because the binding constraints are manufacturing yield, capital formation and procurement institutions rather than knowledge. This is a general pattern and superconductivity is one of its cleanest instances: the corpus's recurring question of what stands between a demonstrated capability and a built world is answered here by yield curves and buyers, not by physics.

Frontier Where it does become civilisationally significant, the mechanism is fusion rather than transmission. High-field REBCO magnets are what makes a compact tokamak geometrically possible, because confinement scales steeply with field and field is what the tape buys. If commercial fusion arrives on anything like its stated schedules, it arrives because this material became manufacturable, and the tape industry will have been the rate-limiting input to the largest energy transition of the century. That is a much larger claim than anything in the grid case and it rests on a programme that has slipped once already.

Speculative The transmission case remains genuinely open at civilisational scale. Lossless or near-lossless long-distance transmission would loosen a constraint that shapes where generation can be built and therefore which renewable resources are worth developing at all. The material can support it. Nothing in the deployment record suggests it is being built, and the reasons are institutional and are the sibling brief's subject.

Established A quieter significance, which the price series makes visible. Ten years of tenfold volume growth with flat cost-per-current is a measurable counterexample to the assumption that manufacturing scale reliably drives learning curves. Where the cost driver is a physical property of a deposited layer rather than a function of throughput, volume does not deliver what volume usually delivers. Any industrial policy that assumes otherwise — and several currently do — is assuming a mechanism this material does not have.

Established What already happened, because the timeline usually starts too late. REBCO coated conductors have been commercially available in useful lengths since roughly 2010; the material's adequacy is a fifteen-year-old fact, not a pending milestone. World capacity passed 5,000 km a year by 2025. A national research agency began funding fusion-specific conductor development, and $80 million of United States public money went to a single new production facility in October 2024, with construction from 2025.

Frontier 2026 to 2028: the capacity build-out and its price test. An eightfold volume increase is projected by 2028, and a 2027 grid delivery is contracted at 50 €/kA·m. This is the interval in which the sticky-price question resolves one way or the other, and it requires no technical breakthrough to resolve.

Frontier 2026 to 2027: SPARC. First plasma targeted 2026 after one slip, net energy 2027. Treat as intentions rather than dates; what is verified is that the first of eighteen toroidal field magnets was complete and on its jig in January 2026.

Frontier Late 2020s: kilometre-class piece length at yield, or not. This is the milestone that would actually change the economics of everything downstream of this brief, and it is the one nobody publishes a date for — which is itself informative, since manufacturers publish capacity roadmaps freely and yield roadmaps never.

Speculative 2030s: an ambient-pressure nickelate conductor. Getting from three-unit-cell epitaxial films to anything wound on a spool requires several unsolved steps in series — thickness, substrate independence, zero resistance well above operating temperature, and a deposition route that runs continuously. Each is individually plausible; the conjunction on a decadal timescale is speculative, and no source consulted proposes a development path.

Handwave Any date attached to ambient-pressure room-temperature superconductivity. There is no candidate material, no theoretical route to one at ambient pressure, and the field's own stated near-term target has retreated to roughly 100 K at reduced pressure. Dates in this category are not forecasts and this brief declines to supply one.

• Depends on Nothing on this map. REBCO's properties are known, reproducible and adequate with margin, so nothing the working material needs is a pending physics result. The ambient-pressure high-temperature search is a genuine open question in condensed-matter physics and is not on the critical path for any application described here — which is the cleanest statement of this topic's position on the map: a room-temperature superconductor would be transformative, and its absence is holding nothing up.
• Requires (not on this map) Thin-film deposition process engineering at kilometre scale, and the yield discipline that goes with it. Quench protection engineering for a conductor with intrinsically low normal-zone propagation velocity. Rare-earth deposition target supply, currently a documented lead-time constraint at under 50% utilisation. The capital to build deposition capacity ahead of demand. And an anchor procurement institution willing to buy at volume — today a fusion programme, not a utility, which is the single most consequential institutional fact about this material. All five are industrial, financial or institutional capabilities rather than discoveries, and their absence is what the field is actually waiting on.
• Enables High-field magnets, and through them the compact-tokamak approach to fusion. That relation runs to a programme page rather than to a Frontier Research brief, so no typed enabling edge is claimed here — the edge is real and the endpoint is not a brief.
• Adjacent Superconducting infrastructure, the sibling slot and the consumer of everything priced here; commercial fusion, the largest customer and the price-setter; exotic materials for propulsion, which assume high-field magnets as an input; and energy corridors, where superconducting transmission is one option among several and currently the least deployed.

“We are waiting for a room-temperature superconductor.” Established For the applications in this brief and the next, we are not. The material available since roughly 2010 has approximately a thousandfold field headroom over grid duty, carries over 300 A/cm-width in long production lengths at liquid-nitrogen temperature, and works at 20 K and 20 T for magnets. Nothing that has failed to be built was blocked by critical temperature. A room-temperature ambient-pressure superconductor would be a magnificent scientific result and would remove the cryogenic plant, which is real but is not, on the evidence in the sibling brief, the binding cost. Framing the field as waiting on a discovery relocates a manufacturing and procurement problem into a laboratory where it cannot be solved.

“REBCO tape costs $X per kilo-ampere-metre.” Established This is the single most common error in writing on this subject, and it is an error of omission. Four figures are in circulation and all four are defensible: $15 to $30 per metre; 80 to 100 €/kA·m at 77 K in self field; $100 to $200/kA·m at 4.2 K and 10 T; and about $300/kA·m at fusion operating conditions. They are not contradictory. They are different normalisations, because critical current depends on temperature and field, so a price per kilo-ampere-metre is meaningless without both stated. A quoted $/kA·m figure with no temperature and no field attached should be treated as uninterpretable, and this brief states conditions with every price it gives.

“Prices are falling fast as production scales.” Established Cost per metre has fallen. Cost per kilo-ampere-metre at 77 K self field has not: roughly $100 in 2022, 80 to 100 euros in February 2026, across a decade in which output rose about tenfold. The two statements are both true and they are routinely conflated, usually by quoting the falling one against the metric that governs applications. The structural reason is in the engineering section: the superconductor is a few percent of the cross-section, so scale moves the numerator and deposition quality moves the denominator.

“LK-99 was explained by a copper sulphide impurity.” Established Half right, and the half that is missing is the interesting half. It takes two separate mechanisms to explain two separate observations. The resistivity drop near 400 K is attributed to a structural phase transition in Cu₂S at about 385 K. The partial levitation is separately attributed to weak but definitive soft ferromagnetic components found in the levitating fragments — ferromagnetism, not the Meissner effect. Attributing both to Cu₂S is the standard oversimplification, and it obscures the fact that two independent groups found two different artefacts and both were right.

“LK-99 showed that science failed.” Established It showed the opposite, and this brief takes the unfashionable side deliberately. An extraordinary claim was posted, replicated worldwide within weeks by groups in at least five countries, and resolved into two identified impurity mechanisms inside about six weeks, at no cost to anyone but the people who tried. The episode is usually narrated as a hype cycle. The record is of an error-correction system working at unusual speed under unusual publicity.

“The Dias affair was two retracted papers.” Established It was five retracted papers, with expressions of concern on five more. The carbonaceous sulfur hydride claim of 287 K at 267 GPa was published in October 2020 and retracted in 2022. The lutetium hydride claim of about 294 K at roughly 1 GPa was published in March 2023 and retracted on 7 November 2023 at the request of all co-authors except Dias and two students, on the grounds that the paper did not accurately reflect the provenance of the materials, the measurements undertaken, or the data-processing applied. A ten-month institutional investigation concluding in April 2024 found numerous instances of research misconduct, described the data manipulation as blatant and significant and as intentional fabrication, found intentional plagiarism in a federal grant application, concluded that he could not be trusted, and recommended termination. He left the university on 19 November 2024. Understating this record is not generosity; it makes the field's response look weaker than it was.

“Hydride superconductors are nearly practical.” Handwave They exist at pressures above a million atmospheres inside diamond anvil cells, and no consulted source proposes a route to a conductor. The field's own near-term target has moved to roughly 100 K at reduced pressure — which, if achieved, would still need cryogenics and would still be worse than REBCO for every application here unless the pressure requirement disappeared entirely.

“Nickelates now superconduct at 96 K.” Established Under 21.6 GPa, in bulk crystals, with zero resistance at 73 K. The ambient-pressure films are a different result: onsets of 45 to 63 K with zero resistance in the single digits to low teens, three unit cells thick, on one specific substrate under compressive strain. Onset temperature and zero-resistance temperature are not the same measurement, and in this material family they currently differ by tens of kelvin. Reports that quote the onset and omit the pressure or the zero-resistance figure are the reason this family is widely believed to be closer to application than it is.

“Fusion demand is starving the grid of tape and driving prices up.” Frontier The evidence points the other way. Prices are flat rather than rising; capacity expanded roughly tenfold; a grid project obtained a below-market forward price for 2027 delivery at the peak of fusion demand; and trade reporting describes fusion driving scale-up, process improvement and quality control that non-fusion buyers now borrow. The honest counter-reading, which this brief holds simultaneously: one machine consumes about two years of world output, so the absence of crowding-out so far reflects the absence of grid demand rather than surplus supply.