Energy Corridors

An energy corridor is bulk electric power carried across a distance at which alternating current stops being the right instrument. The enabling technology is high-voltage direct current: two converter stations facing each other across a line or a cable, buying the absence of reactive charging current and of stability limits at the price of the terminals. The built unit is the interconnector — a point-to-point link joining two systems that were planned separately. The aspiration is the supergrid: a meshed, multi-terminal, multi-vendor DC network planned as one system across a continent, or across two.

This brief is about the distance between those two objects. Point-to-point HVDC has an operating record running from the 1960s and links in service across every relevant geography and sea depth. A meshed continental DC grid has never been built anywhere, and the reasons are almost entirely not electrical. The argument developed here is that the corridor is a megaproject-governance problem wearing an engineering costume: it is sited before it is designed, it is permitted before it is procured, and it is owned by nobody along most of its length.

A declared hole, stated at the front rather than buried. The research pack behind this brief was assembled under a provenance gate that closed the project-register route entirely. It therefore contains no capital cost for any HVDC scheme, no route length for any flagship link, and no cost-per-GW-km figure of any kind. That gap is named as G-14.1 and it is not filled here from memory. Section 14 explains why the central argument — that elapsed time and ownership, not capital cost, are the binding terms — is stronger without it.

Scope. Generation is Category V. Orbital and planetary-scale power collection is Planetary Scale Energy Systems. Cascading failure, hardening and restoration are Infrastructure Resilience, the sibling brief to this one. Delivery, forecasting and outturn are Megaproject Governance, on which this brief depends. What is left is the wire, the permit and the owner.

Established HVDC does not have a per-kilometre cost, and that single structural fact governs the whole subject. The cost of a direct-current link decomposes into two terms that behave completely differently: converter stations, which are largely distance-independent, and line or cable, which is distance-proportional — with cable roughly an order of magnitude dearer per kilometre than overhead line. Zeng, Hennig and Rohrig set out the converter-station topology and loss structure that this decomposition rests on (Energy Procedia 80:123–130, 2015). Established The consequence is that the governing concept is a break-even distance, not a unit price: below it AC wins because you do not pay for terminals, above it DC wins because you do not pay for reactive compensation and stability margin. Established A network with many terminals pays for many terminals, so the economics of a two-ended link do not extrapolate to a meshed grid. Frontier Where exactly the break-even sits depends on route type, voltage class, terminal count and the AC reinforcement needed at each end, and this brief cannot state it, because no verified figure for any of those inputs survived the pack's retrieval conditions.

Established The built record is real, extensive, and—in this pack—almost entirely unquantified. The engineering literature confirms the existence and identity of the link class: the Moyle interconnector between Scotland and Northern Ireland (Harvey, IET ACDC 2001); Basslink, the Tasmania–Victoria submarine link, with both a system-design paper (Westerweller, 2006) and a control-and-protection paper (Davies, 2006); COBRAcable between Denmark and the Netherlands (Tourgoutian and Alefragkis, 2017); and the Tian-Guang project, an early Chinese long-distance scheme documented as far back as 1996 (Quan). Established The single per-project parameter this brief can state with confidence comes from a paper title rather than a table: India's East–South Interconnector II is 2000 MW at ±500 kV (Prasher, IET ACDC 2001). Handwave Any reader who wants NorNed, NordLink, North Sea Link, Viking Link, Rio Madeira, Belo Monte or the 1100 kV Chinese UHVDC class quantified here will not find it, and should distrust any brief that supplies those numbers without saying where they came from.

Established Point-to-point DC is proven. Meshed multi-terminal DC is a different object, and conflating them is the field's most common technical error. Van Hertem and Ghandhari titled their 2010 review Multi-terminal VSC HVDC for the European supergrid: Obstacles (Renewable and Sustainable Energy Reviews 14(9):3156–3163), and the title states the finding. Established The named obstacle classes are DC circuit breakers, converter interoperability across vendors, and protection coordination — the three things a point-to-point link never has to solve, because it has one owner, one vendor and two ends. Frontier Whether those obstacles have since been closed is a live and answerable question that this brief cannot close: the paper's body was not obtainable, and no verified status report on multi-vendor meshed DC or on DC fault clearing was retrievable (gap G-14.2). Frontier The architecture literature has continued regardless — da Silva, Teodorescu and Rodriguez set out multilink DC concepts for supergrid futures in 2011 — but architecture papers are not commissioning records.

Established Siting and interconnection have become a distinct policy-evaluation genre, which is itself evidence about where the schedule goes. Pacific Northwest National Laboratory published a historical evaluation of United States interconnection policy under the title Waiting in Queue (Boff and Barrows, 2023) — a title chosen by people describing a queue, not a construction programme. Established Transmission-line siting appears in the governance literature as a public-dispute management problem rather than an engineering one (An and Kim, Journal of Governmental Studies, 2014). Established Permitting cost is now treated as an attributable component of levelised cost with its own method, developed for wave-energy devices by Copping, Geerlofs and Hanna (PNNL, 2014) but transferable. Speculative What this brief cannot do is give you a number. No median queue duration, no gigawatts awaiting study, no completion rate, and no permitting-delay duration for Germany or anywhere else survived retrieval (gaps G-14.3 and G-14.5). The proposition that permitting dominates schedule is treated here as a hypothesis with a stated test, not as a fact — see section 6.

Established The modelled case for a continental corridor is a genuine literature, and its flagship study is exactly the one this brief cannot quote. MacDonald, Clack and colleagues published the national-HVDC-macrogrid case in Nature Climate Change 6:526–531 (2016); it is the most-cited quantitative argument for a United States supergrid and it is the source of most downstream advocacy. Speculative Its headline cost and emissions results are unverified in this pack (G-14.4) and are therefore not stated. Established What can be stated is the shape of the surrounding literature. Allard and colleagues frame European grid expansion as a flexibility option — one among storage, demand response and curtailment (Energy Economics 87:104733, 2020), which is the correct comparative framing and the one most supergrid advocacy omits. Established National cost–benefit work on grid expansion exists at country scale, and Germany is the standing natural experiment — high renewable penetration, a north–south corridor requirement, and a permitting regime with a reputation (Gunkel and Möst, EEM14, 2014). Speculative The delay durations that gave it that reputation are unverified here (G-14.5), so this brief describes the German case without quantifying it. Frontier Expansion planning has moved to probabilistic and stochastic formulations as standard — Jadidoleslam, Ebrahimi and Latify maximise wind integration probabilistically (Renewable Energy 114:866–878, 2017), and Ndlela's 2025 doctoral thesis combines tie-open-point optimisation with genetic algorithms, hosting-capacity assessment, quasi-dynamic simulation and quasi-Monte Carlo probabilistic analysis. Frontier That last item matters for a second reason: expansion planning is entering the literature from sub-Saharan Africa on its own terms, not as a case study appended to a European model.

Established The named supergrid programmes are all documented, and all unbuilt. DESERTEC has a technical originator, Franz Trieb of DLR, whose Europe–Middle East–North Africa cooperation account appeared in 2021; its transmission architecture descends from Gregor Czisch's least-cost European–North African optimisation, traceable here only through Boubaker, Colantoni and Allegrini, who write of coupling eastern North African wind into a Czisch-style European HVDC super grid on a 2025 horizon (IJRED 2(2):125–129, 2013). Speculative Czisch's own thesis was not locatable through the pack's instruments (G-14.7), which is worth stating: the intellectual foundation of the most famous supergrid proposal in the world is, on this pass, a work cited but not retrieved. Established The North-East Asian Super Grid — Russia, Mongolia, China, Korea, Japan — was developed technically by Bogdanov and Breyer (Energy Conversion and Management, 2016). Speculative The institutional vehicle most associated with the global version, GEIDCO, produced no retrievable record at all (G-14.8): a major actor is simply absent from the evidence base, and that absence should temper any confident account of what the global-grid programme currently is. Established The strong form has a recent statement of its own in Ivens, Akroyd and Kraft's SSRN paper, whose title carries the whole intuition: The Sun Always Shines Somewhere (2025). Frontier And the hostile position is serious rather than fringe: the chapter Super-grids as pseudo-progressive brakes (2013) argues that DESERTEC and the North Sea project were centralising manoeuvres that foreclosed distributed alternatives.

Established Corridors are the crux of the 100%-renewables dispute, and that dispute is formal, live, and in PNAS. Jacobson, Delucchi, Cameron and Frew coupled weather simulation with power-demand modelling to argue for low-cost grid reliability at 100% renewable penetration across the continental United States without stationary battery storage, prioritising thermal storage in soil and water (PNAS 112(49):15060–15065, 2015). Established Clack, Qvist, Apt and eighteen further authors published a formal evaluation rejecting it (PNAS 114(26):6722–6727, 2017). Frontier The specific technical objections are unread here (G-14.9); what is on the record is that twenty-one authors thought the rebuttal worth co-signing, and that the disagreement is substantially about how much transmission the model is allowed to assume. Established The Jacobson line has continued to publish successor studies, and the dispute has not converged.Established The most useful verified number in this whole topic comes from the same group and is not about transmission at all. Across 149 countries, adding firebrick thermal storage for industrial process heat reduces battery capacity by about 14.5%, hydrogen production by about 31%, heat-storage needs by about 27.3%, wind capacity by about 1.2%, land use by about 0.4% and annual energy costs by about 1.8% against a no-firebrick baseline (PNAS Nexus 3(7):pgae274, 2024). Frontier That is a clean demonstration of the substitution structure that governs corridor value: a corridor's benefit is a residual, and every other flexibility option that improves shrinks it. The complementarity argument supergrid advocacy deploys at continental scale — that wind and solar together cover demand more constantly than either alone — is demonstrated by Barber, Boller and Nordborg at microgrid scale in Switzerland (2019). Whether those gains are super-additive or sub-additive in distance is the actual question, and it is open.

Established A corridor delivers power into a system that has to stay standing while it arrives, and the operability of that receiving system is now the part of the problem changing fastest. A synchronous machine holds the voltage waveform up by physics: its rotor carries inertia, its field produces a stiff internal voltage, and it delivers several times rated current into a fault without being instructed to. An inverter does none of that unless it is programmed to, and even then it is current-limited to something close to its own rating. Established The default control mode of nearly every converter installed to date is grid-following: a phase-locked loop measures the voltage waveform somebody else is holding up, and the converter injects current in phase with it. Established A population of grid-following converters therefore cannot be the entire population, because something in the system has to define the waveform they are all following. Frontier Grid-forming control inverts the relationship: the converter regulates its own internal voltage phasor and lets current fall out of the impedance between it and the network, which is what a synchronous machine does — at a fraction of the fault current, and with the thermal overload envelope of a semiconductor rather than of a steel rotor.

Established System strength is not inertia, and the two are conflated constantly, including by people who ought to know the difference. Inertia governs the rate of change of frequency after a loss of generation; it is a seconds-scale energy-balance quantity. System strength governs how far the voltage waveform moves when something injects current into it, is measured as a short-circuit ratio at a connection point, and lives on a sub-cycle timescale. Established A battery can supply synthetic inertia and leave a connection point just as weak as it found it, because the fault current that sets short-circuit ratio is a property of source impedance rather than of a control loop. Frontier Below some site-specific short-circuit ratio, grid-following converters go unstable: the phase-locked loop and the network impedance close a loop around each other, and the loop gain rises as the network weakens. Frontier The remedies actually in service are synchronous condensers — frequently retired generators with the prime mover removed — and minimum-synchronous-unit constraints dispatched out of merit order. Speculative Both are costs that a corridor's benefit case has to carry and almost never does. A link justified on the spread between two price areas is quietly assuming that the receiving area can absorb the injection without buying strength, and no published corridor appraisal this brief can point to prices that assumption.

Established The evidence here is not modelling. It is an event record, and it is unusually well documented because system operators are obliged to publish. The South Australian black system of 28 September 2016 was not caused by wind generation being present; it was caused by fault-ride-through protection settings inside the installed wind fleet that withdrew after a set number of voltage disturbances — a behaviour the operator did not know was in its fleet until the fleet demonstrated it. Established The Blue Cut fire disturbance in southern California in August 2016 removed a large block of solar generation that was never itself faulted, because inverter controls misread frequency during the fault and disconnected on a protection function that was doing exactly what it had been configured to do. Established The Great Britain event of 9 August 2019 took an offshore wind farm and a gas unit off within seconds of a lightning strike, drove system frequency to 48.8 Hz and shed demand automatically. Established ERCOT's Odessa disturbances in 2021 and again in 2022 each removed more than a gigawatt of solar output through inverter responses that were, once more, as configured. Frontier The pattern across all four is identical and it is not a pattern about renewables: the fleet did what its settings said, and nobody held an accurate model of what the settings said.

Established There is a class of transfer capability that requires no new right of way, and on this brief's own argument it should be the first thing tried. If the binding term is elapsed time, and the largest component of elapsed time is siting and consent, then any instrument that raises the throughput of an existing corridor without touching the right of way skips the part of the schedule that no amount of engineering shortens. Established Four such instruments are in commercial service and carry distinct bodies of evidence: dynamic line rating, topology optimisation, advanced power-flow control, and reconductoring with advanced conductors. They are collectively marketed as grid-enhancing technologies, which is a procurement category rather than a technical one, and the four should be assessed separately because they fail differently.

Established Dynamic line rating is the cheapest of the four and the most revealing about how ratings are made in the first place. An overhead conductor's limit is a temperature, and temperature is set by current, ambient air, solar heating and wind cooling; the static seasonal rating a line is operated to assumes a conservative combination of all four, in effect a hot still afternoon. Established Measuring the conditions that actually obtain — with line-mounted sensors, tension or sag monitors, or by modelling from local weather — usually reveals headroom, and the headroom is largest when the wind is blowing, which is exactly when a wind-exporting corridor most needs it. Frontier Reported uplifts cluster in the tens of percent and occasionally run much higher, but the distribution matters more than the mean and is rarely published: a rating that sits above static most of the year and below it on the worst afternoon hands the operator a new risk on the day the system is tightest, and that asymmetry, not the cost of a sensor, is what adoption actually turns on. Established The United States moved the floor rather than the ceiling: FERC Order No. 881 requires transmission providers to use ambient-adjusted ratings, which is the weakest form of the idea and was chosen because it is the form that could be made mandatory. Frontier Full dynamic rating remains discretionary in most jurisdictions, which tells you where the obstacle is.

Established Topology optimisation and advanced power-flow control exploit the same fact: power in a meshed alternating-current network goes where impedance sends it, not where the operator wants it. A corridor is rarely limited by its own thermal capacity; it is limited by the loading that its flow redistributes onto some other element. Established Topology optimisation searches the switching state of the network for reconfigurations that move flow off the binding element, at the cost of a breaker operation and a study. Established Advanced power-flow control does the same thing with hardware — modular series-injection devices that add or subtract effective impedance on a circuit and push flow onto an underused parallel path — and is in transmission-scale service in Britain and the United States. Frontier Both share an evaluation problem this brief has met before: the benefit is a reduction in congestion cost, congestion cost is a counterfactual dispatch, and a counterfactual is not an outturn. Handwave There is no published register of topology-optimisation or power-flow-control installations reporting measured before-and-after congestion rent, and until there is, the savings claims in this class are vendor arithmetic of unknown bias — which is precisely the criticism this brief makes of supergrid cost estimates, applied evenhandedly.

Frontier Reconductoring is the largest of the four and the one that most directly substitutes for a new corridor. Advanced conductors — composite-core and other high-temperature low-sag designs — carry roughly twice the current of the conventional steel-reinforced aluminium conductor they replace at comparable sag, because they can run hotter without stretching as far. Frontier A 2024 Lawrence Berkeley National Laboratory analysis argued that reconductoring existing lines could deliver a large share of the additional transfer capability the United States needs, at materially lower cost and in materially less time than new build, and it is the most-cited quantitative statement of the case. Speculative Its headline cost and schedule ratios are not verified in this pack and are not repeated here as fact; what survives without them is the structural claim, which is that reusing a right of way deletes the schedule term that dominates. Established The engineering caveats are real and are routinely omitted by advocates: the existing structures have to take the new tension and the new thermal loads, a line that is voltage- or stability-limited rather than thermally limited gains nothing from a hotter conductor, and the outage needed to restring a circuit is itself a security-of-supply problem on a corridor that is already the constraint.

Frontier The live question is no longer whether a continental corridor can be built but whether its value survives everything else getting better. Treat corridor benefit as a residual computed net of storage, demand flexibility, thermal storage and local complementarity, and the residual is a moving target under a moving frontier. Jacobson and colleagues quantified one such displacement precisely in 2024; Allard and colleagues framed the comparison correctly in 2020. Speculative Nobody appears to have published the obvious study: the same supergrid model re-run across successive vintages of storage cost curves, reporting whether marginal corridor value rises or falls. That is cheap, requires no new data, and would settle a decade of argument.

Frontier Load is starting to move to power rather than power to load. The United States Department of Energy's 2026 report Large Load Planning, Siting, and Interconnection marks the shift: datacentre demand is reshaping the problem from “connect remote generation to distant cities” into “site the load next to the generation,” which is a different corridor architecture and in the limit is no corridor at all. Speculative Firebrick process heat is the same move made for industrial heat. Speculative The unmeasured quantity underneath this is the share of total demand that is genuinely location-flexible. It is plainly not zero and plainly not one, and nobody has published it.

Frontier Interconnectors are increasingly justified on resilience rather than on economic dispatch, and that justification is currently unauditable. COBRAcable's design paper was presented at a conference on the Resilience of Transmission and Distribution Networks (2017) — a venue choice that is itself data about how the class is being sold. Speculative The sibling brief on Infrastructure Resilience documents why this is hard to check: resilience lacks an agreed measurable, so a resilience justification cannot be falsified by an outturn. Speculative The sharp version of the worry is that interconnectors are sold on security of supply and bought on price-differential arbitrage, and that the two diverge under exactly the correlated-stress conditions where the security case is supposed to pay.

Frontier Cross-border capacity is being reframed as a common-pool resource rather than a private asset. The framing arrives from the gas side — Transmission Capacity as a Common-Pool Resource (International Symposium for Next Generation Infrastructure, 2015) — and it is the sharpest available lens on corridor under-provision. Frontier Institutional economics entering a domain previously treated as pure engineering is the most important conceptual development in this topic in fifteen years, and it is the one least represented in the modelling literature.

Frontier NorthConnect is the rarest thing in infrastructure research: a fully engineered project that did not happen, with a published design record. Gleadow and Love presented the Scotland–Norway link at IET ACDC in 2019. It did not proceed. Speculative The cause is unverified here (G-14.10), and the reason that matters is methodological: negative outcomes with engineering documentation are the only way to separate “could not be designed” from “was designed and still failed.” A field that only studies its completions cannot learn what kills things. Frontier Smith, McIntosh and Monkhouse's developer-side account of what determines interconnector project success (IET ACDC 2015) is the closest thing to a systematic treatment, and it is one paper.

Speculative The deepest frontier question is one nobody is funding: what actually killed each supergrid proposal. The popular account of DESERTEC's death — the Arab Spring plus European fear of dependence on North Africa — runs into the fact that Lilliestam and Ellenbeck put the energy-weapon question into Energy Policy in 2011, before the collapse. Speculative A competing account is simply that the cost premium evaporated: European photovoltaics fell fast enough that generating in the Sahara and shipping the power thousands of kilometres stopped beating generating locally. Speculative Both accounts are unverified here. Handwave A coded register of twenty or more proposed intercontinental corridors with cause-of-death, tested for whether causes cluster, does not exist and would be the single most tractable original contribution anyone could make to this subject.

Frontier Grid-forming control has moved from demonstration to procurement, and the interesting question is no longer whether it works but who is made to buy it. Grid-forming batteries are in service at transmission scale in at least Australia, Britain, Ireland and Hawaii, and system operators in weak-grid jurisdictions have begun writing grid-forming capability into connection requirements rather than tendering it as an ancillary service. Frontier The specification problem is genuinely hard and is being worked in the open: a performance definition has to say what the converter does in the first cycle after a disturbance without dictating the control law, or it becomes a single vendor's product description. Speculative Black start is the same argument arriving on the day it matters most. Restoration has historically been anchored on hydro and on gas units with diesel starters, energising a dead network into which everything else is resynchronised; a grid-forming battery can in principle perform that role, and demonstrations at distribution scale exist. Speculative Whether an inverter-anchored restoration path has been proven at bulk-system scale anywhere — against a genuinely dead network rather than a sectionalised rehearsal — is not settled in this brief's evidence base, and the thing that would settle it is a published restoration trial report rather than another simulation.

Frontier The sharpest unexamined question in this class is whether grid-enhancing technologies are a substitute for a corridor or a complement to one. The advocacy frames them as a bridge: work the existing network harder while the new line is being consented. Speculative The uncomfortable possibility is that they are a substitute good — that raising the utilisation of the existing network lowers the congestion rent which is the new line's revenue, and that a programme of uprates can therefore make a corridor's business case worse while making the system better. Speculative That is the residual-value argument of this section arriving from inside the wires rather than from storage, and no published study appears to have run it: the test is a single model re-solved with and without a full uprate programme, reporting marginal corridor value in each case.

Established Three things block a continental supergrid, and only one of them is technical. First, meshed multi-vendor DC carries a published, named obstacle list that has not been demonstrably closed. Second, cross-border capacity is a common-pool resource with no owner whose mandate spans it. Third — and this is the one nobody names — there is no published outturn reference class for HVDC, so the cost of the thing is not a forecastable quantity in the sense that Megaproject Governance requires.

Frontier The technical bottleneck is specific and small, which is what makes it interesting. It is not power electronics in general; it is DC fault clearing in a meshed network, protection coordination across a network whose fault currents rise in milliseconds, and control interoperability between converters supplied by different vendors. Established A point-to-point link clears a DC fault by shutting down; a meshed grid cannot, because shutting down is the failure mode it exists to avoid. Speculative Whether a commercially operating multi-vendor meshed DC network with functioning DC fault clearing now exists anywhere is unresolved in this pack and should be checked before anyone asserts either way.

Speculative Underneath the technical bottleneck sits a physical supply bottleneck that institutional reform cannot relieve. HVDC converter manufacturing and high-voltage cable manufacturing are concentrated in a very small number of suppliers, and cable-laying vessels are a hard, countable constraint on how much submarine link the world can install per year. Handwave This brief has no verified capacity data for any of the three (G-14.11), which is itself a finding: the physical rate limit on the energy transition's most-modelled infrastructure is not in the public literature.

Established The binding term, on the argument this brief makes, is elapsed time rather than capital. Not because capital is cheap — it is not — but because a corridor's cost of capital is a function of how long it is held before it earns, and the interval between proposal and energisation is dominated by processes that no amount of engineering shortens. Frontier A link whose consent, litigation and cross-border commercial arrangements take a decade has burned a decade of financing charges before a single tower is raised, and the sensitivity of the whole business case to that interval is larger than its sensitivity to the unit cost of conductor. Speculative The honest form of that claim is a hypothesis with a test attached, and the test is in section 6.

Speculative One bottleneck is epistemic and it is the one this brief keeps returning to. Nobody can currently say whether corridors overrun, by how much, or whether line and converter behave differently. Established The most-modelled infrastructure of the energy transition has no published outturn reference class, and until it does, every argument about whether a supergrid is affordable is being conducted between two sets of estimates neither of which has ever been scored.

Established The operability bottleneck that nobody can route around is model availability, and it is a commercial bottleneck wearing a technical costume. The planning studies that decide whether a corridor is adequate are run in the phasor domain, at fundamental frequency, on positive-sequence models; converter control interactions happen faster than that representation can see them. Established Electromagnetic-transient simulation resolves the actual waveform at microsecond steps and can therefore reproduce control interaction, but it needs manufacturer models of the specific control code in the specific firmware build, and those arrive as encrypted black boxes when they arrive at all. Established The United States Department of Energy convened a workshop on precisely this in 2024, which is itself the diagnosis: the method exists, and the models do not circulate. Handwave No published figure states what share of installed converter capacity anywhere carries an electromagnetic-transient model validated against a measured field response. That number would be the most informative single statistic in the subject and this brief cannot supply it.

Established Protection is the operability problem that receives least attention and is hardest to retrofit. Distance and directional-overcurrent relays infer where a fault is from the magnitude and phase of the current feeding it, and both of those are physics when the source is a rotor and firmware when the source is a converter. Established A current-limited source with a controlled phase angle and no guaranteed negative-sequence injection can leave a relay unable to see a fault, unable to tell which direction it lies in, or able to do both but too slowly to matter. Frontier IEEE 2800, published in 2022, specifies interconnection and interoperability performance for inverter-based resources on transmission systems, including ride-through and unbalanced-fault behaviour, and is the right instrument; it binds only where a connection agreement or a regional reliability rule adopts it, which is a jurisdiction-by-jurisdiction process rather than a switch.

Established The reason the four are not already universal is regulatory rather than technical, and it is the cleanest incentive story in this brief. A transmission utility under cost-of-service regulation earns an allowed return on the capital it places in the rate base. Grid-enhancing technologies are cheap, several of them behave more like operating expense than capital, and a device that defers a line earns the owner less than the line would have. Frontier The regulatory instruments that would fix this exist in draft rather than in force: shared-savings incentives that pay a share of demonstrated congestion relief, and planning rules that require alternative transmission technologies to be evaluated against a new-build baseline instead of assumed away. Frontier FERC's long-term regional transmission planning order requires such technologies to be considered in planning, which is a procedural obligation rather than an economic one. Speculative Whether consideration has changed any outcome is measurable, has an obvious metric — the share of approved transmission projects that are uprates rather than new build — and is not, so far as this brief can establish, being published anywhere.

Established This brief waits on a governance result before it waits on an electrical one. The corridor literature quotes engineering estimates. Megaproject Governance establishes that engineering estimates for projects of this class are systematically low, that the mechanism is at least partly deliberate rather than cognitive, and that the reference class rather than the estimate is the right forecasting object. Established The closest large-n energy analogue is the hydropower outturn study of Ansar, Flyvbjerg, Budzier and Lunn (Energy Policy 69:43–56, 2014); its specific figures are unverified in this pack, but its existence establishes that an energy-megaproject reference class has been built once, for dams, and never for wires.

Established The second dependency is methodological and it links this brief to its sibling. Hines, Cotilla-Sanchez and Blumsack's finding that topological network metrics correlate only mildly with physics-based power-grid models (Chaos 20(3):033122, 2010) constrains what a corridor-security argument may be built from. Frontier Any claim that a meshed continental grid is more or less fragile than a set of separate systems, if it rests on graph topology alone, is resting on a proxy the best available test says is weakly informative. Established Paul Hines is also among the twenty-one authors of the Clack rebuttal, and the same commitment — physical models over convenient abstractions — runs through both. That is not a coincidence; it is a research programme.

Speculative The third dependency is a single unread conclusion. Whether Lilliestam and Ellenbeck answered the energy-weapon question in the negative determines whether the popular account of DESERTEC's death is sufficient or is a myth (G-14.6). Frontier One paper, read, would settle the framing of an entire programme's obituary — which is an unusual position for a research dependency to be in, and worth noticing as a fact about how thin the evidentiary base under a famous story can be.

Speculative The fourth is a dataset rather than a result, and it is the one this brief most wants. A corridor register — proposed intercontinental schemes with approval-stage estimates, ownership structure, and cause of death where they died — would simultaneously supply the reference class, test the ownership-concentration hypothesis, and adjudicate the permitting claim. Established No science is required to produce it. That is what makes its absence a governance finding rather than a research gap.

Established The workback plan from “impossible” to a delivered continental supergrid runs seven links, and five of them are institutional. The target state is precise: a meshed, multi-terminal, multi-vendor HVDC network spanning a continent, planned as one system, with a single capacity-allocation regime, delivered inside its approval estimate.

Established Link 1 — build the HVDC outturn reference class. A public register of at least 100 completed HVDC projects carrying approval-stage estimate, outturn cost, route length, terminal count, route type and schedule. The measurement is the overrun distribution disaggregated by route type, and specifically whether converter stations and line behave differently. Established This binds first because every downstream cost claim is unfalsifiable without it. The constraint is institutional, not scientific: the data exists inside transmission system operators and vendors and is commercially sensitive. Frontier Link 2 — decompose the corridor timeline. Median elapsed time from proposal to energisation, split into permitting, litigation, procurement, construction and commissioning, across at least three jurisdictions. The measurement is the ratio of permitting-plus-litigation to procurement-plus-construction. This adjudicates the claim that currently drives most corridor policy on assertion alone.

Frontier Link 3 — close the meshed multi-terminal obstacle list. A demonstration network of at least three terminals and at least two converter vendors, with DC fault clearing and coordinated protection, in commercial operation. The measurement is fault clearance within specification without collapsing the DC network, and interoperability without vendor-specific control. Established This is the only genuine industrial-capability link in the chain, and the supply constraint beneath it is physical. Speculative Link 4 — a capacity-allocation and cost-recovery regime spanning the route. A rule set under which a jurisdiction crossed by, but not terminating, a corridor is compensated, and under which capacity rights survive a national curtailment override. The measurement is revealed behaviour in a correlated stress event: does the corridor deliver across a border when the exporting system is itself short?

Speculative Link 5 — corridor value as a residual under a moving substitute frontier. One supergrid model, re-run across successive vintages of storage, demand-flexibility and thermal-storage cost curves; report the sign and magnitude of the change in marginal corridor value per vintage. Analytically binding, and cheap. Speculative Link 6 — a single planning entity with balance-sheet capacity and a route-spanning mandate. Speculative Link 7 — reference-class-adjusted supergrid cost estimates entering the policy literature, which is downstream of Link 1 and cannot precede it.

Established Link 4 binds hardest and Link 1 binds first. Link 3 is the only link where the world would have to learn something it does not know; everything else is a matter of choosing to provide what already could be provided. The honest terminal position is that the supergrid is blocked neither on physics nor on money, but on the absence of an entity whose interests span the route and the absence of a dataset that would let anyone say what it costs.

Established The converter station is the engineering object that matters, because it is where the money and the constraints concentrate. Its topology and its losses determine both the break-even distance and the feasibility of a meshed architecture, which is why Zeng, Hennig and Rohrig's power-flow treatment of converter topologies and losses is the right entry point rather than any line-design paper. Established There is a formal loss-determination standard for HVDC converter stations, which tells you the industry considers converter losses a contractual quantity rather than a modelling assumption.

Established Three engineering requirements separate a supergrid from a set of interconnectors. DC circuit breakers capable of interrupting a fault in a network that has no natural current zero; protection coordination across a system whose fault transients propagate faster than conventional protection philosophy assumes; and control interoperability between converters from different manufacturers, which is a specification and conformance problem as much as a hardware one. Frontier All three were named as open in 2010 and none has a verified closure date in this pack.

Established Cable is the other hard engineering term, and it is where geography stops being negotiable. Submarine routes avoid every siting objection on land and pay for it in cost per kilometre, installation rate and repair time. Frontier Basslink's separate control-and-protection paper exists precisely because a long submarine link is a different control problem from an overhead line of the same rating. Speculative A derated link — energised but limited by works incomplete at the far end — is a normal outcome for a project whose two ends are governed by different authorities, and it is the failure mode a single-owner project does not have.

Frontier Planning tooling has caught up faster than delivery has. Probabilistic and stochastic expansion planning is now standard, and Ndlela's thesis is representative of the current toolkit: tie-open-point optimisation with genetic algorithms, hosting-capacity development, contingency reliability assessment, quasi-dynamic simulation and quasi-Monte Carlo probabilistic analysis. Speculative The gap between planning sophistication and delivery performance is the diagnostic fact of the field.

Established Storage is the corridor's principal competitor, not its complement, and the accounting is explicit. Every improvement in batteries, hydrogen, thermal storage or demand flexibility reduces the residual value of long-distance transmission. The firebrick result quantifies one such displacement across 149 countries with six separate reductions reported, and no comparable exercise has been run for transmission itself.

Frontier Local complementarity is adjacent at the opposite end of the scale. Barber, Boller and Nordborg show wind and photovoltaic combination producing more constant coverage than either alone at Swiss microgrid scale. Speculative The supergrid argument is that this effect keeps growing with distance because weather decorrelates; the counter-argument is that the cheapest decorrelation is already captured locally and the marginal decorrelation bought by the last thousand kilometres is small. Both are stated; neither is measured against the other.

Frontier Datacentre siting is the adjacent technology that could make the subject partly obsolete. If a meaningful fraction of new load is location-flexible, the cheapest corridor is the one you do not build. Speculative The 2026 Department of Energy large-load report is the leading edge of this and it is very new, which means the corridor literature has not yet absorbed it. Speculative Note the direction of the effect is not obvious: flexible load could equally justify more transmission, by making remote generation sites viable that no corridor would otherwise reach.

Established Gas interconnection supplies the institutional analogue. The common-pool-resource treatment of interconnector capacity was developed for gas and transfers cleanly, which is a reminder that the hard part of electricity corridors is a problem other network industries solved, badly, first. Frontier And Infrastructure Resilience is adjacent in the strongest sense: a corridor is simultaneously a resilience instrument and a new coupling between systems that were previously independent, and the same brief has to hold both.

Established The institutional diagnosis is the whole brief in one sentence: costs fall along the route and benefits accrue at the endpoints, and no entity's mandate spans both. That is the classic under-provision structure of a common-pool resource, and the 2015 treatment of interconnector capacity in exactly those terms is the sharpest framing available.

Frontier The corridors that got built have concentrated ownership; the ones that did not were consortia. DESERTEC and the North-East Asian Super Grid were association and consortium vehicles rather than balance-sheet owners. Chinese ultra-high-voltage deployment happened under a single vertically integrated, state-backed owner. Speculative The testable form of this is that supergrid completion rate should correlate with ownership concentration rather than with the strength of the resource case or the cost estimate. It is testable against a corridor register, and the register is precisely what does not exist.

Speculative A capacity-allocation regime is the hardest institutional object in the chain because it must survive a bad day. A rule that allocates capacity fairly in normal conditions and is overridden by national curtailment powers in a shortage is not a regime; it is a fair-weather contract. Established The regime has to bind when the exporting jurisdiction's own voters are cold, which is a constitutional commitment rather than a commercial one. Speculative Nothing in the located literature describes such a rule operating and being tested.

Frontier Route compensation is the neglected half of the same problem. A jurisdiction crossed by a corridor but terminating none of it receives the servitude and none of the trade. Speculative Until a mechanism compensates that jurisdiction in something other than goodwill, the rational strategy for any transit state is to hold the project hostage at the moment its sunk cost is highest — which is a fair description of how several proposed corridors have behaved.

Speculative Two institutional facts about the evidence base itself belong here. The dominant institutional actor in global-grid advocacy produced no retrievable record in this pass, which means any account of its current programme is uncorroborated. Established And roughly half the power-systems literature in this subject is published by houses that deposit no abstracts, so a machine-assembled evidence base on corridors will systematically over-represent the publishers that deposit and under-represent the engineering literature that does not. Frontier That is a real fact about how citation corpora get built, and it applies to any assessment report assembled the way this one was.

Established The distributive structure of a corridor is unusually stark: a line imposes visible, local, permanent costs on people who receive none of its benefits. That is why transmission siting appears in the governance literature as a public-dispute problem, and why treating opposition as irrational obstruction misdescribes it. The objection is usually correct on its own terms.

Frontier The centralisation critique deserves to be taken seriously rather than dismissed. The 2013 pseudo-progressive brakes chapter argues that DESERTEC and the North Sea project functioned as centralising manoeuvres that foreclosed distributed alternatives, and that the electricity was not the point. Speculative The mainstream reply is that continental aggregation is a physical resource-diversity gain available by no other means, and that scale is not by itself a political programme. Frontier Both can be true: a real physical gain can be pursued in a form that concentrates control, and the choice of form is not settled by the physics.

Frontier The dependency objection has a peer-reviewed treatment that predates the popular version of it. Lilliestam and Ellenbeck asked in 2011 whether importing electricity would expose Europe to an energy weapon. Speculative The general argument in that direction — that electricity is harder to weaponise than gas because generation is dispersed and substitutable — is not verified here, and this brief does not assert it. What it does assert is that the objection was examined before the project collapsed, so “Europe feared dependence” is at best an incomplete obituary.

Speculative And there is an intergenerational asymmetry. A corridor is a hundred-year asset approved on a twenty-year appraisal by people who will not be accountable at its outturn. That is the same structural problem Megaproject Governance identifies, and corridors have it in a stronger form because the route communities are permanent and the sponsors are not.

Speculative The strong civilisational claim is compact: the sun always shines somewhere, so a sufficiently large grid never has a supply problem. Ivens, Akroyd and Kraft put exactly that title on the energetic-feasibility question in 2025, and the underlying intuition is why every global-grid proposal since Buckminster Fuller has been proposed. Frontier Energetic feasibility, though, is the easiest of the tests. A global grid must also be owned, allocated, protected and repaired.

Speculative An intercontinental grid would be the first physical infrastructure requiring standing cooperation between states with no history of it. The North-East Asian proposal is the clearest instance: Japan, Korea, China and Russia interconnected is a technical problem sitting on top of a trust problem, and the trust problem is not downstream of the technology. Handwave A world with a functioning intercontinental grid is a world that has already solved something else, and the grid is the evidence rather than the cause.

Speculative The counter-civilisational reading is that corridors are a coupling risk dressed as a resilience gain. Joining systems that were independent creates a shared failure surface, and the literature that would tell us how much is the one the sibling brief on Infrastructure Resilience shows is weakest exactly where it is most confident. Frontier The honest statement is that continental interconnection trades a class of local shortage risk for a class of correlated systemic risk, and nobody has priced the trade.

Handwave At the far end, the corridor is the terrestrial half of a question Planetary Scale Energy Systems asks in orbit. Any planetary-scale collection scheme needs planetary-scale distribution, and every argument in this brief about ownership, allocation and route compensation reappears there in a harder form, with no jurisdiction to appeal to at all.

These horizons track institutional capacity and delivered kilometres, not converter technology, which is the term least likely to be limiting.

• 10 yr: Frontier More point-to-point HVDC, in quantity, mostly submarine and mostly bilateral, because that is the form a two-party contract can carry. Speculative A published HVDC outturn reference class is achievable in this window and is the single highest-value thing anyone could deliver; it requires no science and one willing regulator with subpoena-grade curiosity. Speculative A three-terminal, two-vendor meshed demonstration with DC fault clearing is plausible. Frontier A continental meshed grid is not, and no serious proponent claims otherwise. Frontier Most of the transfer capability actually delivered in this window will come from uprating existing rights of way — dynamic ratings, power-flow control and advanced conductors — rather than from new corridors, because that is the only route whose schedule is not governed by consent. Speculative Expect grid-forming capability and validated electromagnetic-transient models to become connection conditions rather than procurement options in the weak-grid jurisdictions first.
• 25 yr: Speculative A regional meshed DC network — North Sea scale, several terminals, one regulatory framework, one currency of settlement — is the realistic near-term form of the supergrid and the test case for everything larger. Speculative Whether it is worth building will depend on the residual-value question, which by then should have an answer if anyone runs the vintage series. Handwave A cross-Mediterranean link at DESERTEC scale requires an owner that does not currently exist and cannot be constituted by consortium.
• 50 yr: Speculative A continental grid with a single capacity-allocation regime is conceivable within one political union and hard to imagine across several. Speculative The binding term remains a rule that holds during a shortage rather than a rule that holds during a surplus. Handwave Intercontinental links would by then be technically routine and institutionally novel, which is the reverse of how the subject is usually described, and the reverse of how it is usually funded.
• 100 / 250+ yr: Handwave A global grid, if it exists, will be the artefact of a political settlement rather than an engineering programme, and its existence will be evidence about governance rather than about power electronics. Handwave The alternative future is equally coherent, gets far less attention, and would look like success rather than failure: location-flexible load, cheap local storage and demand that follows generation make long-distance transmission a niche instrument, and the corridor era turns out to have been a hundred-year transitional technology that solved a problem the next century did not have. Handwave Nothing in the evidence base discriminates between these two futures, and anyone who tells you it does is quoting a model, not a measurement.

• Depends on This brief depends on Megaproject Governance, and the dependency is load-bearing rather than decorative. A corridor is sited, permitted and financed long before it is built, and the numbers that decide whether it proceeds are approval-stage estimates of exactly the class that VII-19 shows are systematically low and at least partly deliberately so. Three specific results transfer. First, the reference class rather than the estimate is the right forecasting object — and no reference class exists for HVDC, which is why this brief can state no cost. Second, the closest large-n energy analogue is hydropower (Ansar and colleagues, 2014), a class whose outturns were measured only after someone chose to measure them. Third, VII-19's finding that a governance reform can move outturns without improving forecasts means a corridor programme can get cheaper without anyone getting better at predicting it, which changes what “success” would even look like. Read the other way: if VII-19 is right, every published supergrid cost estimate is an engineering estimate that has not been reference-class adjusted, and the entire modelling literature is quoting a number of unknown bias.
• Requires (not on this map) Three constraints, none of them a research result and none of them produced by any brief on this map. The third is the one section 4 already names and cannot quantify: converter and high-voltage-cable manufacturing concentrated in a handful of suppliers, and a countable fleet of cable-laying vessels that sets a hard annual ceiling on installed submarine kilometres regardless of how many links are consented. Gap G-14.11 records that no verified capacity figure for any of the three was retrievable. The other two constraints arrive with the operability and right-of-way material and neither is a research result. The first of those is industrial: an electromagnetic-transient model of every converter on and around a corridor, accurate to the installed firmware and inspectable rather than encrypted, is what turns an adequacy study into a statement about the real fleet — and it is a thing manufacturers can choose to release or withhold, not a thing a laboratory can discover. Every post-event report cited in section 2 turns on a fleet behaviour the operator could not model beforehand. The second is institutional: a regulated return that pays an owner for delivered transfer capability rather than for capital installed. Under a rate-base return, dynamic ratings, topology optimisation, power-flow control and reconductoring all lose to the line they displace, which is why the cheapest capacity on the system is the least built. A legislature or a regulator can supply that; no brief on this map produces it.
• Enables Delivers to Planetary Scale Energy Systems. Any planetary-scale generation concept — orbital collection, continental-scale solar, anything whose output is produced far from where it is consumed — inherits the distribution problem whole, and inherits it in a harder form: the terrestrial corridor at least has jurisdictions that can be negotiated with. The transferable results are the converter-versus-line cost decomposition, the meshed-versus-point-to-point distinction, the common-pool ownership diagnosis, and the residual-value framing that says a distribution asset's benefit shrinks as everything else improves.
• Adjacent Adjacent to Infrastructure Resilience, which is the sibling brief and shares an author lineage rather than merely a subject: Paul Hines, whose topological-models paper anchors VII-20 and whose title is a question the field reads as a statement, is also among the twenty-one authors of the PNAS rebuttal that anchors the transmission-assumption dispute here. The shared commitment — prefer the physics to the convenient abstraction — cuts both ways across the two briefs. It says that a corridor's contribution to resilience cannot be established from a graph, and it says that a model's claim about how much transmission a decarbonised system needs has to survive contact with the power flow. Adjacent also to Continental Transportation Systems and Northern Development Corridors, which share the linear-infrastructure siting problem exactly — a right of way is a right of way, and the compensation question is identical whether the thing crossing the land is a conductor or a rail. And adjacent to Civilization Resilience Planning, for which continental interconnection is simultaneously a hedge against local failure and a new coupling that propagates a distant one.

Established “HVDC is cheaper than AC over long distances, so a supergrid is cheap.” The comparison is a break-even distance, not a price. It turns on converter-station cost, which is largely distance-independent, against line cost, which is not. A meshed network with many terminals pays for many converter stations, so the per-kilometre economics of a two-ended link do not extrapolate to the thing being proposed. This is the most common quantitative error in supergrid advocacy and it is an error of structure, not of arithmetic.

Established “The technology is proven.” Point-to-point HVDC is proven, extensively, across sea depths, climates and half a century. Meshed multi-terminal multi-vendor DC is a different object with a named, published obstacle list — DC breakers, interoperability, protection coordination — and a supergrid is by definition meshed. Every sentence that begins “the technology already exists” should be asked which of the two objects it means.

Handwave “It costs $X per gigawatt-kilometre.” This brief refuses to supply that number, and the refusal is the argument. Any single cost-per-GW-km figure is uninterpretable without six qualifiers attached: route type, since overhead line, underground cable and submarine cable differ by roughly an order of magnitude; number and rating of converter terminals; voltage class; year and currency; and whether the figure includes land acquisition, AC reinforcement at both ends, and financing. Established The research behind this brief hit a provenance gate that closed the project register, so it contains no per-project capital cost, no route length and no derived unit cost anywhere — declared as gap G-14.1. Frontier What would settle it is not a better search but a dataset that does not exist: a public register of completed HVDC projects with approval-stage estimate and outturn cost, disaggregated by route type and terminal count. Note what follows. The central claim of this brief — that the binding term is elapsed time and ownership rather than capital — does not need the number. A cost-per-GW-km would tell you what a corridor costs to build; it would tell you nothing about why NorthConnect, fully designed, was never built, or why a common-pool asset with no owner does not get provided. Supplying a plausible-looking figure would have made the brief look more quantitative and its argument no stronger.

Frontier “DESERTEC died because Europe feared dependence on North Africa.” That specific objection had a peer-reviewed treatment in Energy Policy in 2011, before the collapse. Speculative Whether the paper answered it negatively is unverified here, so this brief does not claim the objection was refuted — only that it was not an unexamined show-stopper, which is how the popular account uses it. Speculative The competing explanation, that the cost premium collapsed when European photovoltaics got cheap, is testable against published cost trajectories and is the one this brief would bet on.

Speculative “Interconnectors deliver security of supply.” They deliver energy when the counterparty has a surplus. The security case requires delivery under correlated stress, which is exactly when a surplus is least likely and when export curtailment by the exporting jurisdiction is most likely. Frontier The test is clean and nobody in this evidence base has run it: measure interconnector flows during correlated regional stress events and check whether power runs away from the stressed system. Continental European and British winter events supply the cases.

Speculative “Permitting is the bottleneck.” Widely believed, plausibly true, and not verified anywhere in this brief's evidence base. It is carried here as a hypothesis with a stated test — decompose median proposal-to-energisation time and check whether permitting plus litigation exceeds procurement plus construction across jurisdictions with different technology stacks — rather than as a fact. Frontier The counter-evidence is worth stating: NorthConnect had a completed engineering design and did not proceed, which points at commercial and regulatory failure rather than at planning consent.

Established “More transmission always helps integrate renewables.” Transmission is one flexibility option competing with storage, demand response, thermal storage and curtailment, and its marginal value falls as the others improve. The firebrick study quantifies that displacement for a single technology: one new thermal-storage option cut modelled battery capacity by about 14.5%, hydrogen by about 31% and heat storage by about 27.3% across 149 countries. Frontier Nobody has run the equivalent calculation for a corridor.

Frontier “The 100%-renewables question is settled” — in either direction. It is an active formal dispute in the same journal: Jacobson and colleagues in 2015, Clack and twenty co-authors in 2017. Cite both, in that order, and note that the disagreement is substantially about how much transmission the model may assume — which makes this brief's subject the crux of the argument rather than a detail in it.

Established “China proves a supergrid works.” Chinese deployment proves that ultra-high-voltage point-to-point links work at scale under a single vertically integrated state-backed owner. That is evidence for the ownership-concentration hypothesis, and it is evidence against the multi-state meshed proposition, since it demonstrates the enabling condition that multi-state proposals lack. The exception proves the rule in the original sense of testing it.

Established And one correction that belongs in this brief although it comes from the sibling: a famous paper whose title is a question is routinely cited as though it answered yes. Hines, Cotilla-Sanchez and Blumsack asked Do topological models provide good information about electricity infrastructure vulnerability? (Chaos 20(3):033122, 2010). Across 40 Eastern United States power-grid areas, using characteristic path lengths, connectivity loss and blackout sizes, they found that topological metrics correlate only mildly with physics-based power-grid models. Frontier The paper is a demurral from the complex-networks grid-vulnerability programme, and it is cited as a foundation of it. Established The related quotation error compounds it: the paper's finding that directed attacks cause larger failures than random ones is frequently quoted alone, detached from the surrounding argument that the measures producing it are weakly informative. Quote both or neither. Frontier This matters here because the security argument against grid interdependence is usually a topological argument, and the best available test says topology is a weak proxy for what a power system actually does. Established The same author co-signed the PNAS rebuttal on transmission assumptions, and the through-line — prefer the physics to the abstraction — is the methodological commitment this brief and Infrastructure Resilience share.