Wireless Energy Transmission

Wireless energy transmission means moving usable power from a source to a load without a conductor between them. It divides into two physically distinct regimes that share almost nothing except a name. In the near field, an inductive or resonantly coupled pair of coils exchanges energy across a gap that is small compared with the coils themselves: phone pads at millimetres, machine-tool spindles and subsea couplers at millimetres to centimetres, vehicle plates at ten to twenty centimetres, roadway coils at about twenty centimetres. In the far field, a microwave or optical beam carries power across hundreds of metres to kilometres, is collected by a rectenna array or a photovoltaic receiver, and is rectified back to direct current. The first regime is a shipped, mature, high-efficiency technology. The second is a fifty-year research programme whose best independently reported end-to-end result at any range beyond a hundred metres is under twenty-one per cent.

This brief also carries Category V slot 16, Beamed Energy Systems, because the two slots describe one subject under two names — a sibling collision found by measuring co-occurrence across the corpus rather than by comparing titles. The far-field half of this brief is that slot.

Where this brief stops, stated precisely because the boundary is the one readers cross by accident. Everything here is terrestrial or near-Earth: ground-to-ground microwave links, ground-to-air laser beaming, roadway and depot charging, consumer pads, implants, subsea couplers, industrial machinery. Beaming from orbit — the launch mass, the orbital aperture, the ionospheric transit, the receiving-site land take and the economics of a solar power satellite — belongs to space-based solar power and is not re-argued here. The distinction matters more than it looks, because several of the strongest terrestrial results in this brief were produced by programmes whose stated destination is orbital: the Xidian University ground verification system that holds the best far-field efficiency figure in this brief is a space-solar-power programme running terrestrial tests. This brief takes the terrestrial measurement and declines the orbital extrapolation. Beaming for thrust rather than for power delivery is beam-powered propulsion, also separate.

One convention governs the whole brief and is the reason it is worth reading. An efficiency figure for wireless power is meaningless without its measurement basis, and the three bases in common circulation differ by tens of percentage points on the same hardware. End-to-end or DC-to-DC or wall-to-load efficiency counts everything from the mains or the source direct-current bus to the energy that actually arrives in the battery or load. Coupler-to-coupler or coil-to-coil efficiency counts only the magnetic link and excludes both power-electronics stages. Rectenna conversion or receiver conversion efficiency counts only the last stage — the radio-frequency-to-direct-current or optical-to-electrical step at the receiving aperture — and says nothing about how much of the transmitted beam the aperture caught in the first place, which is beam collection efficiency and is the figure this sector publishes least often of all. A peer-reviewed static charging system quoted at 97.7% is coupler-to-coupler; the same class of system quoted grid-to-battery is 90%. A laser record quoted at “better than 20%” is optical-to-electrical at the receiver and the transmitter's wall-plug loss is not in it. Conflating these is the classic error in this subject, and it is committed by vendors, by trade press and by government press offices alike. This brief states the basis with every number, and where a source did not publish one, it says so.

Established The framing this slot inherited assumes the enemy is the two per cent that a copper conductor loses, and that comparator is wrong. A bare conductor is 97 to 99% efficient; a wired charging chain is not. The Idaho National Laboratory instrumented a 2012 Chevrolet Volt on-board charger and measured 91.7% at 3.0 to 3.3 kW on 208 VAC and 90.6% at about 1.3 kW on 120 VAC; including the supply equipment, wall-to-battery efficiency was 90.8% on level two and 88.8% on level one. Real-world instrumented figures reported in trade press are worse still: a Tesla Model Y drew 92.2 kWh at a level-two supply to deliver 81 kWh, about 86%; a BMW iX on a 22 kW wall box drew 125.2 kWh for 105.2 kWh, about 84%. The small-converter case is worse again: an instrumented iPhone 15 Pro charge put 12.7 Wh into the battery for 18.25 Wh at the wall, about 70%. Using 98% as the benchmark overstates the wireless penalty at short range and understates it at phone scale, which is exactly backwards in both directions.

Established At a few centimetres, wireless is efficiency-neutral against a real cable, and the honest objection to it is capital cost rather than watts. Against the 90.8% wall-to-battery figure above, peer-reviewed grid-to-battery static inductive transfer lands at 90% at 3 kW. SAE J2954 was reissued on 13 August 2024 as a full Standard rather than a Recommended Practice, covering three charging levels up to 11 kVA with 22 kVA foreshadowed, and trade coverage of the publication reports “efficiency of up to 93 per cent” — a figure this brief carries as trade-reported, because the standard text itself was not read. Porsche states it will be the first manufacturer to ship an 11 kW one-box base plate, on the Cayenne Electric in Europe in 2026, claiming “up to 90 per cent” grid-to-battery across a gap of roughly four to six inches from a plate measuring 117 × 78 × 6 cm and weighing a good 50 kg. Three of those four numbers come from an interested party. The peer-reviewed one does not, and it agrees.

Established Beyond about half a coil diameter, the penalty is not two-fold but four-fold to forty-five-fold, and the crossover is a geometric fact rather than a maturity gap. The controlling published statement is that inductive systems achieve better than 90% efficiency only within the reactive near field, “typically a distance less than half the coil diameter,” with efficiency dropping exponentially beyond that unless the coil is made larger. That single sentence sets the architecture of the entire field: wireless power range is bought with transmitter aperture, not with cleverness. A one-metre roadway coil buys roughly half a metre of range. A sixty-millimetre phone coil buys roughly thirty millimetres. Porsche's 117 × 78 cm plate working across four to six inches sits comfortably inside the rule rather than defying it.

Established The best independently reported end-to-end far-field figure anywhere is 20.8%, at just over a hundred metres. A Xidian University ground verification system, reported on 19 May 2026, beamed microwave power to multiple moving targets and delivered 1,180 W at more than 100 m with a DC-to-DC transmission efficiency of 20.8%, plus 143 W to an uncrewed aircraft at 30 m. That is the strongest number in this brief and it deserves both its weight and its caveats: it is DC-to-DC, which is the honest basis; it is reported through Chinese state media and the underlying paper is not identified; the transmit aperture and the operating frequency are not given in the report; and no peer-reviewed source for it could be located. Against a wired chain at about 90%, 20.8% is a 4.3× energy penalty; against a bare conductor at 98%, 4.7×. The arithmetic is this brief's; the inputs are sourced.

Established At kilometre scale the penalty is roughly an order of magnitude, and at the one kilometre-scale microwave demonstration with a published transmitter rating it is plausibly forty-five-fold. The United States Naval Research Laboratory's SCOPE-M programme delivered 1.6 kW at 1 km at 10 GHz at Blossom Point, Maryland in September 2021, exceeding a 1 kW target by sixty per cent, and 1 kW at 1 km at MIT Haystack with higher average power; the receiving rectenna comprised “tens of thousands of X-band antennas,” each with its own rectifier diode. The programme was announced on 20 April 2022 and executed within twelve months. Frontier Its end-to-end efficiency was never published. Trade coverage reports a 100 kW-class high-power amplifier on the transmit side, which puts 1.6 kW out of a 100 kW input at order two per cent wall-to-load — an inference from an amplifier rating, not a measurement, and flagged as such. Neither the transmit aperture nor the rectenna aperture nor the beam collection efficiency was published, so the two per cent cannot be decomposed. Note also that NRL's “exceeded the goal by 60 per cent” is routinely misquoted as “60 per cent efficient.”

Established The laser record is a genuine distance record and is not an efficiency record. DARPA's POWER programme, through its PRAD receiver demonstration at the High Energy Laser Systems Test Facility at White Sands Missile Range, announced on 16 May 2025 that it delivered more than 800 W during a 30-second transmission at 8.6 km — about 24 kJ in that run, with more than a megajoule transferred across the campaign. The receiver is a compact aperture feeding a parabolic mirror onto dozens of commercial off-the-shelf photovoltaic cells. The published efficiency is “better than 20 per cent from the optical power out of the laser to the electrical power out of the receiver” — at shorter distances. Frontier The efficiency at 8.6 km is not disclosed, and neither is the wavelength, the transmit aperture, or the beam collection efficiency at that range. Combining the measured better-than-20% optical-to-electrical receiver with an industrial laser at better than 50% wall-plug ceilings the whole chain near ten per cent end-to-end — again this brief's arithmetic from two cited figures, not a measured system number. The team's own framing is the correct one: “a number of design decisions were made in the interest of building something quickly, not efficiently.”

Established And the worst-behaved case in the whole subject is the one with the installed base. Instrumented measurement with watt-meters at the mains and at the device, logging at five-minute intervals over a full 0 to 100% charge of a 12.7 Wh battery, found 18.25 Wh at the wall wired, 23.33 Wh on a well-aligned 15 W magnetic pad (about 54% wall-to-battery), and 33.93 Wh on a misaligned 15 W pad (about 37%). One vehicle-brand charging platform consumed 113.7% more energy per day than the wired path. Peak battery temperature reached nearly 40 °C on the aligned magnetic pad and stayed above 40 °C throughout on the misaligned one, which is a documented degradation mechanism rather than only an energy story.

Established The standby draw is the larger number and almost nobody counts it. Measured idle consumption with no device present was about 0.2 W for one magnetic charger and 1.4 W constant for a vehicle-brand platform. Across the roughly seventeen non-charging hours in a day this dominates the charging loss entirely: annual waste runs from 5.8 kWh to 17.5 kWh against 1.94 kWh for the wired path. A 1.4 W parasitic load is of the same order as the energy actually delivered into a phone battery in a day. The consortium promoting the standard publishes certification counts — more than 13,000 certified products, with more than 1,200 new certifications during 2025 across more than 330 member companies — and asserts that magnetic alignment “improves energy efficiency” without publishing a single efficiency figure. Handwave Its projection that nearly four billion units will ship within five years is a vendor-consortium forecast with no published methodology, and this brief uses it only to indicate order of magnitude. Even so, the aggregate is the largest energy quantity in the brief, and it has nothing whatever to do with beaming.

Established The second finding that cuts against the framing: the electric-road programmes did not fail technically. Purdue University and the Indiana Department of Transportation delivered 190 kW to a Cummins electric semi-trailer travelling at 65 mph over a quarter-mile of coils embedded in rigid concrete on US-52/231 near West Lafayette, installed in 2024 and demonstrated in the autumn of 2025 with about eight inches of clearance. The A10 motorway trial southwest of Paris ran 1.5 km with more than 300 kW instantaneous peak and more than 200 kW average under optimal steady-state conditions, live from October 2025 with four prototype vehicles. Those are working systems. What defeats them is civil cost and institutional readiness: of a United States estimate of $6.32 to $6.55 million per lane-mile, $3.08 million is civil infrastructure — concrete and trenching — against $2.23 million for the charging system itself; European project reporting puts maintenance at two-and-a-half to four times a conventional road; and Salt Lake City needed about three years to invent a first-of-kind permit because no construction or design standards specific to dynamic wireless power transfer exist in city codes. The binding constraint on electric roads is a pavement-reconstruction budget line, not a coupling coefficient.

The genuinely open questions in this subject are almost all engineering-economic or institutional. The physics that sets the shape of the field — how coupling falls with distance, what resonance buys, where the near-field boundary sits — has been settled for years and is treated as established in the section below. What follows separates what is actually open from what merely sounds open.

Frontier Whether an airborne relay architecture makes long-range optical power delivery useful is the largest funded open question in the field. DARPA's answer to distance is not a better beam. Solicitation DARPA-PS-25-17, “POWER: Subscale Aerial Demonstration (Phase 2),” with abstracts due 30 June 2025 and a project start in October 2025, targets three airborne relay nodes at 60,000 ft delivering 10 kW to a final ground node at 125 miles from a 50 kW source laser, with relay apertures under one metre, using HELSTF and an RQ-4 Global Hawk as the airborne platform; RTX took a $10 million relay-design contract in August 2023. Note what this architecture concedes: the answer to range is to shorten it repeatedly. Note also what it does not concede — a 50 kW source delivering 10 kW to the far end would be twenty per cent end-to-end optical, before wall-plug losses at the source, across three relay hops. That would be a remarkable result and it has not been demonstrated.

Frontier Whether DARPA POWER was completed or truncated is unresolved, and this brief does not resolve it. DARPA's own programme page now reads “This program is now complete… This page is no longer maintained.” IEEE Spectrum, covering the same programme, describes the 2025 record as wrapping up “phase one of a three-phase project,” and defence trade press documents a Phase 2 solicitation with an October 2025 start. These accounts cannot all be right. The most likely explanation is a website archival policy rather than a cancellation, but a reader should know that the possibility of truncation after Phase 1 has not been excluded by anything this brief could verify.

Frontier Whether laser beaming works outside desert conditions is newly open and newly interesting. On 17 June 2026 the Naval Research Laboratory, with Boeing, DEVCOM GVSC and the three services, demonstrated a dual-use laser that beams power and can immediately transition to countering a drone threat, across an airfield in severe weather including snowfall approaching whiteout conditions, until visibility nearly disappeared. It was explicitly contrasted with “previous record-setting power beaming demonstrations conducted in highly controlled desert conditions.” Handwave NRL published no power, no distance, no wavelength and no efficiency for that test. For a subject that turns on numbers this is the most frustrating datapoint of 2026: the most recent United States result in the brief is qualitative, and this brief records the qualitative claim and refuses to attach a figure to it.

Frontier Whether the A10 motorway trial is economically serious is unknown because the two parties running it have published no efficiency figure and no cost. More than 300 kW peak and more than 200 kW average over 1.5 km is a real engineering result. Neither VINCI Autoroutes as operator nor Electreon as supplier published a grid-to-battery efficiency, a cost per kilometre, or the amount of the Bpifrance award, in either of the two releases they issued. This is the single most important missing number in the dynamic-charging story, and its absence after a year of publicity is itself evidence about what the number is likely to be.

Frontier Whether the far-field consumer sector has any remaining path is close to settled, and the honest reading is that it has already resolved. Ossia's own news index shows no items after 2023, the most recent referencing CES 2023; the company appears dormant, and this is an established absence rather than an inference about its finances. Energous survives, but by abandoning consumer charging: financial-year 2025 revenue of $5.6 million, up 633% from $0.8 million, a net loss of $9.6 million that is its best since 2013, 36% gross margin, more than 25,000 PowerBridge transmitters shipped in 2025, and a Fortune 10 retailer expanding from 410 to 4,700 locations. The winning application is powering batteryless RFID and sensor tags and electronic shelf labels — microwatts, in retail aisles, not phones on a desk. That is a real business and it is not the business the sector promised.

Frontier Whether idle draw attracts binding regulation is now a question with a date on it. Commission Regulation (EU) 2025/2052, published 12 October 2025, explicitly extends ecodesign scope to “wireless chargers and related charging pads” and introduces standby power consumption limits for wireless chargers and charging pads, applying from 14 December 2028. This is the first binding regulatory response to the idle-loss problem anywhere. Frontier The specific numeric threshold could not be verified. The EUR-Lex full text would not render on repeated attempts; a compliance secondary source cites general no-load caps “as low as 0.3 watts for certain models,” which this brief records as unverified for the wireless category specifically. Against measured idle draws of 0.2 W and 1.4 W, where the threshold lands between those two numbers determines whether the regulation is a formality or a redesign order for a large part of the installed base.

Frontier Whether the first production wireless vehicle charger is interoperable with anything is open, and it is a bigger question than it sounds. Two Porsche newsroom releases and the trade coverage of them do not mention SAE J2954 or interoperability at all. If the first shipping product is a proprietary single-vendor system, then the standardisation that this field spent a decade achieving has not yet been exercised by the market, and a parallel IEC 61980 track that peer-reviewed review literature already flags as an unresolved standardisation gap acquires more weight rather than less. The price of the Porsche system is also unpublished.

Speculative What is not open, and is regularly presented as though it were: the distance relationship itself. No result in any source consulted for this brief suggests that the near-field coupling exponent, the far-field aperture relationship, or the dependence of maximum achievable efficiency on the product of coupling and quality factor is available for improvement. Higher quality factors, better beam-forming, metamaterial apertures and adaptive alignment all move the constant. They do not move the exponent. A brief that treated “better beam-forming will solve distance” as an open frontier would be misreporting the field's own revealed judgement, which is legible in the fact that its best-funded programme answers distance with relays.

Established The first bottleneck is geometric and it binds everything else. Coupling falls as the cube of separation in the near field, the useful range is about half the transmitter's own diameter, and resonance improves the constant by the cube root of the quality-factor gain rather than changing the exponent. No advance in materials, semiconductors, control or beam-forming reported in any source consulted alters that relationship. The only lever is a larger aperture, and a larger aperture is the cost. This is why the field's most heavily funded long-range programme answers distance with three airborne relay nodes rather than with a better transmitter.

Established The second is civil construction, and for electric roads it is the whole story. In the United States cost decomposition, the charging system is $2.23 million per lane-mile, power distribution $1.01 to $1.24 million, and civil infrastructure $3.08 million — the largest single line item. The coils are not the expensive part. European project reporting adds that maintenance runs two-and-a-half to four times a conventional road, which converts a capital problem into a recurring one. Any proposal to electrify a corridor is in the first instance a proposal to reconstruct its pavement.

Frontier The third is that nobody agrees what a kilometre costs, and the estimates span a factor of twenty. United States Department of Transportation knowledge resources, citing a 2021 Indiana joint transportation research programme study, give $6.32 to $6.55 million per lane-mile, or $3.9 to $4.1 million per kilometre. European Commission final reporting on FABRIC gives €2 to €4 million per kilometre. A 2026 peer-reviewed review cites $1.7 to $3.5 million per kilometre. A separate 2026 peer-reviewed techno-economic paper takes a baseline of €1,000 per metre, or €1 million per kilometre, with a sensitivity range of €0.5 to €2 million — a factor of four below the government-cited figure, and this brief could find no reconciliation between them. At the far end, Korea's online electric vehicle programme was reported at 20 to 30 billion won per kilometre, roughly US$17 to US$26 million, an order of magnitude above everyone else. For comparison, Sweden's Smartroad Gotland spent 116 million SEK on 1.6 km, about 72.5 million SEK per kilometre or US$7 to US$8 million — but that budget is research-inclusive for a first-of-kind installation, so it is an upper bound rather than a construction cost. The arithmetic in that last sentence is this brief's; the inputs are sourced. A technology whose cost estimates span twenty-fold does not have a business case; it has a research agenda.

Established The fourth is institutional, and it is the one nobody budgets for. Salt Lake City took about three years to create a first-of-kind permitting process for dynamic wireless power transfer, for the stated reason that “there are no construction and design standards specific to DWPT systems currently within city codes.” That is not a technology risk and it is not a one-off: every jurisdiction that wants a powered roadway must repeat it until a model code exists, and no source consulted describes one being written.

Established The fifth is human-exposure compliance, which derates real systems by a factor of nearly four. The Chalmers result — 4.8 kW at 96% unconstrained falling to 1.3 kW at 89% once the ICNIRP 8.84 µT limit is respected, at a 0.3 m gap — is the cleanest quantification available. Every vehicle-scale system pays some version of this, and headline power figures that do not state whether exposure limits were enforced should be read as the unconstrained case.

Frontier The sixth is the vendor's balance sheet, which is a bottleneck in the plain sense that a dead supplier ships nothing. The dominant vendor in the dynamic-roadway story reported financial-year 2025 revenue of 15.97 million ILS, down 49.5% year on year, against a net loss of 101.6 million ILS — a loss roughly six times revenue — with a market capitalisation around 636 million ILS. Those figures come from a third-party financial data provider rather than the filings themselves, because the company's investor-relations pages could not be reached and its most recent report, Q1 2026 filed 28 May 2026, is published in Hebrew only. The brief records both the numbers and the provenance limitation.

Established The seventh, for far field, is that the approvals are transmitter-side and small. Regulatory clearances for over-the-air power exist at 1 W and about 5 W of transmitted radio-frequency power. No amount of receiver engineering makes a watt at the transmitter into watts at a device metres away; the aperture relationship stands in between. The bottleneck for room-scale wireless charging is not regulation and not electronics. It is that nobody has proposed a transmit aperture large enough to matter and small enough to live in a room.

Established Near-field wireless power depends on nothing outstanding. It is a shipped, standardised, mature technology whose physics is settled and whose economics are already legible. Nothing in this brief's near-field half is waiting on a discovery, which is why the near-field frontier questions above are regulatory and commercial rather than scientific.

Established Far-field power depends on aperture, and aperture depends on how much structure a buyer will pay for. That is a structural-economics dependency, not a physics one, and it does not resolve into any result on this map. It is the reason the relay architecture exists: three nodes at 60,000 ft is a way of buying range with airframes instead of with aperture.

Frontier The optical route depends on wall-plug laser efficiency, and that single number caps the application space more than any beam-forming improvement could. With source conversion at 40 to 60% and receiver conversion at 40 to 60%, the chain ceilings at 16 to 36% before atmosphere. At the best-case 50% wall-plug figure reported for industrial lasers, combined with DARPA's measured better-than-20% optical-to-electrical receiver, the whole chain ceilings near ten per cent. Doubling laser wall-plug efficiency would be the largest single improvement available to laser beaming, and it is a component-industry problem external to this field.

Established Electric roads depend on pavement reconstruction capacity and on municipal code. Civil works are the largest cost line, maintenance is two-and-a-half to four times normal, and no city had a permitting pathway until one was invented over about three years. Both dependencies are institutional and neither is a research result, which is why the techtree below records them under Requires rather than Depends on.

Established What depends on this is narrower than the field's rhetoric suggests, and it is worth naming precisely. Connector-free power in places where connectors fail depends on it entirely: subsea vehicle docking, implanted medical devices, sealed and rotating machinery, hazardous-area equipment where a slip ring would need explosion-proof certification. Depot and terminal opportunity charging for buses depends on it commercially. Space-based solar power depends on the far-field half of it absolutely — a solar power satellite is a beaming problem with a launch problem attached, and every efficiency term in this brief appears in its ledger. Advanced battery technologies are the direct substitute rather than a dependant: the practical competitor to charging on the move is a larger battery, and every improvement there weakens the case for a powered roadway.

Established The decisive experiments in this subject have largely been run, and several of them returned answers by being taken apart. The deployment record below is the natural experiment, and it should be read as a set of results rather than as a list of projects.

Established The most useful results here are the negative ones, and the strongest is a dismantling. Gotland's road was not merely switched off at the end of the project — “at the end of the project the electric road was dismantled.” Korea's programme is the more complete verdict: all four original wireless bus lines were shut down as infrastructure aged, the 2014 Gumi commercial route was abandoned, earlier prototypes were “left in the garage or scrapped,” commercialisation “has not been successful,” and continued public funding became a controversy in 2019. A technology that works, is deployed, and is then removed has returned a commercial answer, and it is a more informative answer than any laboratory result. Frontier The status of the 2019 Yuseong line specifically is unverified; the four earlier closures are confirmed.

Established The clean natural experiment on measurement basis is already in the literature, run inadvertently by one EU project against itself. Two FABRIC sites, in the same programme and the same era, testing the same function, reported 60.1% at 50 km/h at the French site and better than 80% at 50 km/h at the Italian prototype, grid-to-battery. The peer-reviewed paper on the French site reports about 70% grid-to-battery with the designers projecting “90% with mature technology” — Frontier a projection, not a measurement. That spread, within one project, is the cleanest available illustration of how immature the field's measurement practice is, and this brief reports the disagreement rather than averaging it.

Handwave A negative result about assumptions rather than hardware. A 2026 peer-reviewed techno-economic comparison of dynamic wireless charging against hydrogen builds its case on an assumed 96% grid-to-wheel efficiency for the wireless system. That assumption sits twenty-six to thirty-six percentage points above the same technology's measured field performance in the FABRIC results above, and it drives the paper's entire comparison. This is recorded not to impugn the paper but because it is the exact failure mode this brief exists to guard against: a plausible-looking efficiency number with no stated measurement basis, propagated into an economic conclusion.

Frontier The experiments still genuinely open are three. Whether an airborne relay chain can deliver 10 kW at 125 miles from a 50 kW source through three nodes at 60,000 ft — funded, solicited, not demonstrated. Whether the 20.8% DC-to-DC figure at just over 100 m survives independent replication and peer review, since it is currently the field's best end-to-end number and rests on state-media reporting of an unidentified paper. And whether the 14 December 2028 European standby limit, once its numeric threshold is known, is above or below the 1.4 W idle draw that instrumented testing found on a mass-market pad.

Established One experiment that has been run and is rarely reported as one: the market test. Two decades of over-the-air power have produced regulatory grants for 1 to 8 W of transmitted radio frequency and a commercial market in microwatt-class tag powering, with one vendor dormant since 2023 and the other profitable-ish only after abandoning the original application. That is a result.

Established Coupling collapses with the cube of separation once the gap is large relative to the coil. For coaxial coils at large separation, mutual inductance goes as M ≈ (μ₀/2π) · (D₁D₂)² · n₁n₂ / d³, so the coupling coefficient k = M/√(L₁L₂) falls as 1/d³ in that regime. Peer-reviewed review literature states the same contrast directly: near-field systems operate under an inverse-cube decay while far-field systems follow an inverse-square decay. The far field decays more slowly, which is why beaming exists at all; it starts from a vastly worse constant, which is why it is not competitive.

Established Resonance buys a factor, not a scaling law, and the algebra says exactly how much. Maximum achievable efficiency depends only on the product of coupling and quality factor: η_max = (kQ)² / (1 + √(1 + (kQ)²))². The sources disagree on the form and this brief reports the disagreement. A book chapter on coupled-mode theory renders the same relation as U²/(1 + U)² with U = k√(Q₁Q₂); the tutorial review gives the form above. The two are not equivalent, and the optimal-load result is the one with the nested radical; the practical consequence is small at high kQ and large at low kQ, which is to say it matters precisely in the regime where wireless power is marginal. Because η depends on kQ and k falls as 1/d³, a tenfold improvement in Q buys about 2.15× in usable range. That cube root is the entire return on a century of resonator engineering, and it is why the field's figure of merit is kQ rather than either factor alone.

Established The engineering rule that follows is the most useful sentence in the subject. Inductive systems achieve better than 90% efficiency only within the reactive near field, typically at a distance less than half the coil diameter, and efficiency drops exponentially beyond that without increasing coil size. Range is bought with aperture. A confirming datapoint at the boundary: a 2024 preprint transmitted 109.7 W at 1 m at 47.14% system efficiency, and its authors explicitly note that roughly fifty per cent is the maximum-power-transfer point rather than a maximum-efficiency point — a distinction that is itself a measurement-basis trap, since a system tuned to deliver the most watts is not the same system as one tuned to waste the fewest.

Established Human-exposure limits derate real vehicle systems by a factor of nearly four, before any other consideration. A Chalmers master's thesis optimised a coil pair at a 0.3 m gap and reached 4.8 kW at 96% against a 20 A conductor limit — but only 1.3 kW at 89% once the ICNIRP field limit of 8.84 µT was respected, with coupling coefficients in the range k = 0.09 to 0.15. That is a 3.7× power derating from exposure rules alone. The European FABRIC project independently reported in-cabin fields below 27 µT at 85 kHz, compliant with ICNIRP 2010. Any wireless-power figure quoted without a statement of whether exposure limits were enforced should be assumed to be the unconstrained one.

Established The measurement basis is where most of the apparent disagreement in this field lives. The table below is the peer-reviewed static charging record with its bases attached, and it is the reason the same technology is honestly describable as either 90% or 97.7%.

Established The 97%-plus figures are coupler-to-coupler and must never be quoted against a wall-plug number. The two rows that carry an end-to-end basis land at 90%, which is the figure that belongs beside the 90.8% wired baseline. The bus figure carries a further caveat worth stating: it is coil-to-coil, not grid-to-battery, and it is sourced to a 2014 evaluation study rather than to a current measurement.

Established Far-field hardware has its own basis problem, and it is worse because a fourth term enters. A rectenna array converts incident radio-frequency power to direct current at an efficiency that is genuinely high — but the array only receives the fraction of the beam that its aperture intercepts, which is beam collection efficiency, and that fraction is set by aperture, wavelength and range. Review literature frames the generic microwave limitation as “the Gaussian beam challenge”: much lower conversion efficiency at the beam edges, and difficulty sustaining better than 80% rectification over long distances. The practical consequence for a reader is that a published rectenna efficiency and a published end-to-end efficiency can differ by an order of magnitude on the same demonstration, and none of the three headline far-field demonstrations in this brief published a beam collection efficiency at all. The 20.8% Xidian figure is valuable precisely because it is stated DC-to-DC and therefore has all four terms inside it.

Established The optical chain decomposes into three cascading stages and the arithmetic is unforgiving. Review literature puts laser source conversion at 40 to 60%, transmission losses as variable with atmosphere, and receiver conversion at 40 to 60% — an end-to-end ceiling in the 16 to 36% band before any atmospheric loss. IEEE Spectrum's reporting notes that industrial lasers reach better than 50% wall-plug, and that because optical frequencies still scatter in fog and cloud, microwaves are generally superior for atmospheric transmission. The confirming datapoint from the other direction: NRL's 10 GHz beam lost less than five per cent of its power even in heavy rainfall.

The nearest neighbour, and the one this brief spends the most care separating itself from, is space-based solar power. The relationship is not that orbital solar is an application of wireless power; it is that orbital solar is the far-field half of this brief carrying a launch bill. Every efficiency term set out here — DC-to-DC end-to-end, beam collection at the receiving aperture, rectenna conversion — appears in a solar power satellite's ledger, multiplied together, with atmospheric and ionospheric transit added. Read the two briefs together and the sibling's case is legible as an aperture argument: a satellite can afford a kilometre-scale transmit aperture and a kilometre-scale ground rectenna in a way that no terrestrial link can, which is the only respect in which orbit helps. That the strongest terrestrial efficiency figure in this brief was produced by a programme whose stated destination is space solar power is not a coincidence, and it is the cleanest illustration of why the boundary between the two slots is drawn at the atmosphere rather than at the technology.

Beam-powered propulsion shares the physics and inverts the objective: a propulsion beam wants momentum transfer and can tolerate efficiencies that would be absurd for power delivery, because the alternative is carrying reaction mass. The efficiency arithmetic in this brief does not transfer across that boundary and should not be quoted there.

Advanced battery technologies are adjacent as a substitute rather than a complement, and the substitution is direct. Every improvement in energy density weakens the case for charging on the move, because the electric road's entire value proposition is that the vehicle need not stop or carry as much. That is why the dynamic-charging record should be read against battery progress rather than in isolation.

Continental transportation systems is where the electric-road case would live if it scaled, and planetary-scale energy systems is where wireless transfer occasionally appears as an assumed enabling condition — usually, on the evidence here, at an efficiency it has never demonstrated.

Two non-brief neighbours matter for reading the sector. Ecodesign and standby regulation, which for the first time has a wireless-charging clause with a date on it. And directed-energy defence, which is adjacent in a way the field does not advertise: NRL's June 2026 demonstration pairs power beaming with counter-drone engagement on the same aperture, because a kilowatt-class beam that can power a vehicle can also damage a target, and the shared hardware is not a metaphor.

Established The defining institutional fact about far-field wireless power is that the buyer is a defence ministry, not a utility. The three most significant long-range results in the brief were produced by a United States naval research laboratory, a United States defence research agency, and a Chinese university programme aimed at space solar power. The funded forward programme is a military logistics concept — power to austere or contested locations without a convoy — and its architecture, three airborne relays from a 50 kW source, is shaped by that requirement rather than by any civil use case. No civil institution anywhere is a volume purchaser of beamed power, and this brief found no civil procurement of any kind. Whatever the technology eventually becomes, its specifications are being set by an operational requirement in which a ten per cent end-to-end chain is acceptable because the alternative is a fuel truck.

Established The regulator arrived before the market did, and that is unusual enough to be worth stating. The European Union has brought wireless chargers and charging pads into ecodesign scope with standby power limits from 14 December 2028 — the first binding regulatory response anywhere to a loss mechanism the industry has never quantified in public. This is the one institution in the brief acting on the largest energy quantity in it. Frontier Its effectiveness turns entirely on a numeric threshold this brief could not verify, sitting somewhere against measured idle draws of 0.2 W and 1.4 W.

Established Standardisation has landed for static vehicle charging and has not yet been exercised. SAE J2954 became a full Standard on 13 August 2024, covering up to 11 kVA with 22 kVA foreshadowed and specifying a differential inductive positioning system for automated alignment. A parallel IEC 61980 track exists and peer-reviewed review literature flags the relationship between them as an unresolved standardisation gap. Frontier The first production vehicle to ship with an 11 kW plate does not mention either standard in its manufacturer's own announcements. A standard that no shipping product cites is not yet an institution; it is a document.

Established The institution that does not exist is a municipal code for powered roadways. Salt Lake City spent about three years creating a first-of-kind permitting process because no construction or design standards specific to dynamic wireless power transfer exist in city codes — anywhere. Every subsequent jurisdiction faces the same three years unless a model code is written, and no source consulted describes anyone writing one. This is a cheap, unglamorous, entirely tractable obstacle that has done more to prevent deployment than any technical limitation in the brief, and it sits with a class of institution — municipal public works departments — that has no research budget and no reason to move first.

Frontier The commercial institutions in the dynamic-roadway sector are not durable on current evidence. The lead vendor's revenue halved in 2025 to 15.97 million ILS against a 101.6 million ILS net loss. Its most recent filing is available in Hebrew only and its investor-relations pages could not be reached, so the figures here come from a third-party data provider rather than from the accounts. Its actual revenue base is depot and terminal charging — twenty-three electric buses on inductive pads at one depot from November 2023, expanded to a second terminal eleven kilometres away in 2025 — not open road, and even there the charging power is not disclosed. The institutional reading is that the open-road story is a demonstration business and the depot story is the product.

Established In consumer far field the institutional record is a long sequence of announcements followed by very small approvals, and it has now resolved. Grants exist at 1 W and about 5 W transmitter-side. One company has published nothing since 2023. The other found a real market in retail sensor powering and a Fortune 10 customer expanding from 410 to 4,700 locations, at $5.6 million of annual revenue. The durable lesson for a reader is procedural rather than technical: ask for the approved power at the approved distance, and for the measurement basis of any efficiency figure. In this sector those two questions separate every real claim from every promotional one, and almost nobody asks them.

Frontier What would change the institutional picture, stated as a testable condition. A model municipal code for dynamic wireless power transfer, published by any national standards body, would remove the single largest non-technical obstacle to roadway deployment at negligible cost. A requirement from any journal, regulator or standards body that efficiency figures state their measurement basis would remove the largest source of error in the field's public record. Neither requires a discovery, funding at scale, or anyone's permission, and the absence of both is the clearest evidence available that no institution currently regards this technology as important enough to tidy up after.

The evidence base in this field is unusually dependent on interested parties, and the pattern of what they do not publish is itself the finding. Established The vendor leading the dynamic-roadway sector does not reveal the charging power of its own depot installations. Neither party in the largest motorway trial has published an efficiency figure, a cost per kilometre, or the amount of its public award. The manufacturer of the first production wireless vehicle charger does not publish a price, and does not mention the relevant standard at all. Three far-field companies decline to publish delivered power against distance. Two government laboratories published headline power and distance for record-setting demonstrations and no end-to-end efficiency for either. None of this is dishonest; all of it is selective, and the selection runs one way.

Established Public money is materially involved and the accountability is thin. The Swedish Transport Administration put 91 million SEK of 116 million into a road that was dismantled. The Michigan Department of Transportation put $1.9 million into a quarter-mile whose total cost has never been disclosed, because the vendor's share is undisclosed. The European Union contributed €6.495 million to FABRIC. France's award to the A10 consortium came through a public investment bank and its amount has not been disclosed by any party. Korea's programme became a funding controversy in 2019 after its commercial deployments were abandoned. In each case the public financed an experiment whose result it cannot fully audit, and in most of them the missing figure is the one that would let a taxpayer judge whether the experiment succeeded.

Frontier A specific research-integrity concern, stated as a concern rather than an accusation. A peer-reviewed 2026 techno-economic paper builds a comparison against hydrogen on an assumed 96% grid-to-wheel efficiency for dynamic wireless charging. The same technology's measured field performance in EU project reporting is 60.1% at one site and better than 80% at another, with about 70% in the peer-reviewed paper on the French installation. An assumption sitting twenty-six to thirty-six points above measured performance, unflagged, driving a published economic conclusion, is a failure of review rather than of authorship — and it is the exact mechanism by which an unstated measurement basis becomes policy input. This brief names the pattern because the pattern is the ethical content; it takes no position on the paper's other conclusions.

Established An undisclosed cost borne by consumers. Instrumented testing found peak battery temperature near 40 °C on an aligned magnetic pad and above 40 °C throughout on a misaligned one. Elevated temperature is a documented degradation mechanism, so the price of wireless convenience includes shortened battery life — a cost not disclosed at the point of sale, not covered by any standard this brief could find, and borne by the user rather than the maker of the pad.

Frontier An opportunity-cost question worth posing plainly. A kilometre of powered roadway costs $3.9 to $4.1 million on the most credible government-cited estimate, with maintenance at two-and-a-half to four times normal, and competes directly against larger batteries and static charging that already work. No source consulted answers whether that money buys more decarbonisation than the alternatives, and the one peer-reviewed attempt rests on the efficiency assumption criticised above.

Unresolved questions this brief states as an obligation rather than resolving. Frontier No end-to-end efficiency has ever been published for the NRL kilometre-scale microwave demonstration; the two per cent figure here is inferred from a reported 100 kW amplifier rating and is not a measurement. DARPA has not disclosed the efficiency of the 800 W run at 8.6 km. NRL's June 2026 adverse-weather demonstration has no published power, distance, wavelength or efficiency. Whether DARPA POWER was completed or truncated cannot be reconciled from the available sources. The Xidian 20.8% figure — this brief's headline far-field number — rests on state-media reporting of a paper that is not identified, and no peer-reviewed source for it could be located. The numeric standby threshold in the 2028 European rule is unverified. Whether the first production wireless vehicle charger conforms to SAE J2954 or interoperates with anything is unknown. Milton Keynes ran eight wireless-charged buses on a route announced in September 2012 for a summer 2013 launch, with contemporaneous mention of systems already running in Turin and Genoa, and this brief could verify no power level, no efficiency, no cost and no termination date for any of the three. And no numeric maximum permissible exposure or nominal ocular hazard distance was verified for any beaming system in this brief. These are stated because a reader is entitled to know the shape of what is missing, and because in this particular subject the missing numbers are not random — they are, with striking consistency, the end-to-end ones.

Established The civilisational case for wireless power is not a world without wires, and the physics forecloses that clearly enough that continuing to imply it is a category error. The prize is narrower and real: removing connectors from places where connectors are the failure mode. Subsea vehicle docking runs at 90.87 to 95.6% across 5 to 40 mm gaps at 735 W to 10.6 kW, with a Zhejiang University sea trial at 800 W and 85% in the South China Sea, because wet-mateable connectors are the dominant failure mode in subsea electrical systems. Implanted devices, sealed machinery and rotating spindles are the same argument. Where the gap is millimetres and the alternative is a connector that fails, wireless is both efficient and correct.

Established The general principle this case illustrates is about measurement discipline rather than about power. A technology whose performance figure can legitimately be quoted as 97.7%, 90%, 54% or 2% depending on which stages of the chain are counted will be systematically oversold, not by fraud but by selection. Every party in this sector — vendors, consortia, government press offices, university research offices, peer-reviewed authors building economic models on assumed inputs — has a standing incentive to quote the most favourable defensible basis, and each of those numbers is defensible. The remedy is not scepticism about the numbers; it is a requirement that a basis be stated. That requirement is cheap, is not imposed by any journal or regulator in this field, and would have prevented most of the errors this brief corrects.

Frontier The largest energy quantity in the subject is also the least discussed, and it is a defaults problem rather than a physics problem. A consortium projects billions of Qi2 shipments within five years. At 24 to 104% excess energy per charge against a cable, plus 0.2 to 1.4 W of continuous idle draw, the aggregate is large, diffuse, and invisible to every individual user who has simply put a phone down on a pad. It is the classic case for efficiency regulation, and Europe has now written the first binding rule with effect from 14 December 2028. Whether the threshold is set at a level that bites is, on the evidence in this brief, worth more kilowatt-hours than every beaming programme ever funded.

Speculative Where wireless power would genuinely change civilisational arrangements, it does so through orbit rather than through the ground. A terrestrial beam is an expensive way to move power a distance a cable already covers. An orbital beam moves power to places no cable reaches, and it is the only architecture in which a kilometre-scale transmit aperture is affordable. That argument is the sibling brief's to make, and every efficiency term set out here is one of its multiplicands. This brief's contribution to it is the discipline: any orbital case built on a receiver conversion efficiency rather than a DC-to-DC efficiency should be read again.

Established What already happened, because the chronology usually starts too late and skips the closures. KAIST's online electric vehicle ran from a campus line in 2009 through Sejong in 2015 to two Gumi lines in 2016 and Yuseong in 2019, and all four original lines have since shut down. The European FABRIC project ran from 2014 to 30 June 2018 on €6.495 million of EU contribution and produced the field's most-cited grid-to-battery numbers. Smartroad Gotland ran to autumn 2023 and was dismantled. Michigan's quarter-mile has been operational since November 2023. SAE J2954 became a full Standard on 13 August 2024. Purdue installed its quarter-mile in 2024 and demonstrated 190 kW at 65 mph in the autumn of 2025. The A10 went live in October 2025. DARPA set the laser distance record at 8.6 km in May 2025, NRL demonstrated adverse-weather dual-use beaming in June 2026, and Xidian reported 1,180 W at more than 100 m at 20.8% DC-to-DC in May 2026.

Frontier 2026: the first production wireless vehicle charger, in one market. Porsche's 11 kW system is stated for Europe in 2026 on the Cayenne Electric, with other markets later. No other production vehicle was verified as shipping with factory wireless charging. Whether it is J2954-conformant is unresolved.

Frontier 2026 to 2027: the roadway expansions that will either happen or quietly not. Detroit's three-quarter-mile Michigan Avenue extension was targeted for spring 2026 and has not been reported complete. Utah's Inland Port section was functional in May 2026 with full rollout slated for late summer or early autumn 2026. Handwave These are the schedules of programmes whose funding has already fallen sharply in at least one case, and this brief records them as intentions.

Frontier From October 2025: DARPA POWER Phase 2, three airborne relay nodes. Project start October 2025, targeting 10 kW at 125 miles through relays at 60,000 ft from a 50 kW source. No demonstration has been reported. Frontier Whether the programme is running at all is itself unresolved, since DARPA's own page describes POWER as complete.

Established 14 December 2028: the first binding standby limits for wireless chargers. This is the only hard, dated, legally binding milestone anywhere in the brief. The regulation is published; only the numeric threshold is unverified.

Speculative Late 2020s to 2030s: whether powered roadways exist as anything but demonstrations. Local reporting on the Utah programme quotes a horizon of “ten to twenty years.” Nothing in the cost record suggests a route to corridor-scale deployment at $3.9 to $4.1 million per kilometre with maintenance at two-and-a-half to four times normal, and the technology's own vendor is losing six times its revenue. The plausible path is opportunity charging at depots and terminals rather than open road, which is where the vendor's actual revenue already sits.

Handwave Any date attached to room-scale consumer wireless power. There is no candidate architecture, no proposed transmit aperture, and no vendor publishing delivered power against distance. The sector's two-decade record is one dormant company and one that pivoted to microwatt tag powering. This brief declines to supply a date.

• Depends on Nothing on this map. Near-field wireless power is shipped, standardised and mature, and its physics has been settled for years; far-field power depends on aperture, which is structural economics rather than a pending result. No brief on this map is holding either half up, and stating that plainly is the honest position — the constraints are recorded below because they are real, not because they are discoveries waiting to happen.
• Requires (not on this map) Transmit aperture for anything at range, since usable near-field distance is about half the coil diameter and far-field collection is set by aperture against wavelength and range. High-efficiency wall-plug lasers for the optical route, where a 50% device already ceilings the whole chain near ten per cent end-to-end. Pavement reconstruction capacity, because $3.08 million of a $6.3 million lane-mile is concrete and trenching and maintenance runs two-and-a-half to four times a normal road. Municipal construction and design standards for dynamic wireless power transfer, which did not exist anywhere until one city spent about three years inventing a permit. And a market willing to pay for convenience at parity efficiency, which is the entire business case for static vehicle charging. All five are industrial, civil, institutional or commercial capabilities rather than discoveries.
• Enables Connector-free power wherever connectors are the failure point: subsea vehicle docking at 85 to 95% in sea trials, implanted devices at 88% for a pacemaker, rotating machine-tool spindles where a cable blocks automatic tool exchange, and hazardous-area equipment where a slip ring would need explosion-proof certification. Those relations run to application classes rather than to Frontier Research briefs, so no typed enabling edge is claimed. The one edge to a brief on this map runs to space-based solar power, which is the far-field half of this subject with a launch problem attached.
• Adjacent Space-based solar power, the orbital case and the scope boundary this brief draws most carefully; beam-powered propulsion, the same physics used for thrust rather than for power delivery; planetary-scale energy systems, where wireless transfer appears as an enabling condition; advanced battery technologies, the direct substitute for charging on the move; and continental transportation systems, for the electric-road case.

“A copper wire is 97 to 99 per cent efficient, so wireless always loses badly.” Established That is true of the conductor and false of the charging chain. A national laboratory measured wall-to-battery efficiency at 90.8% on level two and 88.8% on level one for a production electric vehicle; instrumented real-world tests report about 86% for one vehicle on level two and about 84% for another on a 22 kW wall box, with direct-current fast charging losing about ten per cent. A phone on a cable measures about 70%. Using 98% as the comparator overstates the wireless penalty for static vehicle charging and understates how badly a phone pad performs against its own wired equivalent. It is the first correction because every other number in the brief is read against it.

“The NRL demonstration beamed power 1.6 kilometres.” Established The units are transposed. It was 1.6 kilowatts at one kilometre, at 10 GHz, at Blossom Point, Maryland, with a second site at MIT Haystack delivering 1 kW at the same 1 km, announced 20 April 2022 after twelve months of execution. The related error is reading “exceeded the goal by 60 per cent” as “60 per cent efficient”; the goal was one kilowatt and the achievement was 1.6, and no end-to-end efficiency was ever published for this demonstration.

“DARPA demonstrated a laser power relay at 800 metres.” Established Three errors in one sentence. It was 800 watts, at 8.6 kilometres, and there was no relay. The May 2025 result was a 30-second direct transmission at White Sands, with more than a megajoule transferred across the campaign, into a compact receiver aperture feeding dozens of commercial photovoltaic cells. The relay demonstration is a future programme — solicitation DARPA-PS-25-17, Phase 2, started October 2025, targeting three airborne nodes at 60,000 ft and 10 kW at 125 miles from a 50 kW source. Nothing has been flown.

“The DARPA result was better than 20 per cent efficient.” Established This is a measurement-basis error and it is the most instructive one in the brief. The published figure is better than 20% from the optical power out of the laser to the electrical power out of the receiver, and explicitly at shorter distances. It excludes the laser's own wall-plug conversion, it excludes whatever fraction of the beam missed the aperture at 8.6 km, and it is not quoted for the record run at all. Put a best-case 50% wall-plug laser in front of it and the end-to-end chain ceilings near ten per cent — this brief's arithmetic from two cited figures, and still an upper bound because beam collection at 8.6 km is unpublished. A rectenna or receiver conversion efficiency is not an end-to-end efficiency, and in this sector the gap between the two is routinely an order of magnitude.

“Static wireless charging reaches 97 per cent.” Established Coupler-to-coupler, yes: 97% at 22 kW, 100 kHz, 135 mm gap, and 97.7% at 100 kW, 25 kHz, 125 mm gap, both in the peer-reviewed literature. Those figures count only the magnetic link and exclude both power-electronics stages. The same literature's grid-to-battery figure is 90% at 3 kW. Quoting a coupler-to-coupler number against a wall-plug baseline is the standard way this technology is oversold by people who are not lying about anything.

“Wireless charging wastes half the energy.” Handwave Not at vehicle scale, where measured grid-to-battery inductive transfer is 90% against 90.8% for a real cable and the honest objection is capital cost. At phone scale it is nearly true and the framing still misses the larger number: 23.33 Wh at the wall for 12.7 Wh into the battery on a well-aligned 15 W magnetic pad is about 54%, and 33.93 Wh misaligned is about 37% — but the 1.4 W of constant idle draw measured on one mass-market pad wastes more over a year, at 17.5 kWh against 1.94 kWh wired, than the charging inefficiency does. The waste is mostly in the seventeen hours when nothing is charging.

“Electric roads failed because the technology does not work.” Established They did not fail technically. A quarter-mile of United States highway delivered 190 kW to a Cummins electric semi-trailer at 65 mph. A French motorway delivered more than 200 kW average over 1.5 km. They fail on $3.08 million of civil works in a $6.3 million lane-mile, maintenance at two-and-a-half to four times a normal road, three years to invent a municipal permit, and a lead vendor whose revenue halved in 2025 against a loss six times revenue. Sweden's road was dismantled and Korea's four bus lines were shut down — commercial verdicts, not engineering ones.

“Sweden and Michigan have kilometres of working electric road.” Established Sweden's Smartroad Gotland electrified 1.6 km of a 4 km stretch, ran to autumn 2023, and was dismantled. Michigan Central is a quarter-mile; one mile was the contract ceiling, not the build. The Michigan Department of Transportation contributed $1.9 million with the vendor covering an undisclosed remainder, and project funding fell to about $689,000 in early 2025 from $2.7 million the year before. The total installation cost has never been disclosed, and a $70 million figure attached in local reporting to the Michigan Avenue expansion almost certainly refers to the whole street reconstruction rather than to wireless charging; this brief did not verify a wireless-specific cost and does not quote one.

“France is running Europe's wireless motorway, continuing VEDECOM's work.” Established Those are two separate things a decade apart. VEDECOM's Satory work was the EU FABRIC project, which ended 30 June 2018: a 100 m lane at 20 kW, 85 kHz, 175 mm nominal air gap, about 70% grid-to-battery, at roughly €3 million per lane-kilometre. The current story is the A10 “Charge as You Drive” trial, a different consortium on a different road, live from October 2025, with no efficiency figure published by anyone.

“Korea solved dynamic charging years ago with OLEV.” Established All four original wireless bus lines were shut down, the 2014 Gumi commercial deployment was abandoned, earlier prototypes were “left in the garage or scrapped,” and commercialisation “has not been successful.” The reported cost was 20 to 30 billion won per kilometre, about US$17 to US$26 million — an order of magnitude above every other estimate in the literature — at about 17 kW per pickup, a maximum transmission rate of 85%, and a required gap under 20 cm.

“The FCC has approved unlimited-distance wireless power.” Established That is a statement about what a grant restricts, not about what a device does. One company holds a Part 15 grant of 19 October 2021 for a 900 MHz, 1 W transmitter with no distance restriction attached; another holds a Part 18 authorisation of 7 March 2022 at about 5 W with a prior one-metre limit removed. Both figures are transmitter-side. Current products are rated at 1 W EIRP and 8 W EIRP / 2 W conductive on 900 MHz and 2.4 GHz. Frontier No vendor in this sector publishes delivered power against distance, and this brief could find no third-party measurement; the physics places delivery in the microwatt-to-low-milliwatt band at metres, and the only quantitative anchor any vendor offers is one FAQ's “low microwatts.”

“Room-scale wireless charging is coming.” Established It has already been tried and the sector has resolved, not the way its promoters intended. One of its two leading companies has published nothing since 2023. The other reached $5.6 million of 2025 revenue by shipping more than 25,000 transmitters for batteryless retail tags and shelf labels — microwatts on aisles, not watts on desks. An infrared entrant claims “several watts at several metres” under IEC 60825-1 Class 1, but that figure comes from a tertiary source rather than a datasheet, and no FCC or FDA grant for it is documented.

“Lasers are the future of power beaming because they focus better.” Established In atmosphere the evidence points the other way, and the safety asymmetry is structural rather than incidental. NRL's 10 GHz microwave beam lost less than five per cent even in heavy rainfall and was at a power density “sufficiently low that it was intrinsically safe.” Optical frequencies scatter in fog and cloud, and reporting on DARPA's own programme states plainly that microwaves are generally superior for atmospheric transmission. Microwave beaming at demonstrated power densities is intrinsically safe; kilowatt-class laser beaming is administratively safe — it depends on interlocks, keep-out volumes and an operator in the loop, not on the beam being harmless. That is why the 2026 NRL demonstration pairs power beaming with counter-drone engagement on the same aperture. Frontier This brief could not verify a numeric maximum permissible exposure or nominal ocular hazard distance for any beaming system, and describes the eye-safety constraint qualitatively only.

“Better beam-forming will solve distance.” Established Maximum achievable efficiency depends on the product kQ, coupling falls as the cube of separation in the near field, and a tenfold quality-factor improvement buys about 2.15× in range. In the far field, collection is set by aperture against wavelength and range. The field's own revealed judgement is the decisive evidence: the best-funded long-range programme in the world answers distance with three airborne relay nodes. When a field's leading effort solves range by repeatedly shortening it, that is the field's verdict on beam-forming.