Advanced Air Mobility

Advanced air mobility is the attempt to build a new class of short-range aircraft — electrically powered, taking off vertically, carrying one to six people over roughly 20 to 150 kilometres — together with the vertiports, airspace procedures, charging hardware, pilot pipeline and operating rules that would let them fly on a schedule. The aircraft are the visible part and the least uncertain part.

Established The object at the centre of this brief is a legal category, not a vehicle type. An aircraft that takes off on rotors and cruises on a wing is neither an aeroplane nor a rotorcraft under codes that have governed civil aviation since the 1940s. What reserve energy is required, what a pilot must be rated on, what failure probability must be demonstrated, what a landing site must look like: all of it flows from treating these machines as powered-lift, a category that existed on paper for decades with almost no civil population. This brief uses “powered-lift” where the legal question is at stake and “eVTOL” where the engineering is.

Four unit conventions, because most disagreement here is incomparability rather than disagreement. Energy density quoted for a cell and for an installed pack differ by roughly a third, and aviation packs carry containment and thermal mass that automotive packs do not. A decibel figure is meaningless without the phase of flight, the slant distance, the weighting, and whether it was measured or modelled. Dollars per available seat mile, per passenger mile at an assumed load factor, and per block hour are three different quantities. And a fleet that cannot fly in cloud has a dispatch reliability set by local weather, not by the aircraft.

Established Where this brief stops. Uncrewed delivery aircraft appear only where they share airspace or infrastructure. Cell chemistry belongs to advanced battery technologies; this brief takes cell performance as an input and asks what it buys in the air. Surface alternatives belong to continental transportation systems.

Established The claim this brief lands. The technology is real, the certification path is real and partly built, the first shakeout has already happened, and the binding constraint is a four-way squeeze: pack-level energy density against mandated reserves, vertiport throughput against urban land, certification throughput against a regulator’s finite specialist staff, and a fare that must fall roughly an order of magnitude below helicopter charter to reach the market the projections assume. None of the four is a physics problem, and all four are harder than one, because none can be solved by a single company.

Established A note on sourcing. This brief was commissioned in September 2026 from the Institute’s research base. Reading-list entries without links are cited from the bibliographic record rather than re-fetched, and claims are dated no later than early 2026 unless carried by a linked source.

Established The United States created a new civil aircraft category in October 2024, the first since the helicopter. The Federal Aviation Administration’s Special Federal Aviation Regulation on powered-lift resolved what had blocked everything downstream: how a pilot becomes qualified on a type with no training fleet, how operating rules written for aeroplanes and helicopters apply to a machine that is alternately both, and what energy reserves it must carry. It did so by allowing powered-lift to be trained and operated under aeroplane or rotorcraft rules depending on configuration, with performance-based alternatives on application. It is the most consequential event in the field’s history and it is administrative, not technical.

Established Europe set the design bar first, and set it very high. EASA’s 2019 special condition for small-category VTOL aircraft created an “enhanced” category for flight over congested areas or with commercial passengers, requiring a catastrophic-failure probability no greater than 10⁻⁹ per flight hour — the large-transport-aeroplane standard, imposed on a machine one-fiftieth their mass — and continued safe flight and landing after failure without relying on autorotation, which most distributed-propulsion configurations cannot perform. It is the most demanding airworthiness target ever set for an aircraft this size, and the reason every schedule published before 2021 was missed.

Frontier As of early 2026 no eVTOL held a full type certificate from the FAA or EASA. The leading United States programmes had published airworthiness criteria under the special-class provision and were in late-stage certification flight testing with agency involvement. The dating is deliberate: certification is a discrete event that can occur in any month, and a reader in late 2026 should check the current status. What the brief asserts is the shape of the record — every entry-into-service date announced between 2018 and 2022 passed uncertificated, and the slippage was measured in years.

Established China certificated an autonomous two-seat multirotor first, and the terms are the interesting part. The Civil Aviation Administration of China type-certificated the EHang EH216-S in October 2023, with production and standard airworthiness certificates following and operator certificates issued to sightseeing operators in 2025. It is a pilotless multirotor with no wing, endurance in the range of 20 to 25 minutes, approved for defined sightseeing areas rather than point-to-point transport. A genuine first, and a poor proxy for the transport mission.

Established The first shakeout has happened, and it took the well-funded. Lilium, which raised well over a billion euros for a ducted-fan jet, entered German insolvency in October 2024, obtained brief rescue financing, and wound down in early 2025. Volocopter, the oldest European multirotor programme, filed for insolvency in December 2024 and its assets were sold out of the process. Vertical Aerospace lost a prototype in an uncrewed test accident in 2023 and recapitalised. Two of the four best-funded European entrants were gone inside a quarter, and the failures were financial rather than aerodynamic: the aircraft flew, and the money ran out before the certificate arrived.

Frontier For the survivors the certification date is a solvency question. Company-reported quarterly operating outflows for the two leading United States developers ran in the range of roughly US$100 million to US$180 million each across 2024 and 2025 against essentially no aircraft revenue; the range is wide because the disclosures differ in what they capitalise. The strategic investment sustaining them is real money, but the airline “orders” in the press releases are predominantly non-binding letters of intent with refundable or nominal deposits.

Established The energy arithmetic is the honest heart of the subject and fits on one line. The only defensible long-run series, harmonising more than ninety data series into 1,716 cell records from 1991 to 2018, puts commercial cell-level gravimetric density at the 98th percentile rising from about 80 to over 250 watt-hours per kilogramme. The intergovernmental benchmark for 2025 deployment puts the dominant lithium-iron-phosphate cell near 205 watt-hours per kilogramme and sodium-ion at up to 175. Installed pack density is roughly 60 to 75 per cent of cell level in automotive practice and lower in aviation, where containment and thermal management are certification requirements rather than design choices. A 200 watt-hour-per-kilogramme aviation pack is a good pack today.

Established Reserve energy eats the small battery twice. Hover power scales with disc loading, and a compact multirotor sized to fit an urban pad spends several times the specific power in hover that it spends in cruise. On top of that, commercial passenger service requires landing with enough energy to reach an alternate and hold. Because a battery’s reserve is carried as mass for the whole flight rather than burned off like fuel, the penalty compounds: the mass guaranteeing the diversion raises the energy needed for every prior minute. Published range and payload figures for the same airframe are frequently not simultaneously achievable, and a range number without its reserve, payload and alternate distance is not a number.

Frontier Noise is where vendor figure and public evidence diverge most. Developers publish overflight figures around 45 to 65 A-weighted decibels at stated altitudes against helicopters at 80 to 90. The mechanism is real: many small rotors at low tip speed move more air more quietly, and distributed propulsion spreads acoustic energy into frequencies that attenuate faster. The evidence is thin where it matters. Measurements are overwhelmingly of single aircraft in controlled conditions, the annoyance response to a novel multi-tonal source is not the same function as the response to a helicopter at equal decibels, and nobody has published the cumulative contour of a vertiport running at design throughput over housing.

Established The vertiport standards exist and they cap throughput hard. The FAA’s vertiport design engineering brief and EASA’s prototype specification for VFR vertiports both date from 2022 and both build on a final approach and take-off area sized to the aircraft’s controlling dimension, with obstacle-free surfaces at prescribed slopes. The consequence is arithmetic: one such area handles one aircraft at a time through approach, land, unload, charge, load and depart. With charging in the loop that is a handful of movements per hour per pad. A city-centre site with one pad is a curiosity; a site with throughput that matters is a small airport, and urban land does not readily produce small airports.

Established The founding economic claim was a factor-of-twenty cost reduction and it has never been demonstrated. The 2016 white paper that launched the modern wave projected near-term operating cost near US$1.84 per passenger mile falling to roughly US$0.44 at scale, against helicopter charter economics then and now around US$8 to 10. The reduction was to come from electricity replacing turbine fuel, simpler powertrains, high utilisation and eventually removing the pilot. Each mechanism is plausible alone; none has been measured on a certificated commercial fleet, because none has flown a year of scheduled service.

Established The demand forecasts inherit scepticism the transport-economics literature has already earned. The unretracted ancestor of the canonical forecasting result found cost overruns of 50 to 100 per cent common on large transport projects and traffic forecasts wrong by 20 to 60 per cent; later work showed the most-quoted urban rail cost band travels with its five qualifiers stripped. Air mobility sizings are the same species of analysis under weaker constraints, with no operating fleet to calibrate against. When one large bank cut its own 2040 sizing from roughly US$1.5 trillion to about US$1 trillion within three years with no observable change in the fleet, that was information about the forecast.

Established The public demonstrator record carries a warning. NASA’s X-57 Maxwell, its distributed-electric-propulsion testbed, ended in September 2023 without ever flying, on late-emerging mechanical issues and unavailable critical components, with the agency presenting the outcome as lessons and standards. A well-funded public programme with no commercial pressure and a conventional airframe could not close a distributed-electric flight demonstration on schedule. That does not show the commercial programmes will fail; it shows integration is harder than component maturity suggests.

Frontier Whether dispatch reliability in a real metropolitan climate supports scheduled service. Initial operations will be visual-flight-rules, single-pilot and weather-limited. Helicopter emergency medical services, the closest analogue with published statistics, report weather-related mission refusals in the tens of percent by region and season. An operator cancelling one flight in four has a different business from one cancelling one in twenty, and no figure exists for any eVTOL operation. The instrument-flight-rules path — icing certification, redundant sensing, vertiport approach procedures — is identified and institutionally unfinished.

Frontier Whether packs survive the aviation duty cycle economically. An air taxi pack is charged fast, discharged deep, cycled several times daily and retired at a state-of-health floor set for safety. Automotive cycle-life data come from gentler conditions. The open question is the replacement interval and therefore amortised pack cost per flight hour — one of the largest lines in any honest operating cost model, and not publicly measured for a certificated aviation pack.

Frontier Whether the pilot comes out, and on what evidence. The long-run cost case depends on reducing or removing the crew. Regulators have no accepted means of compliance for autonomous passenger-carrying flight in controlled airspace, and the intermediate step — simplified vehicle operations, or one remote supervisor for several aircraft — is a research programme rather than a rule. The Chinese pilotless certification shows a path for a narrow mission profile; it does not transfer to transit over a city.

Speculative Whether the Gulf launch model generalises. The most advanced deployments announced by early 2026 were in the United Arab Emirates, where one authority controls land, airspace, the airport, the regulator and much of the capital, and the weather is hot and predictable rather than icing-prone. That removes most of the institutional obstacles this brief identifies, which makes it a strong test of the aircraft and a weak test of the system.

Established Pack-level specific energy, not cell-level. The constraint is the installed figure after containment, cooling, structure and the unusable bottom of the discharge curve. Moving from roughly 200 to roughly 300 watt-hours per kilogramme at pack level is the difference between a 25 kilometre hop with a legal reserve and a useful 100 kilometre stage. No shipping chemistry delivers that under aviation qualification today.

Frontier Charge power at the pad, and its grid consequence. Turning an aircraft in the time it takes to unload and load implies several hundred kilowatts to over a megawatt. A four-stand vertiport is therefore a multi-megawatt customer in a dense urban network — an interconnection queue problem before an engineering one. Buffering with oversized ground batteries converts it into a capital and footprint problem rather than removing it.

Established Land, not air. The scarce input is a parcel in a dense place with the obstacle-clear approach surfaces the standards require, the electrical service, the ground access, and a local authority willing to permit movements above the neighbours. Rooftops are structurally and operationally harder than they look, and existing heliports were sited for a fleet an order of magnitude smaller than the projected one.

Frontier Assurance of flight-critical software and high-voltage systems. Distributed propulsion replaces mechanical linkages with a control computer, and the assurance case at the failure probabilities EASA set is expensive and slow. High-voltage architectures add arc-fault, insulation and thermal-runaway modes with no aviation precedent and no mature means of compliance.

Established The pilot pipeline. Type ratings are normally earned in the aircraft or a qualified simulator, neither of which exists in quantity before entry into service. The 2024 rule addressed this by permitting training in aircraft with a single set of controls under defined conditions and crediting aeroplane and rotorcraft experience. That solves the first cohort; scaling to the thousands a real network implies, in an industry already short of commercial pilots, is unsolved.

Frontier Emergency landing and third-party risk. Without autorotation, survival after propulsion failure rests on redundancy and, in some designs, a ballistic parachute needing altitude to deploy. Urban routing must therefore keep a landing option within reach at all times, which constrains the city-centre paths that generate the time saving the product sells.

Established This brief depends on cell-level results it does not own. Everything in the range and payload argument is downstream of gravimetric and volumetric cell performance under aviation duty, which is the subject of advanced battery technologies; that brief’s finding that density stopped being the binding automotive constraint does not transfer here, because in aviation it never stopped being binding.

Frontier A published aviation-qualified cycle-life dataset. Nobody outside the developers holds the curve of usable capacity against cycles for a certificated aircraft pack at aviation charge rates and depths of discharge. Without it every direct operating cost model in the field contains an unfalsifiable term.

Frontier The decisive result is a metered year, and it is a measurement rather than an experiment. The single result that would most change this brief’s assessment is one full year of scheduled revenue operations by a type-certificated powered-lift fleet, with four quantities published on a common basis: dispatch reliability against schedule including weather cancellations, direct operating cost per block hour with the pack amortisation shown separately, movements per hour actually achieved at the busiest vertiport, and noise complaints per hundred movements from the surrounding population. Every contested claim in the field resolves against that dataset and none of them resolves without it. No such year has been flown anywhere as of this brief’s cut-off.

Frontier The cheapest useful experiment is acoustic and nobody has run it at scale. Instrument a residential area around an existing heliport, run a certificated eVTOL on a representative profile at design movement rates for a season, and publish both the measured contours and the surveyed annoyance response beside the helicopter baseline from the same site. This needs one aircraft, a microphone array and a survey instrument that already exists. Its absence is a choice.

Speculative The natural experiment already partly running is the Gulf deployment. If commercial operations proceed there, they will generate the first real data on passenger willingness to pay, turnaround times and reliability — under institutional and meteorological conditions so favourable that a good result establishes an upper bound rather than an expectation for temperate, multi-jurisdiction cities.

Established Three configurations, three different bets. Multirotors are mechanically simplest and aerodynamically worst in cruise, which caps range. Lift-plus-cruise aircraft carry separate rotors for hover and a propeller and wing for forward flight, paying a permanent drag and mass penalty for hardware that is dead weight in one regime. Vectored-thrust designs tilt rotors or nacelles to do both jobs with one propulsion set, buying efficiency with a transition that is the hardest control and certification problem in the category. Every surviving transport-mission programme chose one of the latter two, and the choice determines the certification burden more than it determines the performance.

Established Disc loading is the master variable and it trades directly against noise. Power required in hover rises with the square root of disc loading, so small rotors sized to fit an urban footprint cost hover power, while large rotors cost tip speed and therefore acoustic energy. Distributing lift across many small rotors at low tip speed is the field’s central acoustic idea and it is bought with hover efficiency, structural complexity and a control system with many more actuators to fail.

Frontier Thermal management is a certification problem disguised as an engineering one. Fast charging and high-rate discharge both heat the pack, and the aviation requirement is not merely to cool it but to demonstrate that a single-cell thermal runaway cannot propagate to the pack or breach the cabin within the time needed to land. Containment adds mass, and mass is the thing there is least of.

Established The high-voltage architecture has no civil aviation heritage. Hundreds of volts distributed through a composite airframe with segregated redundant channels raises arc-fault detection, insulation coordination, lightning protection and bonding questions that the existing advisory material does not answer. Each has an engineering answer; the cost is that each must be argued from first principles with the regulator.

Established Batteries are the upstream dependency and the honest comparison. Advanced battery technologies documents that the automotive market moved toward lower-density, cheaper, safer chemistry once density stopped binding. Aviation is the counter-case in the same corpus: it is the application where the industry’s discarded objective function is still the right one, and it is small enough that it will not steer cell development on its own.

Established The surface alternatives set the price ceiling. Continental transportation systems establishes that the binding constraint on new-mode proposals is usually a missing number rather than a missing technology — nobody publishes the incumbent’s forward cost curve. The same defect appears here: air taxi business cases are compared against today’s congested road journey, not against the road journey as it will be with the improvements already funded.

Frontier The urban-instrumentation record is a warning about the deployment model. Smart cities finds two decades of urban technology programmes with no evaluated city-level outcome attributable to the technology. Advanced air mobility is arriving with the same municipal-partnership playbook and the same absence of a pre-specified evaluation, and will produce the same evidentiary vacuum unless the operating data above are published.

Established Three regulators, three philosophies, and no mutual recognition worth the name. The FAA built the category through a special class airworthiness path plus an operations rule; EASA built a bespoke special condition with a fixed quantitative safety target; China certificated a narrow autonomous mission first. Bilateral validation agreements exist but validating a type in a category the partner authority has treated differently is slow. A developer that certificates in one jurisdiction does not thereby have a global product, and the cost of the second and third certificate is a material fraction of the first.

Established The binding local authority is not the aviation regulator. National regulators control the aircraft and the airspace. Vertiports are controlled by municipal zoning, building codes, environmental review and, in many jurisdictions, a local right to object to noise. The historical precedent is exact: a fatal 1977 rooftop accident ended scheduled helicopter service from the top of a Manhattan office tower, and heliport closures across major cities since have been driven by local politics rather than by aviation safety findings. The aviation certificate is necessary and nowhere near sufficient.

Frontier Public money is arriving before public evidence. National programmes in the United States, the United Kingdom, Japan, Korea and the Gulf have funded corridors, demonstration sites and procurement commitments. The pattern the corpus documents elsewhere — large programmes with no pre-specified evaluation producing no attributable outcome — applies unless the funding instruments require the operating data this brief asks for. As of early 2026 they largely do not.

Established Noise is a regressive externality with a well-documented distribution. Aircraft noise burdens fall on the population under the approach path, which in most metropolitan areas is poorer than the population in the aircraft. An air taxi network makes this sharper than a conventional airport does, because the flights are numerous, low, and serve a small and wealthy ridership. Any honest assessment treats the noise question as a distributional question first and an acoustics question second.

Frontier The equity argument offered by the industry is weak but not empty. Proposals to serve medical transport, organ delivery, underserved airports and disaster response are genuine uses where the value of speed is high and the ridership is not a premium commuter. They are also small, and they do not produce the movement counts that the cost curve requires. A programme justified on emergency medical grounds and financed on commuter economics should be assessed on the commuter economics.

Established Public subsidy for a premium service is the live political question. Vertiport land, grid upgrades and airspace modernisation are public or quasi-public costs. If the resulting service is priced at helicopter-adjacent fares for a decade, the transfer runs from general taxpayers to a small ridership, which is a defensible industrial policy and an indefensible transport policy, and the two arguments are usually conflated.

Frontier The realistic upper bound is a premium mode, not a transformation. The travel behaviour literature’s most durable regularity is that average daily travel time budgets are roughly stable across societies and centuries; faster modes buy distance rather than time, expanding metropolitan footprints instead of freeing hours. Applied here, a successful air taxi network extends the commutable radius of a wealthy sub-population and induces its own demand. That is a real economic effect and it is not the civilisational one the promotional material implies.

Established The helicopter is the control group and its adoption curve is the honest prior. Civil turbine helicopters have been technically capable of urban point-to-point transport for sixty years. The worldwide civil fleet remains in the tens of thousands, urban scheduled service is a niche in a handful of cities, and the binding constraints were cost, noise and landing sites — the same three constraints this brief identifies. An eVTOL that improves all three by a factor of two ends up as a slightly larger niche. Transformation requires improving them by an order of magnitude, which is what the field promises and has not shown.

Handwave The claim that air mobility relieves ground congestion is assertion. At any plausible fleet size the number of trips removed from the road network is a rounding error against daily vehicle trips in a large metropolitan area, and the induced-demand mechanism works in the other direction. The brief flags this as handwave rather than merely speculative because the arithmetic is available to anyone who makes the claim and is never shown.

These horizons track the system, not the aircraft: certification of a type is a discrete event that could occur in any year, while the network, the throughput and the cost curve move slowly and observably.

• 10 yr: Frontier Type-certificated powered-lift aircraft in scheduled commercial service in a small number of metropolitan areas and Gulf states; fleets in the hundreds rather than thousands; single-pilot, largely visual-flight-rules, weather-limited; fares at or somewhat below helicopter charter; the first published dispatch reliability and block-hour cost figures settling several arguments in this brief.
• 25 yr: Speculative Instrument-capable operations, multi-pad vertiports at airports and a few dense urban sites, pack-level energy density in the region of 300 watt-hours per kilogramme, and a consolidated industry of two to four airframers; regional thin-route operations plausibly larger than the urban commuter market that motivated the field.
• 50 yr: Speculative Reduced-crew or supervised-autonomous operation certificated in at least one major jurisdiction, with cost per seat mile approaching the low end of the 2016 projections; network scale still set by urban land and community noise tolerance rather than by aircraft performance.
• 100 / 250+ yr: Handwave Any claim that vertical flight becomes a mass urban mode rests on solving land and noise constraints that are political rather than technical, and on an energy density no announced chemistry currently supports; the brief declines to project past the point where the binding constraints stop being measurable.

• Depends on cell-level performance and cost results owned by advanced battery technologies, the urban electrical interconnection and siting machinery in energy corridors, and the autonomy assurance methods in agentic autonomy and control; nothing else on the map blocks the aircraft itself.
• Requires (not on this map) a published capacity-against-cycles curve for a certificated aircraft pack under aviation charge rates, which would replace the largest assumed term in every cost model; a production line qualified to aviation standards for those packs rather than automotive ones; a training and type-rating pipeline that works beyond the first few hundred pilots; a repeatable municipal route to permitting a vertiport over objections; a paying ridership large enough at launch fares to fund the years before unit costs fall; and charging hardware standardised so that one stand serves every operator.
• Enables thin-route regional connectivity between secondary airports and remote communities, faster emergency medical and organ transport where minutes are decisive, and a certification precedent for high-voltage electric propulsion that conventional electric aviation would inherit.
• Adjacent to continental transportation systems for the surface alternatives that set the price ceiling, smart cities for the municipal deployment and evaluation model, and frontier aerospace programs for the demonstrator failure pattern.

Handwave “Flying cars are finally here.” No programme in certification is a car. These are small aircraft needing a rated pilot, a prepared landing site, controlled airspace access and a maintenance organisation. The roadable aircraft — the literal flying car — remains a curiosity because the compromises that make a vehicle a good car make it a bad aircraft. The phrase survives because it sells, and it consistently misleads readers about what is being certificated.

Frontier “eVTOLs are a hundred times quieter than helicopters.” The mechanism is real and the multiple is not. Comparisons frequently set a cruise-overflight figure for the eVTOL against a hover or low-altitude figure for the helicopter, and because decibels are logarithmic a hundredfold energy claim corresponds to about 20 decibels — the very top of any measured difference and not sustained across hover and transition. The unanswered question is cumulative: many quiet movements can produce a worse community outcome than few loud ones.

Established “The order book is worth billions.” Announced order values are list prices multiplied by non-binding letters of intent, often with refundable deposits in the low tens of thousands of dollars per aircraft and conditions precedent including certification. Treating them as backlog is a category error; the useful number is cash received, and it is in the filings.

Frontier “Removing the pilot fixes the economics.” Crew is one of several large lines alongside pack amortisation, maintenance, insurance, vertiport fees and energy. Removing it entirely from a four-seat aircraft improves cost per seat mile materially but not by the order of magnitude the long-run projections need, and it requires an autonomy certification basis that does not exist. The projections reaching very low fares assume the pilot is gone and utilisation is high and maintenance is cheap and fees are low, all at once.

Frontier “Energy density will reach 500 watt-hours per kilogramme before 2030.” Laboratory cells at those figures exist; qualified, cycle-life-demonstrated, thermally contained aviation packs do not, and the historical commercial improvement rate does not reach there on that schedule. The honest statement is that the aircraft now in certification were designed around the packs that exist, and better cells would extend their missions rather than being a precondition for service.

Established “Certification is just paperwork and political will.” The European quantitative target, the absence of autorotation, novel high-voltage failure modes and the software assurance burden are substantive demonstrations, not forms. The 2024 operations rule shows what happens when the genuinely administrative obstacle is removed: it was removed, in one rulemaking, and the aircraft still had to finish flight test.