Let’s Revolutionise Aviation
October 2025
When I was in the US for two months this summer, I went to what is perhaps my favourite museum in the world: the Smithsonian Air and Space Museum in Washington DC. I am still dumfounded that the Wright brothers’ flight at Kitty Hawk was only 65 years apart from Neil Armstrong’s steps onto the moon’s surface and the SR-71’s Mach 3+ flights. Our species’ brief chapter of heavier-than-air-flight is truly unique. There isn’t another story that so dramatically captures man’s mastery of his environment; his ability to persevere, understand, and conquer realms previously thought of as fantastical. This rapid rate of progress was not necessarily inevitable, nor has it seemingly continued. Bear in mind that those vehicles from 1969 are, in some regards, still the best mankind has ever produced. The SR-71 remains the fastest manned, production aircraft of all time (and still looks like something from the next century let alone our own one), and that the Saturn V rocket remains the most powerful ever used in a crewed mission and the only launch vehicle to take humans out of LEO. Both were designed on pen and paper, and succeeded with minimal computational capabilities by today’s standards. It might seem that in the 60 years before 1970 aerospace was a burgeoning exponential, but that in the 60 years since has petered out with diminishing prospects. With a world far wealthier, better-educated, more populated, and 10^9 times more computationally powerful, why has nothing seemingly improved that much? Am I being nostalgic?
Part of this is perception. There have been some very welcome improvements: notably the cost and safety of civilian aviation. Inflation-adjusted ticket prices are down about 5x since the second world war which is pretty remarkable for an industry that is both highly capital-intensive and heavily regulated. Last year saw a rate of 1 accident per ~900,000 flights and a fatality risk of 0.06 per million flights, orders of magnitude safer than in the early jet age. These are brilliant developments but certainly less headline-making than the size or speed of a vehicle. Unsurprisingly, millions of aviation geeks around the world have posters featuring the Apollo missions or great combat aircraft of the 20th century, not many have a graph showing the decline in accident rates per million miles or airline ticket deflation.
As brilliant as these cost and safety developments have been for humanity, they are essentially an optimisation. We’re near the point where aircraft in their current form can’t get much safer statistically, with just the occasional freak accident skewing the averages. Nor is there some magical pricing scheme yet to be engineered that will allocate seats at vastly cheaper prices than current levels, because load factors are almost entirely squeezed. Continuing to make engines x% more efficient, reliable and quieter each year for a few decades yields the noticeable improvements, but it won’t drive an order of magnitude shift in possibilities. There are asymptotic limits to the efficiency of the current tube-and-wing, fossil-fuel, turbofan engine standard for modern passenger jets. What is instead needed is not an x% improvement but a step change improvement with an entirely different technology. This can come from two broad domains: engineering by the manufacturers (ie. ground-breaking new planes with totally novel propulsion and structure), or, from operations by the airliners (ie. increased automation of flight routes, aided by benign regulations and taxes).
Looking for such a dramatic degree of progress does beg the question, why does this matter? A more luddite view might think that aviation has already been made much safer and cheaper, and that there are far more important issues to solve elsewhere. It’s estimated that only about 20% of the world have ever been in a plane, that 10% fly annually and around 2-4% fly internationally annually. A lowly single digit percentage of the world has the financial (let alone political) freedom to fly abroad easily. The cynical and de-growther conclusion is that aviation is an industry for the world’s richest that has already benefitted them enough. Any further gains would just accrue to the small wealthy fraction of the world’s population that account for most air traffic.
This scarcity mindset is looking at the market the wrong way round. The fact that aviation still remains prohibitively expensive for most people and freight implies the opposite: that there is vast latent demand waiting to be captured by moving billions of people and products at lower prices. In a future world where aviation is 10x cheaper, it would become a mainstream part of the world’s middle and lower classes and just not a privilege for a tiny few. It is by the same logic that taxes on government flight are so woefully misguided. The idea goes that because flying is dominated by the wealthiest individuals, it must be a progressive tax. As such, market distortion is minimised because the tax does not dramatically impact the inelastic demand from richer customers. This may have been more true decades ago when – with or without taxes – flying was much more expensive. But for budget airline flights in UK and Europe nowadays, taxes can constitute over half and perhaps 75%+ of a ticket on some occasions. This extinguishes demand, with large cost to society and the economy, all for the benefit of just £4.7bn in tax receipts in the case of the UK government. This is less than half of a percent of government receipts; enough to fund the NHS for a week, or the variable amount that interest payment costs oscillate by over a few weeks. The same problem of government taxes bumping up consumer prices exists in the US but to a lesser extent, in the range of 10-20%.
Flight costs are down about 10x in real terms since the advent of commercial aviation in the 1920s. That underestimates the real growth in consumer surplus because it is dramatically safer, more comfortable, more convenient, and more plentiful in terms of options. But for simplicity we’ll stick to cost per mile since those above factors have largely been exhausted.
This brings us to the question of price from the consumer perspective. What is the average breakdown of a ticket? Let's model a standard US domestic flight as a $100 one-way ticket.
It can be divided into 3 chunks;
$25 of that is for fuel
$25 is for labour
$20 is the plane itself (leasing and maintenance)
$30 is a mix of overheads, air traffic control services, wafer thin profits, and government taxes.
We can distil these down further into two, roughly evenly-sized groupings:
Part I: That means that $50 (fuel + aircraft costs) are ultimately a function of aeronautical engineering.
Part II: The remaining $50 of labour, taxes, airport fees, business overheads and profits is a matter for the airlines, aka, the operators.
Fuel straddles these two definitions. On a day-to-day basis fuel is paid, hedged and consumed by the airline very much as an operational cost. But in the long run, fuel’s fundamental contribution to eventual consumer prices is driven by propulsion efficiency, which is a matter for the engineers who make the planes. It is the operator’s job to hedge fuel price volatility as they see fit – something I will assume cancels itself out over the long run.
So there we have are two questions to address:
Part I: How do we make brilliant new aircraft that are lighter, quicker, more efficient, and, above all, cheaper?
Part II: How do we operate these aircraft in the most efficient, low-cost and consumer-friendly market possible?
Part I
In transitioning from propeller in the 1920s to turbojet and the onto turbofan (and turboprop), fuel efficiency has increased from 10 to 90 passenger miles per gallon. Of the ~80pmpg gains, this was driven by two major changes: going from propeller to jet, and optimising the jet engine away frmo turbojets towards turbofans with wider bypass ratios. From now on I'll refer to efficiency in passenger miles per megajoule to accommodate comparisons with non-fossil fuel propulsions. 100pmpg = 0.76 pgmj.
On the current trajectory of aviation innovations, we can expect certain features emerging: blended-wing body fuselage, open-rotor or wider bypass turbofans, and greater use of composites, for example. These improvements will continue to drive efficiencies and ultimately consumer surplus, although they are not dramatic enough for our intended purpose of radical improvement.
I will entertain revolutionary ideas for civil aircraft propulsion methods. Disclaimer: while I have a fair intuition of aeronautical engineering as a pilot and amateur physics student, I am not formally taught in the field. I am not in a position to truly champion or condemn different technologies.
I’ll start with a technology that is very much physically possible although unfeasible in current circumstances: pure battery and electric propulsion. Electric systems are ~85% efficient tank to shaft, vs only ~45% for combustion engines. Electric systems could also easily be carbon-neutral in their impact, they could produce far less noise, and they are mechanically far more simple. So why don’t we have use electric planes? Effectively because of one reason: energy density.
During takeoff, an A380 consumes the equivalent of up to 130MW in thrust, which is enormous for something only 70m long with 500 people on board. Bear in mind that we are talking about output here, not power input from tank. As tank-to-shaft efficiency for fossil fuel turbofans is 40% as mentioned, we can estimate tank-to-propulsor at around 35%. That implies ~400MW worth of fuel is being consumed at any given moment (ie 400MW = 400,000,000 J/s = 12L of kerosene consumed per second). 400MW is enough to power the baseload demand of an entire city the size of Kansas or Glasgow. To use batteries for that would take us to some absurd requirements.
Because I think batteries could have an important role in aviation in the current decades we will go further into the maths here specifically so we can really understand the trade-offs at play. Bare in mind that this is for induced power only (the power required to overcome induced drag as opposed to parasitic drag which dominates at higher speeds).
We can start with a few axioms for steady and level flight:
L = W = Mg
Dynamic pressure: q = ½ * 𝜌 * V^2
Aspect ratio = (b^2) /S
The induced drag coefficient can be represented as:
CDi = Cl^2 / (π * e * AR), where CL (coeff. of lift) = L/qS
Substituting in CL and the AR we get:
CDi = L^2 / (π * e * q * b^2)
We know that power is the rate of doing work over a given distance, which we can model as overcoming induced drag at a given velocity: P = D * V. Substituting in that and dynamic pressure for a function of P we get:
P = 2L^2 / (π * e * 𝜌 * b^2 * V)
And finally putting in the assumption of steady and level flight:
Induced power = Pi = (2 * M^2 * g^2)/(π * e * 𝜌 * b^2 * V)
Where M on the numerator represents mass off the aircraft, and b represents the wingspan.
The M^2 is what drives batteries to be so difficult for aircraft. Larger batteries result in a higher mass, which necessitate higher induced power, more batteries and so forth. The L^2 could help but there is a limit to how long wingspans can reasonably get with airport infrastructure and we’re already close to that limit. Even if that wasn’t an issue wingspan is a linear concept that would increase far slower, proportionally, than mass which is a cubic concept. If we take leading edge energy densities of ~0.3 kWh/kg, a C-rate of 3, then an A380 would require about 150 tonnes of batteries just to hit take-off energy requirements of 130MW. Then you’d be staying in the air for 20 minutes at best. This tells us that power (kW/kg) isn’t the bottleneck, it is energy dense storage (kWh/kg). It’s also worth noting that both specific energy and energy density scales poorly; they are ~30% lower at the pack level rather than the cell level as thermal management becomes an additional challenge.
Either way, a pure-play electric jumbo jet is off the cards any time in the next few decades or until we see some radical innovation in highly energy dense electrical storage. In the history of electrical storage, we have gone from 2Wh/kg to about 350Wh/kg. Getting the 40x improvement to reach the 12000Wh/kg of kerosene is a bold ask. It may happen, but current rates of doubling density every 30-40 years are hardly Moore’s Law standards of growth. Any diminishing marginal returns will make that even harder. But even on current trends electricity will likely still have a key impact in the nearer term via other means.
I’ve been following the startup Astro Mechanica closely since it launched publicly. The company, which most recently raised $27m in April 2025, proposes a novel type of engine design they call the turboelectric adaptive engine. The key novelty of this design is using an exogenous, electrically-powered compressor that adapts to the current speed, rather than having one mechanically in-built in a conventional jet engine.
If you plot specific impulse against speed for different propulsion methods, you’ll see that fuel efficiency is generally high at low speeds and vice versa. I say generally because this consists of piecewise distributions for each cycle – it isn’t one single clean 1/x distribution. There are local fluctuations within each of these, for example a ramjet’s maximum Isp is at ~Mach 3 despite being used between Mach 2-4, but across all air-breathing propulsion techs from Mach 0 to Mach 10+ there is a broad trade off between Isp and speed.
What AM proposes to do is ride the most efficient frontier of this trade-off between all speeds from Mach 0-3 by using an exogenous electric compressor that adjusts the pressure ratio to the appropriate speed. Critically this means this isn’t mechanically tied to shaft rotation speed, and so can effectively act as a ‘gearbox’ as it transitions between turbofan, turbojet, and ramjet.
Clearly there are big engineering risks here: physical limits of the Brayton cycle, cost/benefit of electroadaptive engines versus the added weight, complexity of temperature and flow management, among others. It’s not clear yet that the added weight and challenges of having a separate electric compressor will yield sufficient efficiency gains.
What does make me fairly bullish is this additional electric system will almost certainly continue to get more dense, and cheaper, and so AM has positioned itself to benefit from positive learning rates in electrical engineering. The most pertinent question for the business’ survival will be if they timed it correctly.
Reaction Engines was a British aerospace startup developing Synergetic Air-Breathing Rocket Engine (SABRE) engines which has since gone into administration last year. The company was founded in 1989 to carry on the work of the Horizontal Take Off and Landing (HOTOL) programme that the British government had withdrawn support for. The basic premise was a single stage to orbit (SSTO) ‘spaceplane’ that promised a more ambitious version of what Astro Mechanica plans taking it one step further – a plane that evolves from turbojet through to ramjet, before opening LOX valves and transitioning to a rocket engine as the plane leaves the atmosphere.
It is important to emphasise that this was not a totally insane idea. It was one that garnered the attention from DARPA, Boeing, ESA, the British government, Rolls-Royce and BAE, and no doubt plenty of other actors secretly. It was valued at around £300m at its peak. The company, although never completing a full SABRE engine, passed most subsystem tests strongly and received positive feedback from independent tests. Its founders were the best in British industry, and the next generation who had been taught by the builders of the Black Arrow and early British pioneers in nuclear rocketry research.
The main bottleneck on the engineering side was thermoregulation of air at such speeds. A nickel and inconel tubing system would reduce incoming air temperatures from over 1000*C to under 150*C in about 50 milliseconds, allowing safer compression later on in the cycle and saving the use of on-board oxygen supplies to the latest possible moment when the exiting the atmosphere.
The system promised a lot. Halving oxidiser mass to reach LOE. Per kg payload costs about equal to the Falcon Heavy. If economies of scale could be realised and reusability optimised, it could have tremendous learning rates.
My scepticism would be that the company requires a lot of capital ($12bn by estimates 10 years ago) and technical risk to produce a fully functioning, economical fleet of Skylon spaceplanes (a product that is fundamentally unproven). The payload costs were a lot more convincing in a world without SpaceX; now the risk-adjusted difference in payload costs is meagre for such a speculative venture. What excites me on the project if it carries on in a new form is that a spaceplane could be a step change in aerospace for the fact that such bold combined-cycle design has never been pioneered. It is possible to imagine spaceplanes outcompeting parts of the launch market if the most optimistic (but physically possible) forecasts can be reached. And whilst a Skylon-equivalent spaceplane would be uncrewed, if it could produce another step improvement in launch costs it would unlock even larger markets beyond what SpaceX has opened up. With regards to civilian aviation, the development of SABRE (or at least of its subsystem components like rapid precooling) would offer new progress in propulsion.
There are other high profile startup in the aerospace world like Boom, Hermeus. Unlike the aforementioned companies, these supersonic companies are facing more market risk rather than engineering risk. Boom has already built a prototype aircraft that goes faster than Mach 1. It’s impressive to see, and it’s definitely catalysed some political and financial interest in aerospace which is a positive thing. However, this fundamentally isn’t a huge step forward given that we had Concorde flying back in the 1960s. Even if these companies do survive, they will only ever cater for a tiny percentage of people who can afford to value their time so highly. It will be a romantic sight for any aerospace geeks but it's not a step change for the sector. Since the vast majority of people are sensitive to price above all else (especially those among the latent demand who cannot afford to fly yet), it is lower cost at scale that would be the real breakthrough in aviation – not speed.
There’s a few other ideas worth considering:
Hydrogen is 2.7x as energy dense as kerosene, requires considerable but not total alterations to current turbofan engine designs, and can be produced in a sustainable and scaled manner. But it is dangerous and difficult to handle for the reason that, as the smallest naturally occurring molecule, it constantly leaks from any storage. And in its liquid form it needs to be stored at –253*C. While energy dense by mass as stated, it is not energy dense by volume (four times more volume for the equivalent energy of kerosene). I’m very pessimistic on hydrogen’s feasibility. Its inherent characteristics are punishing for any engineering project as it is. Even if it had very attractive traits as a fuel right now, you still have an enormous, multi-trillion dollar effort that spans every airport in every country and completely rewires global energy grids to produce and then transport this fickle molecule. It’s hard to see hydrogen every becoming so attractive that it warrants essentially unanimous international cooperation and investment. And while this world I am envisaging with 10x more air traffic would demand huge increases in airport and energy infrastructure, that is easier in the sense than any given country is incentivised to build more planes and airports independent of other countries. No intricate cooperation is required, the industry incrementally grows capacity in a quick and dispersed manner because individual incentives are aligned to build capacity regardless of the behaviour of other actors. No country is going to build a hydrogen-based aviation ecosystem unless every other country agrees to as well.
Materials are obviously a crucial part of any aerospace endeavour. Major breakthroughs would be crucial to hydrogen as above, for the mentioned reason of storage and transport. A new material could also overcome the trade-offs between weight, durability, strength, and temperature. Clearly breakthroughs in battery density and superconductors would revolutionise electric contribution to flight as mentioned. There are also other more specific considerations for materials such as lightning protection and tank crashworthiness which can prove to be serious bottlenecks for regulators.
Lastly there are other more wacky ideas you hear about in aerospace but do not properly apply to civil aviation: nuclear, plasma, ion etc. These aren’t being considered seriously by any government, company or regulator.
Here’s a conclusion of what exactly we should want to see on the manufacturing and engineering side.
1-2 order of magnitude of improvement in electrical energy storage density over current leading Li-ion batteries.
SABRE or turboadaptive technology to drive large efficiency gains in combined cycle propulsion.
A materials breakthrough either to allow easy handling of hydrogen, or one that allows for radically stronger, temperature-tolerant and lighter airframes.
Part II
Let’s rewind all the way back in this piece. Despite the miracle technologies we might see introduced into civil aviation this century, their actual impact will be decided by the economic and regulatory environment in which they operate. Even if, absurdly, the cost of the aircraft and its fuel tended to zero your plane ticket would barely halve in price. We need to address how the companies operate and how their variable costs could be reduced.
Automation would be a boon to an airline’s cost structure. Unlike the stochastic and intuitive nature of driving cars on roads which has proven extremely hard to totally automate, piloting commercial aviation is far more procedural and therefore ripe for total automation. Flights now are largely flown on autopilot anyway, and the crashes that do occur are often due to human errors. If standby pilots are still required for regulatory reasons, then one pilot can remotely monitor many planes staggered by route timings and take-offs/landings where extra attention is required. The infrequency of intervention needed means multiple flights could be assigned to one pilot. Likewise routine flight attendant duties could be automated easily given the systematic layout of a plane. Again, for regulatory reasons, airliners may still provide one attendant for health and safety, order, or edge cases of customer service. Ground handling could be automated in a similar fashion with edge cases handled by a small on-site team, but the bulk of handling and sorting down automatically.
All of this could quite feasibly be 90%+ automated by 2050 with where trends in robotics are going. The main barriers would not be technical, but regulatory and social. We would have to hope that regulatory authorities would eventually tolerate near full-automation of flights, with sufficient trialling data. This is where cargo planes would make for great testing ground, with all the demand to achieve lower transport costs and none of the risk of passenger safety. And as for social acceptance, it may be an odd or unpopular sight at first to have automated and teleoperated pilots or automated flight attendance. But the price sensitivity of consumers would make this transition inevitable for economy-class fliers.
Sitting somewhere between engineering costs and operational costs, intelligent diagnostics systems would cut costs further. You could have these diagnostics systems improve with more training data, or digital twin engines running in real time. You could squeeze unplanned downtime as close as possible towards zero.
One aspect so far ignored is the airport infrastructure itself. The largest airports around the world are already squeezed at 80% to 95%+ capacity in terms of flights per runway per hour. Building additional runways is only a small part of the problem. Even if that could be achieved, new bottlenecks would then emerge with the capacity squeezed for terminals, transport, security and baggage handling.
The solution is a vast build out of purpose-built, modern airports required to facilitate a 10x in civil aviation passengers and cargo. With this means of designing airports from scratch rather than papering over the cracks of airports from almost a century ago, as happens entirely in developed countries. Airports that could be built to allow blended-wing body aircraft to navigate them for example, for high voltage charging points for electric features on the planes, or to integrate the automation of tugs and passenger processing from first principles into the architecture of the site. Having airports operating at a larger scale but with automated services would drive down their costs. Yet having more of these airports would also increase competition between the, driving down fee pricing power over airlines, freight companies, and consumers and ultimately passing on these lower costs onto the passenger in the form of lower fees.
Clearly there are regulatory barriers at the national level as building even just one new airport is perhaps the most ambitious infrastructure project possible in many countries and is a multi-decade project. Yet this is easier to solve that, say, the infrastructure challenge of building out hydrogen refuelling, because incentives align at the national and regional level – no international cooperation is necessarily required. If civil aviation becomes so widespread, and if air freight costs are driven down sufficiently, then the incentives to build out such infrastructure will grow enormously as the loss of economic competitiveness would become too great. Countries wouldn’t have to come together on any complex framework and timeline, they would simply set out about building dozens of new airports of their volition.
Conclusion:
Short term: build BWB aircraft, cut taxes on airline tickets to zero, build out more runways, accumulate maximum data on all piloting and airport service jobs.
Long term: electrify aircraft to maximise propulsive efficiency, explore adaptive jet and combined cycle engines, automate all flying and service jobs as far as regulations and technology allows, build out hundreds of vast and automated airports.