A jet airliner of 2026 cruises no faster than the Boeing 707 that opened the jet age. Engineers did not run out of ideas: near Mach 0.85 the drag curve meets the fuel bill, and every time the industry was offered speed, it bought a cheaper seat instead. Physics set the price of speed; the market declined to pay it. That price is now being reopened.
The Airliner That Broke the Sound Barrier
On 21 August 1961, a Douglas DC-8 climbed high over California, pushed its nose into a shallow dive and touched Mach 1.012. It was a production airliner, the same type that would soon carry holidaymakers to Honolulu, and for a few seconds it flew faster than sound. A few seconds in a dive is not a supersonic cruise, and nobody at Douglas claimed otherwise. But the flight made a point that would matter for the next sixty-five years: the sound barrier was not what kept airliners slow.
Here is the strange part. The fastest airliners in regular service in 2026, the Boeing 787, the Airbus A350 and the A380, cruise at about Mach 0.85, roughly 900 kilometres an hour at altitude. The Boeing 707 that Pan Am put on the New York–Paris route in October 1958 cruised at about Mach 0.80 to 0.82. That is a gain of roughly six percent in almost six decades. The single-aisle jets that now carry most of the world’s passengers, the Airbus A321neo and the Boeing 737 MAX, have typical cruise speeds of around Mach 0.78 to 0.79. These are different classes of aircraft built for shorter routes, and the comparison is between typical cruise speeds, not averages of every flight. Even so, many of the newest single-aisle jets passengers board are, on paper, slightly slower than the first.
Most machines of the twentieth century kept getting faster. Commercial aviation did so from the Wright brothers to the 707, and then the line went flat. The one exception, Concorde, flew at twice the speed of sound for twenty-seven years, retired in 2003 and has had no successor.
The usual explanation is that engineers hit a wall. That explanation is incomplete: the walls are real, but they are walls of cost, not of possibility. Engineers did not stop. Newly delivered jets burn about 43 percent less fuel per tonne-kilometre than the new jets of 1970, according to the International Council on Clean Transportation. The progress happened. It went somewhere other than speed.
This Dispatch argues one thing. Cruising speed has been flat for half a century because, every time speed and the cost of a seat pulled in different directions, the industry chose the seat. The physics determined what faster flight would cost. The market determined whether it was worth paying. The plateau is the record of that second decision, taken again and again by airlines, manufacturers and regulators, in 1970, 1973, 2002 and 2003. Speed is one of the clearest cases in modern technology of a capability priced out rather than engineered out, and the argument makes a prediction that can be checked: speed comes back where, and only where, the value of the time saved exceeds what it costs to save it.
The Wall That Is Really a Slope
Start with the physics, because it sets the price.
An aircraft’s speed is usually expressed as a Mach number, its speed as a fraction of the speed of sound in the surrounding air. For an airliner, though, the number that matters is not the speed of the aircraft but the speed of the air over its wing. As part of producing lift, a wing accelerates the air flowing over its upper surface. So when the aircraft is cruising at Mach 0.8, the air over the top of the wing can already be moving faster than sound.
When that happens, small shock waves form on the wing. They add a new kind of drag, called wave drag, and they can make the airflow behind them break away from the surface. The result is a steep rise in drag as the aircraft approaches the speed of sound. Engineers call the point where it begins the drag-divergence Mach number. Below it, a little more speed costs a little more fuel. Above it, a little more speed costs a lot more.
This is a slope, not a wall, and people climbed it. Military jets and Concorde went through the sound barrier with thin, sharply swept wings and very large engines. The question was never whether an airliner could go faster. It was what each extra knot would cost, and who would pay for it. That is the rule for reading everything that follows: physics sets the price of speed; the buyer decides whether to pay it.
The most revealing chapter concerns a wing that could have bought speed and was used to buy something else. In the 1960s the NASA engineer Richard Whitcomb developed the supercritical wing, a profile with a flattened top and a curved trailing edge that delays and weakens those shock waves. NASA flew it on a modified F-8 from March 1971. That test wing was optimised for cruise at Mach 0.98, a whisker below the speed of sound. NASA described the gain as one that could be spent three ways: on more speed, on more range or on less fuel.
The industry spent it on fuel. Supercritical aerofoils became standard, but they were used to build wings that were thicker, lighter and roomier for fuel tanks, cruising at the old speeds while burning less. Mach 0.98 stayed in the NASA archive. Mach 0.85 went into service.
The reason is that an airline’s bill is written at the margin. Flight computers carry a setting called the cost index, the ratio between the cost of time and the cost of fuel. The cost of time is mostly crew paid by the hour and maintenance charged by flight hour; the aircraft’s capital cost counts only when flying faster lets the airline fit more trips into the aircraft’s day. The computer picks the cruise speed that minimises the total. In practice airlines cruise just above the speed that gives maximum range, because past that point the fuel curve rises fastest.
Published performance data for the Airbus A321neo show it burning about 2,200 kilograms of fuel an hour at Mach 0.76 and about 2,440 at Mach 0.80, at the same altitude and weight. Five percent more speed costs eleven percent more fuel per hour, or about five percent more fuel per kilometre. In 2008 Northwest Airlines reported saving 162 gallons on a single Paris–Minneapolis flight by cutting its average speed from 542 to 532 miles per hour. The flight took eight minutes longer.
The other half of the calculation is what those minutes are worth, and here the economics is less obvious than it looks. Time saved has value only when it changes something. Twenty minutes off a flight that arrives at a hub to wait for the same connection is worth almost nothing to the passenger. Twenty minutes that let an aircraft fly an extra rotation in a day are worth a great deal to the airline. A faster flight that lands before an airport opens, or departs at an hour nobody wants, can be worth less than a slower one. Speed is sold not in minutes but in schedules, and most schedules have little room to use it.
Engines push in the same direction. The high-bypass turbofan, which entered airline service with the Boeing 747 in 1970 and now powers every large airliner, works by moving a very large mass of air at modest speed. That is the most efficient way to make thrust when the exhaust is not much faster than the aircraft, which is the situation at Mach 0.8. At Mach 2 the opposite holds: you need a slim engine that throws a smaller mass of air out very fast. Concorde flew on turbojets, and Boom’s proposed Symphony engine is described as a medium-bypass design. Every gain in turbofan efficiency since 1970 has widened the cost gap between subsonic and supersonic flight. Airframe and engine locked each other in.
The Bill That Chose
The order of events matters.
The plateau began before oil became expensive. On 22 January 1970 Pan Am flew the first commercial Boeing 747 service from New York to London, with 332 passengers, in 6 hours and 14 minutes in the air. The 747 cruised at about Mach 0.84, barely faster than the 707, but it carried more than twice as many people on high-bypass engines that burned far less fuel per seat. The bet on bigger and cheaper, rather than faster, was placed while oil was still cheap.
Then fuel became expensive. During the 1973 embargo crude went from about $2.90 a barrel to $11.65 by January 1974, according to the Federal Reserve’s history of the shock. In 1979 it happened again: the price of kerosene-type jet fuel sold by US refiners rose from about 40 cents a gallon in January to 72 cents in December, an increase of 80 percent in a year, on the Energy Information Administration’s series. Airlines lowered cruise speeds, planned flights by computer and filled their aircraft more densely. In 1976 the US government launched NASA’s Aircraft Energy Efficiency programme, worth about half a billion dollars, split between engines and airframes and aimed at burning less fuel. Nobody set up a programme to fly faster.
1973 brought a second decision that is easy to forget. That year the US Federal Aviation Administration banned civil supersonic flight over American land, in the rule now numbered 14 CFR 91.817. The cause was the sonic boom, the double crack a supersonic aircraft drags along the ground beneath it. The ban did not prohibit supersonic aircraft; it prohibited them from doing over land the one thing they were built for. Speed was now expensive in fuel and illegal over the largest domestic aviation market in the world.
The efficiency gains that followed were real but uneven. The ICCT’s long-run series shows average fuel burn of new aircraft falling about 2.6 percent a year in the 1980s, barely moving between the mid-1990s and the mid-2000s, then improving again as the 787, the A350 and the re-engined single-aisle families arrived. A genuinely new design has typically burned about fifteen percent less than the one it replaced. Since 2020, the ICCT notes, improvement has stalled again, largely because few new types have entered production. Fuel has not become a smaller concern: the International Air Transport Association’s June 2026 outlook puts it at 31.4 percent of airline operating costs this year.
The most revealing evidence comes from 2001, when Boeing put a choice in front of its customers. Its Sonic Cruiser was a delta-winged airliner designed to cruise at Mach 0.95 to 0.98, up to fifteen percent faster than any jet in service, while matching the per-passenger fuel efficiency of the best twin-engine widebodies of the day. In effect, Boeing proposed to sell speed at roughly the old price. The September 2001 attacks then devastated airline finances, and the answer came back in stages. By October 2002, according to Flight International, not one airline in Boeing’s advisory group gave a high rating to the aircraft’s cruise speed. Boeing cancelled the programme on 20 December 2002 and in January 2003 announced the 7E7, a conventional airliner using the same advanced technology, whose baseline design was expected to burn about ten percent less fuel than the Sonic Cruiser. The 7E7 became the 787 Dreamliner. It cruises at Mach 0.85.
This was not a controlled experiment; the timing, in the worst airline downturn in decades, weighted the scales. But as commercial tests go it is unusually clean: one manufacturer, one set of technologies, one group of customers, offered two products that differed mainly in speed and in seat cost. The customers took the cheaper seat.
The Machine That Proved It
Concorde is usually treated as the exception to the plateau. It is better read as its price tag.
Concorde entered service on 21 January 1976, with British Airways flying from London to Bahrain and Air France from Paris to Rio de Janeiro. It cruised at Mach 2, about 2,170 kilometres an hour, and crossed from London to New York in about three and a half hours, half the time of a subsonic jet. Its fastest crossing, on 7 February 1996, took 2 hours, 52 minutes and 59 seconds. As engineering it worked, and kept working for twenty-seven years.
As a business it found a market, but a narrow one, and never one large enough to carry the programme without public money. Airlines had once held options on dozens of Concordes; the Smithsonian’s account of the programme says they dropped them once they calculated the operating costs. Twenty were built and fourteen entered airline service, seven each with British Airways and Air France, the state-owned flag carriers of the two countries that had paid to develop it. Both airlines acquired their aircraft from their governments at virtually no cost. The programme cost more than £1 billion in the money of the 1970s, roughly ten times the 1962 estimate of £150 million to £170 million.
The reason shows in one number. At full load, Concorde delivered about 16 passenger-miles per US gallon of fuel, against about 33 for the Boeing 707, 46 for the 747 and 54 for the DC-10, in figures compiled by the US Office of Technology Assessment in 1980. Concorde carried a hundred people at roughly a third of the fuel efficiency of the jumbo jet that had entered service six years earlier. The overland ban confined it to ocean crossings, which in practice meant the North Atlantic.
It is often said that Concorde never made money. In fact British Airways ran it profitably for long stretches, selling seats to people for whom three and a half hours were worth almost any price. By 2003 the standard return fare between London and New York was £6,636, according to the BBC. But those profits were possible only because the taxpayers of two countries had already written off the cost of the aircraft. Speed paid its way only once someone else had paid for the machine.
The end came through demand and maintenance, not physics. On 25 July 2000 Air France Flight 4590 crashed after take-off from Paris, killing 113 people, the type’s only fatal accident. The fleet was grounded, modified and returned to service in late 2001, weeks after the September attacks had hollowed out premium travel. In April 2003 Air France’s chairman, Jean-Cyril Spinetta, said the airline had recently been filling about 20 percent of its Concorde seats and that its cost per seat-kilometre had risen 58 percent since the crash. Airbus, by then the type’s manufacturer, warned that keeping it flying would require an enhanced maintenance programme. Air France flew its last service on 31 May 2003, British Airways on 24 October.
What Got Faster Instead
The strongest objection to this argument is that air travel did get faster. It did, just not in the cruise. To see where, it helps to separate three clocks: air time, the minutes the aircraft is actually flying; block time, from gate to gate, which is what the timetable shows; and journey time, from departure city to arrival city, stops included.
Consider Paris to Tokyo. Air France’s first service, in November 1952, flew a southern route with stops across the Middle East and Asia and took about 50 hours from departure to arrival, of which about 38 and a half were spent in the air. By 1958 a polar route via Anchorage had cut the journey to nearly 30 hours, still on piston-engined aircraft. In April 1970, Air France and Japan Airlines began flying across Siberia with Boeing 707s, stopping briefly in Moscow; Le Monde reported the journey at about 13 hours 45 minutes including the stop. Cruise speed barely changed after the first jets. What changed was the route.
That is the general pattern. The largest gains in journey time since the 1960s came from range: aircraft able to fly far enough to skip the refuelling stop at Gander, Anchorage or Bahrain. On a 5,500-kilometre crossing such as London to New York, flying at Mach 0.85 instead of 0.80 saves only about twenty minutes of air time. Removing a stop on a long route saves hours. Qantas plans to start daily nonstop flights between Sydney and London in October 2027 with modified Airbus A350-1000s, on flights of up to 22 hours, and expects to save up to four hours against today’s one-stop journeys. The aircraft will cruise at Mach 0.85.
The other great gain was access. In real terms the average US domestic fare, fees included, has fallen by nearly half since the late 1970s, by the count of the industry body Airlines for America. Whether deregulation caused that fall is disputed, since real fares were already falling before 1978. That bigger, more efficient aircraft made the seat cheaper is not. For most people, the speed gained in the jet age was the change from not flying at all to flying regularly.
Some of what looks like slowness has nothing to do with the aircraft. A study of 43 million US domestic flights by researchers at Northwestern’s Kellogg School and Washington University found that published flight times grew 8.1 percent between 1997 and 2017, though the aircraft were not flying more slowly: airlines were lengthening block times so that more flights would arrive “on time”. OAG found that in 1996 no flight between Los Angeles and San Francisco was scheduled for longer than 90 minutes; by 2015 nearly half were scheduled for 91 to 110. Since January 1973 every passenger on a US flight has passed through screening, and airlines now advise arriving two to three hours before departure. Cruise speeds have barely moved; the ground has grown.
Geography has also turned against speed. Since 2022 the closure of Russian airspace to most Western airlines has pushed flights between Europe and East Asia onto longer paths. The two Air France nonstops from Paris to Tokyo checked for this Dispatch are scheduled at 13 hours 35 minutes and 14 hours 20 minutes. The 1970 journey, Moscow stop included, sits within minutes of them.
The objection, in the end, describes the trade rather than refuting it: range, frequency and cheap fares are exactly what an industry optimising the cost of a seat-kilometre would produce.
Repricing Speed
If the plateau is a price rather than a wall, it should move when the price moves. As of 4 November 2026, three efforts to move it are under way: on the law, on the noise and on the aircraft.
The law first. On 6 June 2025 Executive Order 14304, “Leading the World in Supersonic Flight”, directed the FAA to repeal the 1973 overland ban and replace it with a noise-based standard. The FAA’s proposed rule, “Enabling Supersonic Overland Flight”, was published in the Federal Register on 2 July 2026. It is not a simple opening of the skies. An operator would have to show that no sonic boom with an overpressure above 0.11 pounds per square foot reaches the ground, demonstrate it to the FAA, receive a finding from the Administrator and then fly within FAA-set conditions. The comment period closed on 17 August 2026. No final rule had been issued by the date of this Dispatch, and until one is, the 1973 prohibition stands; the FAA has said it aims to finalise its supersonic rules by mid-2027. Even then, the rule would remove a legal barrier. It would not certify any aircraft.
The noise next. NASA’s X-59 is a single-seat research aircraft with a long, needle-like nose, shaped so that the shock waves it creates do not merge into one loud boom but reach the ground as a softer thump. It first flew on 28 October 2025, went supersonic on 5 June 2026 and reached Mach 1.53 in September. By its first anniversary it had flown 41 times. Its next phase, acoustic validation, has not yet begun; after it, NASA plans to fly over selected American communities and ask residents what they heard.
Then the aircraft. Boom Supersonic flew its XB-1 demonstrator past the speed of sound on 28 January 2025. Its proposed airliner, Overture, is designed to cruise at Mach 1.7 with 60 to 80 passengers over 4,250 nautical miles. Boom lists 130 “orders and pre-orders”, a total that includes options and conditional commitments; United’s 2021 agreement covers 15 aircraft plus 35 options, and American’s up to 20 plus 40 options. On 22 October 2026 Boom completed the first prototype core of its engine. Six days later United’s chief executive, Scott Kirby, said of the target to enter service in 2029: “It’s not going to happen in ’29.”
Overture’s specification carries the thesis inside it. Boom’s original design called for Mach 2.2. The current one flies at Mach 1.7, slower than Concorde in 1976, because a slower supersonic aircraft can use a more efficient engine, burn less fuel and aim at business-class rather than first-class fares. The revival, too, is optimising cost per seat. It is drawing the line at a different point on the slope.
One more detail completes the picture. Boom has adapted the core of its Symphony engine into a 42-megawatt gas turbine called Superpower, to generate electricity for AI data centres. Its launch customer was the data-centre developer Crusoe, which agreed to take 29 turbines in a deal valued at about $1.25 billion; in February 2026 Baker Hughes announced that Boom had ordered the generators to go with them, bringing the package to about 1.21 gigawatts. At the end of September 2026 Crusoe walked away from the launch partnership. Boom says it will still deliver about 250 megawatts to other sites next year and is targeting a gigawatt in 2028. The engine meant to buy back the speed of travel is looking for revenue elsewhere, and finding that the side business has a demand problem of its own.
| Front | Status, 4 November 2026 | Still missing |
|---|---|---|
| Law FAA, 14 CFR 91.817 | Proposed rule published 2 July 2026: boom overpressure at ground ≤ 0.11 psf, operator demonstration, FAA finding and conditions; comments closed 17 August | Final rule: none issued; FAA target mid-2027 |
| Noise NASA X-59 | 41 flights; supersonic since 5 June 2026; top speed Mach 1.53 | Acoustic validation; community overflights |
| Aircraft Boom Overture | Design Mach 1.7, 60–80 seats; 130 orders and pre-orders incl. options; first engine core completed 22 October | Airframe, engine tests, certification; United’s CEO: “not going to happen in ’29” |
| Engine, other use Boom Superpower | 42 MW turbines for data centres; launch customer Crusoe (29 turbines, about $1.25bn) left the partnership in late September | Replacement customers; Boom targets about 250 MW in 2027 and 1 GW in 2028 |
Reading the Plateau
The frame travels well beyond aviation. Carry it into the next headline about a faster aircraft, a quicker train or any technology that seems to have stalled, and it resolves into three questions.
First: who pays for the minute? Every increment of speed has a price in fuel, noise and capital, and a value that depends on what the minute saved changes. The question to ask of any revival is not whether it can fly faster, but who will pay for the minutes, and how many of them there are.
Second: which clock? A longer flight on a booking screen does not prove that aircraft have slowed. It may reflect a padded schedule, a longer route or a queue on the taxiway. Cruise speeds have barely changed since the 707; block time and journey time have moved for other reasons, in both directions.
Third: price or permission? Physics sets a slope, not a wall. Regulation, as in 1973, can add a wall of its own. The current American rulemaking would lower one barrier, under conditions, and leave the others in place.
The argument also says how it could be wrong, and where it stops. Three explanations compete for the plateau. The first is economic: demand has not paid for the extra cost of speed. The second is regulatory: speed might pay, but rules such as the 1973 overland ban prevent anyone from selling it. The third is industrial inertia: even where demand exists, transition costs, risk and capital already sunk into subsonic fleets slow adoption. The evidence assembled here shows that the first has dominated. It does not show that it acted alone; the overland ban is regulation, and the speed with which airlines retreated from the Sonic Cruiser after 2001 owed something to risk.
The claim is therefore not that speed will never return, but that its return depends mainly on its price. A profitable supersonic airliner, built on cheaper engines and quieter shapes, would be consistent with that. What would count against it is the opposite case. If an overland rule is finalised, an aircraft like Overture is certified, its cost per seat falls within what premium passengers already pay, and airlines still decline to put it into scheduled service, that would be strong evidence that regulation’s residue, risk or inertia matters more than this Dispatch allows. Not a definitive refutation, since network costs and passenger preferences can hide inside any cost estimate, but a clear signal that the price explanation has been given too much weight. On present evidence, with no final rule, no certified aircraft and a launch date already doubted by a launch customer, that test will not be run before the end of this decade.
Progress is always a choice about what to optimise. In aviation the choice went to the cheaper seat rather than the shorter trip, and it was made by people who understood the trade. NASA had a wing for Mach 0.98 by 1971. A DC-8 had gone supersonic in 1961. A Concorde ticket was on sale for twenty-seven years. The industry looked at all of it and bought the 747, then the 787. If the price of speed falls far enough, the same industry may buy differently. That would not contradict the story. It would be the story, told once more.
- Context
- Sixty-eight years after the 707 entered service, the fastest airliners cruise at about Mach 0.85 and the newest single-aisle jets slightly slower, while the FAA, NASA and Boom try to reopen supersonic flight.
- Causes
- Transonic drag sets a steep fuel price for speed above about Mach 0.85; the oil shocks, the 1973 overland ban and the economics of the cost index made airlines spend every efficiency gain on cheaper seats.
- Actors
- Airlines and their fleet planners; Boeing, Airbus and the engine makers; NASA; the FAA; British Airways and Air France with Concorde; Boom Supersonic and its customers.
- Perspectives
- Read as a price rather than a wall: physics determined what faster flight would cost, the market determined whether it was worth paying.
- Relationships
- Links to the Domains on energy and infrastructure: fuel prices, airspace closures and AI power demand all reprice how fast people can move.
