An instructive moment came on April 3, when Iranian defense forces managed to shoot down an American F15E fighter over the Zagros Mountains, deep inside Iran. There followed an anxious couple of days as the Islamic Revolutionary Guards Corps and American combat search-and-rescue personnel raced to locate the two surviving aviators. The pilot was picked up within seven hours. The injured weapons-system officer, meanwhile, hid out in the mountains with little more than a pistol for protection. After what one senior military official described as “one of the most challenging and complex in the history of US special operations”, involving  four bombers, 64 fighters, 48 refueling tankers, 13 rescue aircraft and hundreds of ground personnel, the stricken officer was eventually picked up by a couple of second-hand Airbus propeller planes, executing a low-tech Plan B. The raiding force had got stuck on a soft runway.

President Trump could claim “WE GOT HIM!” — but like the war itself, the rescue came at an extraordinary cost. Numerous expensively tricked-out special-ops airplanes and helicopters were destroyed in the process, and America seriously hindered its own ability to make any such rescues in future. The escapade compounded what had already been an expensive few weeks for the US Air Force. Days earlier, the service had suffered its heaviest loss since the Vietnam War, when cheap plastic unmanned airplanes powered by commercial electronics raided Price Sultan Air Base in Saudi Arabia, damaging a number of KC-135 tanker planes and blowing a E-3G Sentry airborne radar plane in half. Another lesson from Iran is that the best fighters can be ambushed: hit by infrared-homing missiles at close range, launched from 4×4 vehicles that can quickly blend into civilian traffic.

Taken together, these incidents appear to confirm a narrative about 21st-century warfare that has built throughout the Russia-Ukraine war and is now taken as read by many armchair generals. The days of expensive, high-tech machinery like the F-15E (estimated cost: $100 million apiece) are numbered. The future belongs to unmanned aerial vehicles — or drones, as no one in the military likes to call them.

Suggested reading

Britain is not ready for war

By Aris Roussinos

The death of the fighter jet is not a new prediction. After another US pilot, Scott O’Grady, was recovered after his F-16C came down over a Bosnian mountainside in 1995, advocates for a new class of armed, unmanned air vehicles, or UAVs, quickly found a new slogan: “UAVs don’t have mothers.” It looks as if we’re going to have that discussion again.

Fighters are not only painfully expensive to buy; they cost tens of thousands of dollars per hour to fly and the air-to-air missiles that they carry cost $1 million to $2 million each. We must also consider the amount of time it takes the US Air Force to train people to use these machines: two years for a frontline pilot, five more years to raise them to a formation leader, and a decade or more to “command pilot”. Salaries and benefits have to compete with a healthy airline industry. Major live warfare exercises are not only expensive, but increasingly limited in scope. Air combat is reliant on electronic networks — but this makes it easy for an adversary to eavesdrop, making realistic live training even harder. These are painful costs. Designing and developing a new fighter, however, is more than painful. The US and many of its allies have bet the house on the Lockheed Martin F-35, a project that celebrates its 30th anniversary this November. However, troubles with new computers and software mean that F-35s delivered since September 2023 are still not combat-ready. Research and development for the fighter’s new electronics upgrade was most recently estimated at $16.5 billion — well more than half of the R&D tab for the all-new B-21 bomber, which learned lessons from the F-35 and other programs and has, by all accounts, been far better managed

And because fighters take decades to develop, the adversary will adapt and the world will change. The Eurofighter Typhoon was designed in the Cold War for supersonic missile jousts against Soviet fighters. It’s still being developed and modified, mostly to make it a better bomber — because until recently there was no high-end air threat. Britain, Italy, and Japan are meanwhile getting started on the Edgewing, a new stealth-fighter project under the Global Combat Aircraft Program, or GCAP. One of the challenges: none of the companies involved has started a new fighter project since the early 1990s, so most of the people who did it are long retired.

So it’s no wonder military planners and budgeteers are thinking about motherless unmanned aircraft again. Leading the charge in the United States are tech bros like the Hawaiian-shirt-wearing Palmer Luckey, founder of defense startup Anduril. Luckey talks about using Silicon Valley culture to build lethal new weapons “without bankrupting the United States in the process.” He promises to be able to move from the design stage to the front line in weeks, not decades.

But before you ask whether drones can replace fighters, you might want to ask what a drone is. The question is complicated by the fact that air-force people, by and large, hate the word “drone” — which actually applies to three very different things.

The first type of drone is a first-person video (FPV) drone, which borrows consumer technology to create a high-tech equivalent of an infantry mortar. Cheap enough to be used against individual soldiers, precise, lethal, but limited to the range of the control link and their batteries.

Then there are one-way attack (OWA) drones like Iran’s Shahed — which is not really a drone but a cheap cruise missile. Their airframes are molded plastic shells, and their petrol engines are little more sophisticated than a lawn-tractor motor — but what they do have is precise guidance. Israeli missile-defense pioneer Uzi Rubin said in the 2010s: “If you have an iPad, you have the parts for a guidance system.” Missile designers can get around local jamming of satellite navigation, since everyone can access terrain pictures of the entire world. OWA drones use these and cameras to find their targets.

The third class of drones, under active development in America, France, Sweden, and elsewhere, are much larger. They resemble small jet fighters, fly at more than 600 miles per hour, and carry the same weapons as today’s manned combat aircraft.

The three classes pose different threats to the fighter. In June 2025, over a hundred FPV drones were smuggled deep into Russia in ingeniously modified trucks, whereupon they emerged to destroy at least ten aircraft in five locations. Still, repeat successes seem unlikely now that the element of surprise has ceded.

Shahed-type OWA missiles are a threat to fighters because they can do much of their job cheaply and without putting pilots at risk. While there is a running argument over the unit cost of the Shahed and its Russian variants, most estimates suggest that it’s in the tens of thousands of dollars — in other words, about the cost of flying a fighter for an hour. These missiles are not stealthy, but their size makes them difficult to shoot down, and because they are cheap and programmable, dozens of them can “swarm” on a target, striking simultaneously from different directions. They also have a long enough range to reach air bases well behind the lines, so they can directly threaten fighter aircraft on the ground. And if you launch fighters to intercept them, you’re on the wrong side of a “cost-imposition” problem — it costs you more to shoot one of these things down than your adversary paid to launch it.

The “small jet fighter” class of drone has meanwhile been pitched as more of a like-for-like replacement for manned fighters, matching them in range and speed, other than short high-speed sprints. In theory, they should be cheaper over the long-term, too, because a fighter airplane spends most of its flying hours training pilots.

However, military aviators have diligently sought to keep them in their place: they have doubts as to whether they are ready, and fear that budget-cutters will push them prematurely to the front. The US Air Force has named them Collaborative Combat Aircraft (CCAs) to make it clear who’s the boss. Other anodyne names include “remote carriers” or “autonomous collaborative platforms”. The prototypes being tested in the US are about one-fifth the size of an F-35. But that’s an awkward compromise. They’re big enough to need runways, bases, and shelters — but too small to deploy over long distances, except a few at a time in the Air Force’s overburdened cargo fleet.

So right now, the momentum is with the second class of drone — the  “cheap cruise missile” group. Dozens of companies are springing up. Some are producing Shahed look-alikes. Others (such as the Netherlands’ Destinus, produced in collaboration with the German arms giant Rheinmetall) are developing their own jam-resistant navigation systems, molding their own airframe parts, and designing their own jet engines.

These weapons will undergo less rigorous testing than precision cruise missiles from established companies (like the American Tomahawk or the Anglo-French Storm Shadow) and will be built to commercial (i.e. less-than-military) standards. That might mean more failures — but at much less than a tenth of the cost. What makes them work is cheap, generic computing. They can navigate when GPS is jammed and identify their targets and pick aimpoints with simple daylight or infrared cameras — inverting the way their costly predecessors were made.

So, game over for fighters? Not so. The fighter is going to do what it’s done repeatedly over its 110-year history: adapt. Fighters from 1915 to the 1940s were pretty simple: one pilot, engine in front, fixed machine guns. Bigger and more complex aircraft emerged in WW2, some with radar and some, like the legendary Mosquito, big enough to carry bombs. The late 1950s brought supersonic speed and guided missiles, and the guns disappeared until Vietnam experience showed that both guns and maneuverability were still needed, resulting in the specs for today’s F-16 and F-15.

So fighters are versatile. The only target that has never been assigned to a fighter is a submarine. Teamed with a tanker and/or a cargo aircraft — itself a modified airliner, not exotic or expensive — the fighter can strike targets more than 1,000 miles away from its base, or deploy itself around the world while a warship is still pulling out of port.

Cheap cruise missiles, meanwhile, are not alternatives to the fighter but, rather, complements to the fighter’s flexibility and reach. Fighters have carried precision cruise missiles since the early 2000s.

Suggested reading

Military drones will upend the world

By James Kingston

As for the specter of large, reusable pilotless fighters, we should be skeptical of the US Air Force’s preference to keep them small and cheap and instead consider where Sweden seems to be headed. (Despite not fighting a war since 1814, Sweden has done an amazing job of keeping up with, and sometimes surpassing, much richer nations in combat air technology.) Saab’s advanced-projects team, dubbed Rainforest, is leaning toward a fighter-sized unmanned vehicle. Sweden is also the acknowledged leader in linking fighters into a network. A group of four Gripen fighters operate as a real-time network, compiling signals from four sets of electronic and infrared sensors, and from airborne radar surveillance aircraft, into common high-fidelity target tracks.

Consider how this might work: if an unmanned aircraft has half the lifetime cost of a Gripen (a reasonable goal), six aircraft could fly in formation, two with crew and four unmanned, with similar sensors, software and weapons, for equal cost. The formation spreads over tens of miles of sky. The uncrewed aircraft may be out in front at low altitude, not transmitting on their active radars, but prepared to launch weapons with a flight time in seconds. Last summer, Saab demonstrated an AI system to execute an air-to-air missile attack, calculating the attack path, missile launch and evasion; the same technology could support command pilots in a mixed manned and unmanned formation.

Having a crewed aircraft in the mix helps to deal with an ethical dilemma: autonomy. The idea of a weapon that searches for targets and uses its own computers to decide who lives and dies still makes people squeamish. I have engaged in simulations where a group of unmanned vehicles was operating with almost complete autonomy, detecting and identifying targets and assigning different tasks to different aircraft. Nevertheless, a high-resolution radar image of the target was transmitted to the human mission controllers so they could authorize the attack. In military parlance, this is known as having a human “on the loop” rather than “in the loop” — but it hinges on the assumption that data can reliably reach the human controller, in the face of jamming.

There is one way to make communications virtually impossible to jam, however: keep them to line-of-sight between two points. That’s where the manned fighter may remain indispensable:  operating within a cloud of unmanned adjuncts and missiles that simultaneously provides physical defense and attack, as well as creating, in military parlance, a robust line-of-sight network.

So that next-generation fighter might look very different from today’s aircraft, but the fighter has changed shape many times since the First World War. And that’s why it doesn’t surprise me, and rather encourages me, that the aircraft at the center of the new British-Italian-Japanese Edgewing project looks the way it does.

It’s a big beast, with a broad delta wing that recalls Britain’s famous Vulcan,  a 100-ton atom bomber,  more than a turn-and-burn fighter. It will hold what is technically known as a metric bleepton of fuel. It’s designed to be stealthy, but it’s going to be more than that — the modified Boeing 757 that’s being used to test its electronics is festooned with big antennas that won’t fit on a small unmanned aircraft, for detection, jamming and even generating beams powerful enough to reset the adversary’s electronics. The engines, for the first time in any aircraft, are being designed from the outset not just to provide thrust for flight, but power and cooling for electronics.

So GCAP is a fighter in the conventional sense — but it’s also a carrier for autonomous weapons, a command node, and a data center. A good time to remember that the most influential and versatile combat aircraft of WW2, the de Havilland Mosquito, was three times the size of a Spitfire and went after battleships.

There is a bigger picture, too. Those battleships took the lion’s share of defense spending up to the Second World War — but only one battleship put on contract after 1940 was ever completed, as an era of military technology based on steel and steam gave way to an age of jet engines, missiles, nuclear weapons, and electronics.  That was also a time when defense technology led all other developments and gave us mass air travel, satellites and computers.

We’re now in a time where commercial tech is in the lead, and wars are fought with information as well as explosive weapons. If aircraft like the Edgewing jet are going to replace the thousands of fighters in the world, they will be forced to evolve: and reluctance to accept change will not stop it happening. The Iran war will result in the demise of neither the IRGC, nor the jet fighter, nor the unceasing development of warfare.

Britain is no exception. Here, Labour MPs are agitating over their government’s plan to cut disability benefits — mere days after Kier Starmer, announced in Brussels that the UK will acquire 12 nuclear-capable fighter jets for the RAF at a cost of around £1bn. These jets are the Lockheed Martin F-35A Lightning II Joint Strike Fighter (JSF) aircraft, which can be armed with nuclear bombs. The purchase is incongruous for a party of the Left, and its announcement is of a piece with the existing Starmer narrative that the UK is under threat. The new jets, then, are the latest of a long list of populist signals that wrap Labour in the Union Jack. Starmer continues to paint himself as a politician with little to nothing in common with his party-leader predecessor.

No doubt there was also inter-service lobbying behind this announcement. The RAF has always lamented the decision to acquire another heavier, more expensive and less capable variant, the F-35B, to share with the Navy. The F-35B can take off and land vertically, rather than via a runway, thus making it an appropriate fit for the relatively small size of the UK’s two aircraft carriers. These aircraft were to be flexibly shared with the Navy, meaning that the RAF could lose precious squadrons on naval deployments for months at time. The purchase of the F-35A jets means that the RAF gets a dozen devoted aircraft as well as upgrades to its base upgrades at Marham. So the Airforce will be happy — up to a point.

The intriguing angle is the not-so-welcome nuclear mission these aircraft will be given. They have been bought with a single purpose in mind: to deliver American B61-12 free-fall nuclear bombs under Nato’s nuclear sharing arrangements — an understanding that already operates in Germany, the Netherlands, Italy and Belgium. Accordingly, the host state buys the American aircraft and provides a base, support, and a crew. A wider set of Nato states provide support operations, including essential refuelling and force protection under the so-called SNOCAT missions.

But the bombs remain under the control and protection of the United States until they are released to allied nations. This can happen only after an order, issued in wartime, from the American Nato Supreme Allied Commander, and such an order would presumably enable strikes against Russian targets. Given that the UK is a nuclear weapons power with its own, ostensibly independent, Trident deterrent, the logic of the procurement is unclear.

As well as being redundant in these terms, their deployment is a risk. Aircraft can be shot down far more easily than a submarine can be sunk. With this in mind, the MoD, when considering alternatives to Trident in 2013, rejected the far cheaper air-delivered option on the grounds that this method risks the planes being destroyed before they can deploy their bombs. Were the President to decide to use nuclear weapons, it seems far more likely that the United States would use its submarine-launched ballistic missiles, fired from the sea, or land-based intercontinental ballistic missiles, fired from a silo. Missiles are far quicker and more reliable. They do not require complicated international arrangements whose durability has rightly been questioned.

More from this author

Putin knows we are spreadsheet warriors

By Wolfgang Munchau

It is true that these arrangements are diplomatically fragile. They are regularly criticized for violating the ban on Non-Proliferation Treaty member states transferring or receiving nuclear weapons. Nato argues that these transfers would only occur in wartime, when — in Nato’s view — the Treaty would no longer apply. But this disagreement exacerbates the already festering resentment felt by rival or smaller countries that have surrendered any ambition to deploy nuclear weapons. To them, it smacks of discrimination and double-standards.

Nato nuclear-sharing, then, is essentially symbolic and political. It ensures that non-nuclear members “dip their hands in the blood”, as some reports put it, of nuclear deterrence, while expecting that the heavy lifting of nuclear deterrence will rest on the nuclear-armed strategic ballistic missiles of the United States, Britain and France.

So why is Starmer doing this — buying planes that amount to little more than a diplomatic signal — when this particular consideration clearly does not apply to the UK? Is this announcement really just another Starmer expression of flattery and subservience to Donald Trump?

Or could it be that MoD has finally lost confidence in its Trident system?

Technical problems over the renewal of Trident, through which the Vanguard class of submarines will be replaced by Dreadnoughts, have been widely predicted for many years. Dominic Cummings has warned darkly of “rotting infrastructure” draped in state secrecy. Now, in public, the Infrastructure and Projects Authority (IPA) has warned of severe problems with Rolls Royce’s production of the nuclear propulsion system, the PWR3. All other major elements of the renewed submarine system are under review, and present a cause of serious concern over delivery. These assessments are likely to be harbingers of further delays and costs. This is a major problem, because there were already concerns that there would be a break in the availability of the submarines to patrol in the transition period. The Royal Navy places great value on its claim to have achieved a “Continuous At Sea Deterrence” posture since the first Polaris submarines were launched on 15 June, 1968. Could the F-35A procurement be an expensive insurance policy to cover that gap?

But the problems may go even deeper. The submarine system itself suffers a number of vulnerabilities that will only worsen as technology advances. Because the vessels are large, complex and singular, and because their production took an extraordinary 25 years, much of their design is inflexible, and will be outdated by the time they enter into service. In contrast, more modular, networked anti-submarine warfare assets can be produced far more quickly. They benefit from continuous tech development and renewal, and units are far cheaper, more expendable and replaceable. In the well-established and unending game of cat-and-mouse, the submarine has significant weaknesses.

The rapid development of autonomous drones above and below the water, as well as communication systems, sensors, and fully networked interception capabilities, together mean that the new Dreadnought submarines could be redundant before their first patrol, scheduled to be around seven years from now (unless there are significant delays).

If an adversary has the ability to track and destroy a submarine on patrol, this would not only neutralize the submarine’s deterrent effect. It would also lead to a highly destabilizing dynamic at the worst possible time. Faced with the risk of imminent loss of Britain’s nuclear submarines in a crisis, would a prime minister be tempted to preemptively launch an all-out attack? It could feel like a use-it-or-lose-it situation. Within a few years, it could be the case that the very worst place for one’s nuclear arsenal would be in international waters.

We should pay close attention, then, to the little-reported corner of the Strategic Defense Review in which we are told that the Government should “define the requirement for the post-Dreadnought nuclear deterrent within this Parliament”, 30 years before the submarines’ planned retirement. 

Remember for a moment the UK’s international legal obligation as a member of the Non-Proliferation Treaty: “To pursue negotiations in good faith on effective measures relating to cessation of the nuclear arms race at an early date and to nuclear disarmament.” The failure of the nuclear weapon states to make progress on their disarmament commitment is the core reason why the nuclear non-proliferation regime is in a continuous crisis, and why states resist stronger non-proliferation arrangements. Signaling an inflexible commitment in this Parliament to new a nuclear weapon system planned for deployment between 2060 and 2085 hardly sits comfortably with that legal obligation.

In any case, committing to and nailing them down this far ahead is strategic and technological madness. Such an inflexible commitment to particular platforms and delivery systems is extraordinarily reckless at a time of rapid technological change and likely redundancy. 

More from this author

Britain is not ready for war

By Aris Roussinos

Unless, that is, MoD is anticipating an earlier requirement for a new strategic deterrent system. The new submarines’ planned lives run to at least the late 2050s, but if there is concern that they could be compromised well before then, this alone might justify a relatively modest additional capital investment of a couple of billion on this additional F-35 system.

But the F-35A free-fall bomb system currently proposed is no substitute for Trident, for it surrenders sovereign decisions on warhead control and nuclear release to the Americans. If the Trident submarines come to be judged a liability and the UK’s nuclear capability then came to depend upon the aircraft, the UK would need to buy more F-35As and either join the ranks of the other NATO non-nuclear weapon states or instruct Aldermaston, the British nuclear weapons facility, to develop the UK’s own free-fall nuclear weapon.

Not counting the cost of warhead modernization, the Government has so far spent in the region of £25 billion on the Trident renewal program. Whatever the challenges, it seems highly unlikely that officials would contemplate abandoning the project. But if there is a significant possibility that confidence in the system would be in doubt, we could be witnessing the start of an investment in a limited and unsatisfactory plan B. 

Kier Starmer may talk tough and make big announcements, but the underlying reality could be very different. His announcement of a sustained increase in the UK defense budget has captured the headlines. But even were these plans to survive contact with the enemy — that is, competing public spending priorities and the state of the economy — the top-line budget dimension may be less critical than the need to ensure that taxpayers’ money is spent well and that waste is tackled head-on.

Britain’s submarine-based nuclear deterrent system is at high risk of becoming another example of an expensive failed infrastructure project that is not fit for purpose, a legacy of a past strategic context, and an reflective of an obsession with symbolism over substance. As competitors race ahead with accelerated technology developments and the world becomes more volatile, we need to reassess our priorities.