Back in 1898 Hilaire Belloc could quip, regarding European explorers and Africans

Whatever happens, we have got
The Maxim gun, and they have not.

The machine gun was revolutionary (and not just in a Gatling gun sense) by massively improving the capability of defensive fire power against mass attacks. It killed off the cavalry (and infantry) charge and produced a significant change in tactics both defensively and offensively. It was in many ways the culmination of 19th century industrial progress.

A 2000 era Belloc might have said something similar regarding dones, perhaps

Whatever happens, we have got
Predator Drones, and they have not.

By 1920, by 1914 even, the Maxim gun and its friends and relatives were not limited to just the great powers wielding them primarily against ignorant natives. In much the same way a hundred years on the drone is no longer a weapon wielded by high tech superpowers and allies against ignorant tribesmen. In the last few years we’ve seen drones built by 2nd or 3rd level powers such as Iran or Turkey used by even lower level militaries in Africa (Ethiopia) and Arabia (Yemen). But IMHO the Turkish and Israeli drones used by Azerbaijan against Armenia in 2020 were the wake up call that drones have transformed warfare even against generally technologically competent opponents. Mind you the ability of IS and other Syrian groups to deploy homemade drones that could disable aircraft on the ground should have been equally instructive except that the Russians/Syrians with the airforce mostly won that war. This is not dissimilar to the way that Russia successfully deployed the machine gun against Japan in the Russo-Japan war but this was missed because Japan’s navy kicked the crap out of the Russians and rendered the defensive land victories by the Russian fortifications moot.

However unlike the 1900s, we haven’t had to wait a decade for the missed lesson of Syria to become obvious.

In the current Ukraine conflict, the Ukrainians have come up with a defensive doctrine using drones against a non-drone equipped aggressor that complements the Azerbiajani doctrine for using drones against a non-drone using defender. As far as I can tell, the Azerbaijanis didn’t do a huge amount of innovation compared to what the US and Israel have developed, but they deployed the drones against a technically competent and well-armed defender (Armenia). The Azerbijanis showed that the use of drones must radically change the tactics required for air defence by using sacrificial drones to allow other drones to identify the locations of radars and SAM batteries that can then be taken out by the next wave of drones. That’s important but, in a way, it’s just a small upgrade from prior doctrines involving manned aircraft.

The Ukrainians, on the other hand, seem to have not just come up with a counter to that – mobility (which of course also works against traditional air attacks) – but to have come up with inventive ways to use drones against attacking forces. This is new. It is also likely to be as revolutionary as the machine gun once the use of that was figured out. As we all have seen, the key Ukrainian tactic has been to go after the Russian logistics support. They have not of course just used drones for this, but drones have proven to be very, very valuable in this kind of a war where you have special forces teams and resistance cells behind enemy lines blowing things up so as to starve the front lines of food, ammunition and fuel.

Drones are key for this because they are fairly hard to detect, let alone shoot down, but yet can loiter for hours providing remote intelligence or waiting in semi-ambush for a suitable target opportunity. They are of course also cheap, relatively speaking, as well as being a lot safer for the operator compared to a helicopter pilot or similar. If a drone operator makes a mistake and sends his drone into a power line or tree then that’s a loss of several hundreds of thousands (low millions at the high end) of dollars of hardware, but that’s it. The operator is still alive and the several hundreds of thousands of dollars is still a small fraction (certainly under 10%, possibly under 1%) of the cost of a new attack helicopter – not to mention the cost and time of training the replacement pilot, gunner etc. etc.

The cost means that you can deploy dozens of drones for the price of a single piece of manned equipment and that means you can do things like swarm a supply convoy or defended chokepoint (e.g. a bridge) so that even if a few drones are shot down others get through and hit the target. You can also equip every SF team with a drone or two so that they all have the ability to recon enemy forces from a safe distance and so on. One of the major implications for this, as the Russians are discovering, is that logistics convoys require a lot of defence. Now it is certainly true that the Russians have made this easier than one would have expected. Lack of maintenance as well as a failure to have an explicit second wave of invaders to protect supply lines have meant that the actual Russian logistics convoys are a lot more vulnerable than, say, US ones were in Iraq or Afghanistan but I suspect the doctrine would have had some success even there. Admittedly the US had their own reconnaissance drones and clear air superiority, neither of which the Russians really have in Ukraine right now (though they seem to be deploying drones a bit and this Orlan drone that Ukrainians mock is close to the sort of drone I expect to be commonplace in the fairly near future).

Cheap And Plentiful

As I mentioned above drones are (relatively) cheap for military equipment. The Turkish Bayraktar TB2 costs US$1-2 million, according to Forbes (but I’ve similar pricing reported elsewhere). US Reaper drones are up to 10x that, but the single use switchblade 300 kamikazi drones are ~US$70,000 and the anti-armor switchblade 600 somewhere between $100k and $300k. A TB2 can fly for over a day, as can those really expensive US drones and in that time they can fly a thousand miles or more. But that could be overkill for a lot of scenarios. Even (see below) a switchblade 600 is arguably more capable than required.

The basic use case for drones against logistics or static defenses only needs something like the switchblade 600 (perhaps a slightly improved one that can be recovered and reused after launching a missile like an NLAW or Javelin ) that just flies for perhaps 10-20 miles total and spends no more than an hour in the air before being retrieved. That is something that could be built today using commercial agricultural drones (such as the Yamaha helicopter one that sprayed the fields behind my house a week or two ago) that cost under well $100,000. But even that’s possibly high end. In Ukraine a number of drone hobbyists have been recruited to create very cheap drones from hobbyist ones.

On a cluttered table the x-shaped frame of one drone stands among bundles of plastic propellers and bags of minuscule screws.

Soon it will take flight with its payload — a wine bottle-sized anti-tank grenade designed to plunge on Russian vehicles.

Two other drones are already affixed with quad propellers, their squat bodies gaping with miniature bomb bays to rain explosives on Russian infantry challenging Ukrainian defenders to the north and east.

One more — the shape of a stealth bomber, the size of a bird of prey — will conduct reconnaissance missions for artillery squads, spotting targets and marking them for incoming fire.

These drones are really cheap (“An iPhone costs more than this equipment”), and are probably more like homegrown switchblade 300 drones, but they are undoubtedly effective. I suspect that a drone made from a toy radio controlled aircraft and a budget smartphone could meet most of the mission parameters at a price comfortably under US$500 (plane kit $200, phone $100, custom control circuitry maybe $50). You might need to spend a bit more to get the lift capacity for a decent size bomb/rocket – an NLAW weighs ~12kg/25lbs and a Javelin plus launch tube weighs about double that – but common standard hand grenades weigh about 0.5kg/1lb and I can certainly see numerous use cases for a small drone that can drop/launch hand grenades and, for reconnaissance you don’t need anything other than a camera with a telephoto lens. Quite possibly you could use the camera on the cheap smartphone with a suitable external telephoto lens (price $15-$50 depending). On twitter I see this Ukrainian drone which appears to be a standard agricultural spraying drone adapted to drop a couple of anti-tank grenade or similar and elsewhere I’ve seen estimates that the hardware cost of it is in the US$10,000 or less range.

Image

But even that is potentially expensive. Presuming you can get the motors, control circuitry, smartphone and have access to 3d printers and/or carpentry supplies you could build your own drone for even less. With a little care on the design front you could even dump the smartphone and replace it with, say, a Raspberry Pi zero ($5 for the board, add $5-10 for USB connectors etc.). Now you are probably looking at less than $100 in materials for a small drone that is custom designed for your precise use case. A benefit of a drone made out of plastic or wood and paper is that it would have a tiny radar signature and likely an equally small infra-red one. If you gave it an NLAW that would be the largest part of the signature, and presuming you had the target-finding part sorted out it could launch the NLAW from a quarter to half a mile away (say 400-800m). A squad of infantry armed with that could take out entire columns of tanks or artillery while the squad itself is safely dug in under cover a couple of miles away. I’m struggling to see how a tank, with its limited visibility, would even spot the drone approaching and with the engines running you probably wouldn’t be able to hear it either. In my experience of agricultural drones and RC planes, if you can hear them (which you can if you are within half a mile or so typically), you will find it very very hard to spot them unless they deliberately fly relatively high, and even then it is easy to mistake them for crows, eagles or other similar relatively large birds. It is quite possible that drones will end the century of tanks, and the half millennium of artillery because they completely change the calculation of cost of attacker versus defender.

If we assume the drone can accurately launch an NLAW or similar and that “An iPhone costs more than this equipment” for the drone itself then we’re looking at $30,000 or so for a system that can take out multi-million dollar tank of artillery piece (even a Russian T-72 starts at ~US$1 million, modern Western military tanks need another zero on the end, artillery is similar) and its 3-4 person crew at next to zero risk for the attackers who are several miles away. The US Switchblade 600, while more expensive (US$100k-$300k each including javelin warhead), is pretty much the top of the line version of this concept with an expected use range of ~20 miles.

It is true that in Ukraine the Russians seem to have forgotten the point of combined infantry and tanks where the infantry act to spot the attackers with their ATGMs and keep them out of range of the heavy metal, but drones make that a whole lot harder. Traditionally that range has been of the order of a quarter mile or so. The NLAW and the Javelin have changed the calculus by doubling (NLAW) or quadrupling (Javelin) that range, but even that means that launchers can be spotted and attacked so that the anti-tank attacker will stand up, acquire the target, launch the missile and run like hell.

With a drone it doesn’t matter if the missile launch is spotted because the drone isn’t the person sending the shoot command. In theory dismounted infantry could shoot the drone down, but that requires the infantry to be able to shoot a low-flying relatively fast moving object. Hunters experienced at bird shooting may be able to do that, but that and clay pigeon shooting (and relatives) is not something that everyone can do even if they train for it – and most military weapons training involves shooting at static targets. Exactly the same calculus applies to logistics convoys and fuel or ammo dumps. A couple of drones can destroy enormous amounts of supplies or disable – by blocking the front and back of the convoy – entire convoys.

Fundamentally any one thing within a 20 mile radius of the operator can be destroyed by a Switchblade 600. Presuming the operators have more than one and they can be used wisely then a small team armed with such drones can destroy or degrade entire tank or artillery brigades. Almost certainly someone can deliver 80-90% of a switchblade 600’s capabilities for a tenth the price, call them “switchblade 60″s or “Orlan 5″s (since they could also be cut down offensive versions of the Orlan 10). A 20 person team armed with say 50 of these “switchblade 60″s, which has an equipment cost of ~US1 million for all 50, would likely be able to engage an entire artillery battery, plus its armored defenders and its munitions/fuel resupply convoy, and expect to disable and/or destroy most of them. That’s 50-100 times the equipment cost plus at least 100 trained personnel (and quite possibly more). No military can afford such a lopsided fight on a sustained basis – even if they are able to neutralize some or all of the attackers. The only way the armor side wins is if the drone side doesn’t have enough drones to do the job. We may be seeing a version of this in Ukraine today where it looks like some Ukrainian assaults have been limited simply because they don’t have the supplies to take out the entire attacking force but that’s not something that any army, not even the well-funded US one, can afford to rely on.

New Doctrines

The doctrine mentioned just above where a small team can take out dozens of armored attackers using relatively short range drones is just the start. It is likely effective, but based on the cost and availability of the drones currently deployed (and the cost/availability of the munitions the drones deliver on target) it isn’t anywhere close to where we will end up.

Where I think we will get to in the next few years (by 2030 for sure, probably earlier) is drones that cost between $100 and $5000 (i.e. prices between the basic infantry firearm/mortar and a sniper rifle or machine gun) that can accurately deliver munitions that cost about as much as a mortar round (they could be an adapted mortar round, or a few .50 sniper rounds or a few grenades or a basic anti-armor rocket or something else small but deadly) over a range of 20-30 miles with a total flight time of a few hours. These drones will likely be retrievable and reusable but the expectation is that in any mass usage a significant percentage (up to 50%) will not be recovered. In fact this article suggests that Ukraine has already deployed early versions of these drones to some effect and there are videos posted on twitter that show them in action – I think the Ukrainian drones are shorter ranged but they are able to deliver an anti-tank grenade and fly back for a reload so the basic concept is there. However the cost and ease of manufacturing and distribution mean they can be treated as disposable as the average infantry weapon – i.e. please don’t break/lose the damn thing but usually there’s replacements you can get in a day or so if you do. The reason why I specify that the drone be re-usable is that for operations behind enemy lines, where resupply is limited, it will be valuable to have a weapons platform that can be reused and not be a single shot throwaway. In fact (see below) one of the key doctrinal changes is going to be that logistics become ever more key in drone warfare.

There will also be more expensive drones that cost 1000 times that (i.e. $100,000-$5million) that can fly for days and deliver far more than the current 4-8 missiles and so on. The missiles these drones deliver will tend to be smarter and larger than those of the little drones. And I imagine there will be intermediately priced drones (US$20,000-50,000) that are slightly more capable than the cheap drones but not as good as the high end ones. Note that the my fictional switchblade 60 fits into this category as does (at the top end bordering on the expensive drones) the existing switchblade 600. Since a large chunk of the switchblade 600 seems to be the javelin missile (cost ~$200,000), swapping that out for the cheaper NLAW (~$30,000) gets you precisely this drone capability, albeit a single use version. In a few years someone will come up with a reusable version. There will also likely be recon drones, dual use drones that can do both recon and attack and eventually defensive drones optimized to kill attacking ones. I expect people will come up with drones that can fly in at night, rest on building roofs (or utility poles, cellphone towers, disused parking areas…) until a spotter drone finds something for them to attack and then take off and attack. No doubt there will be  other drone types I haven’t considered in depth, including entirely terrestrial (drone tanks) or aquatic ones (a.k.a. smart mines or torpedoes).

One of the keys to deploying drones will be drone-drone intercommunication and flocks of drones. According to wikipedia the Orlan 10

is usually used in groups of two or three, in which the first is used for reconnaissance at a height of 1,000 to 1,500 metres (3,300 to 4,900 ft), the second for electronic warfare and the third as a transponder which transmits information to the control center. One system can include up to five vehicles.

Groups of 5 drones is in fact on the small side. If you look on youtube or other social media you’ll see numerous videos of drone flocks with dozens of members using lights to make pretty patterns and pictures in the sky. This one at Elon Musk’s Texas gigafactory opening is a good example. The same control techniques can be used to create drone swarms that would be extremely difficult for existing anti-aircraft fire to destroy, or at least to destroy cost-effectively. That cost-effectiveness thing is going to be key to drone warfare because eventually, not even the US has unlimited budget. And, even if the budget is unlimited, the supply-chains for more are not.

Ad Hoc Mesh Networks and Flocks

The combination of flocks and relatively low cost means that it becomes possible to swamp conventional defenders with attacking drones. Imagine a wave of say 50 drones attacking your defended point. Well you can use radars to spot and then anti-aircraft guns to take out, but in doing so you have identified the locations of both. The surviving drones may attempt to kamikaze on those attackers or a second wave of drones a little way back armed with more powerful missiles can then launch those missiles on your radars and AA guns as the Azerbaijanis did. The Israelis have a loiter munition/drone that can detect a search radar and autonomously attack it. One could easily imagine a flock of drones that locate radar emitters, infra-red hot spots or other emissions of interest (such as encrypted radio in certain frequencies)  and both report the information back but also decide to spontaneously attack targets that meet the criteria for a threat.

Drone flocks are great for defense too, even against other drones. Think of how small birds mob crows and raptors and apply that to drones. A flock of drones may be the best way to spot and then intercept attacking drones as well as other relatively slow moving things, both in the air and on the surface. Of course drones won’t work against high speed aircraft or missiles, but drones could well be used to protect the SAM batteries and radars from attacks or detection by other drones. In the Azerbaijan/Armenia case if Armenia had had the right drones they could have used those drones to kill Azerbaijan’s sacrificial drones without needing to unveil the SAMs and Radars. Then other drones could have been used to fly back into Azerbaijani territory to spot the drone operators and attack them. A flock of attack drones that communicate with each other can, as I noted above, react to some of its members being shot down by identifying the source of the weapons used and then either attacking or passing the data back for others to use. This would have neutralized the Azerbaijani drone advantage and then, probably, would have allowed the Armenians to defeat the de-droned conventional invaders.

The technology to allow drone flocks to communicate with each other and to pass information to and from the controllers is pretty much already developed. I suspect there would need to be some adjustments to off the shelf mesh networking protocols to handle things like radio jamming and obviously you’d need to layer encryption on top and so on, but none of this is new technology and it is the sort of thing that would be quite quick to develop using already existing standard open-source code libraries. Indeed it may have already been developed by some enterprising drone manufacturers/operators.

A flock of drones running continuous defense/recon would need protocols to handle parts of the flock rotating back to refuel/rearm and obviously you need the fuel and ammunition for that. But it is worth comparing the fuel/ammo usage of a drone flock compared to a tank battalion. Tanks, even on good roads, consume diesel on the order of a gallon per mile and gallons per hour even when static. When shooting artillery and tanks can burn through dozens of rounds quite quickly without necessarily destroying their targets. A drone like the Russian Orlan 10 or those Ukrainian agricultural drones have fuel tanks of a gallon or less and can expect to remain airborne/active for hours based on that. A Javelin or NLAW or equivalent weighs 10-30lbs (5-15 kg) and, in theory at least, can be precisely targeted so that only one needs to be used to take out an enemy tank or truck. A single truck with say 300 gallons of fuel in jerrycans plus a few hundred count of munitions would be entirely sufficient to support a 20-30 drone flock for a week or so. This is under 10 tons of total fuel and munitions so it would fit in a medium truck or couple of vans or similar. The same would keep a tank or artillery battery going for less than a day.

And I’m not even going to get into comparing drones with aviation assets. Helicopters and jets cost thousands of dollars an hour to operate, use comparatively large amounts of fuel and typically have to be based a considerable distance from the front lines because if the enemy attacks your home airport it can ground entire squadrons because of a lack of fuel or take-off runway or …. That means that they often have to fly tens of minutes, perhaps an hour or more to reach their operational area (as well as the same to return) and that in itself is a cost as well as a tactical liability because they have less fuel available to loiter and they can be attacked on the way. The sorts of drones I’m talking about which are made of sheet metal and lawnmower engines at the top end can take off from maybe 50 ft of flat road – or which can be launched like a kite or similar at the lowest end – don’t have any of these problems.

Cheap drones using cheap fairly dumb munitions are an economic game changer. Would you like 10,000 drones or a couple of tanks?

Defensively you can use 10,000 drones to cause total havoc on an invader. What if you launched 10,000 drones at an invader’s supply routes and they carried a mortar shell/incendiary grenade each, that might cost $10-20million in drones, munitions and fuel. Fly them at night, assume 1-2 hours flight to destination(s) – 50-100 miles – drop grenades and mortar shells on IR/radio hotspots and predefined GPS coordinates (bridges, identified bases etc.). Assume you get most of the drones back. The next night you repeat with ~8,500 drones. The next with ~7,000. The next with ~5,500. Your cost for this exercise is about half your drones, a couple of tanker loads of fuel and ~31,000 munitions. Your results are likely to be several thousand enemy trucks destroyed, lots of fuel, munitions etc. blown up and, due to the damage, roads/railways blocked so that simple repairs are hard, let alone additional resupply.

If you attack with 10,000 drones in the same way you can swamp defenses and destroy all sorts of things. Radars, fuel dumps, airfields… picture the Russian naval yards at Sevastopol, and consider what 10,000 incendiaries would do to it. Or even, assume the defenders get lucky, and only 5,000 make it. 5,000 phosphorous or napalm fires would wreck the place even assuming none of them hit something really exciting like the fuel or ammunition stores. You might not sink the ships but many of them would likely have damage that would make them unwilling to leave port. If you hit the repair slips then the ships probably can’t be repaired until the repair slips are repaired. If you hit power grid/transformers etc. they don’t have electricity to run all the things from computers to lights. And so on.

Or you could just go for softer targets and destroy the power and water for the entire city on night one and then on night two go for the places that still had the lights on. Lather rinse repeat. Odessa to Sevastopol is ~300km/200miles as the drone flies so if you wanted to recover the drones you’d need them to have flight endurance of 6-8 hours, but it would be a nightmare to defend against, especially if the first day’s drones included a few that took out radars, and even more especially if you had decoy threats that attracted attention away from the drone swarm.

Mind you, you’ll need considerable assets and personnel to deploy a 10,000 drone flock. Figure perhaps a hundred. standard shipping containers for the drones, plus fuel, munitions plus a few hundred people to prepare and launch them. But I imagine that it should be possible – indeed recommended – to break the fleet down into small groups of say 50-200 drones plus fuel munitions and up to ten people to prepare them for launch

Assassins and Sniper Drones

Another application for drones is to act like a remote higher altitude sniper. Imagine an adapted agricultural octorotor like the one in the photo above only with telephoto camera and a Lapua or Barrett sniper rifle attached securely below it and a separate loitering spotter drone. The spotter (and operator) could identify senior officers, critical resources and so on and then the sniper could be guided to a position a mile away where it can pop up above the skyline, hover and acquire the target and shoot. Not every shot would be a kill but it would be very disruptive even if, say, only one in three were completely on target. The difference between this and a switchblade 300 is the re-usability and price. Sure a flock of switchblade 300s could do a similar job but such drones can only be used once and generally cannot be recovered and reused if they are unable to find a suitable target. This sniper drone on the other had can be recalled, refueled and rearmed multiple times and the cost of a sniper round at around $10 is far far less than the cost of a switchblade 300.  Moreover, the sniper drone could carry a significant number of bullets and, possibly, have a choice of munitions (armor piercing, regular, tracer…). This would allow such a drone to take multiple shots at a variety of targets before needing to be reloaded. In many ways the drone doesn’t have to be that accurate to effectively degrade enemy activity. Simply forcing every enemy convoy and camp everywhere to search for such drones all the time will drastically reduce the resources available for attacking. This is similar to what insurgents did against US forces in Iraq and Afghanistan with IEDs only more so. More over it is far harder to counter because there is no bomb in the road and local trigger. The drone is a up to mile away and the drone operator even further away. The only real defense is going to continuous lowish altitude observation drones that can spot the sniper drone(s) moving into position and/or spot if not intercept the enemy recon drone.

Applications for this concept spread way beyond the traditional battlefield – assuming you sort out the communication and refueling requirements. a sniper drone could potentially fly several hundred miles to an enemy city, take up a position on a roof overlooking a square where political leader was due to make a speech or review troops etc. and pop up and shoot him midway through. Or it could overlook the gate where ministerial limousines enter or exit the presidential palace and shoot a couple of them. Or it could shoot the signalling boxes on rail tracks, the radars or fuel tanks on airfields or cause any number of other accidents. Or it could be used for simple terror by randomly shooting at trains, buses and cars in the rush hour, or at the spectators of sports events or religious ceremonies or…

If (when) a country like Iran figures enough of this out, terrorism is going to get nasty. As it is Iran’s various clients have deployed drones that completely alter the economics of missile defense. Saudi Arabia has used multi-million dollar patriot missiles to take out sub $100,000 drones – and it has needed to because if it doesn’t take them out, the drones will destroy critical parts of Saudi Arabia’s oil industry which costs even more millions of dollars to repair and so on. A country like Saudi Arabia can, more or less, afford this. If (when) Saudi Arabia or Israel decides to respond in kind Iran cannot. I imagine that if (when) Israel feels that Iran is about ready to deploy a nuke a large number of similar drones are going to fly into Iranian airspace and cause utter chaos.

The same goes for bomb dropping drones. A small drone is basically a slow, potentially reusable, cruise missile. Presuming you program its flight path, it can fly a very long way, drop a bomb on something and return to base for another one. How visible to air-radars is a drone flying less than 100 ft above the waves/trees? If you don’t have telemetry you need to spot the radar signature (small) or the IR signature (not much different to a bird) and be sure that you are identifying a drone and not, say, a bird or fishing boat mast. But such a drone could burn down a building, damage a bridge, blow up an electrical transformer or scatter cluster bomblets across a parade ground anywhere. For nations larger than Monaco, the defensive requirements needed to detect, let alone neutralize, drones coming in from any direction are next to impossible.

Cheap drone use is already spreading – this tweet shows Myanmar resistance groups using the basic agricultural drone to drop bombs and I have no doubt they are just one of many financially stretched groups of fighters doing this.

Drone on Drone Warfare

Part of the problem drones cause those they are attacking is that it is relatively hard to shoot them down from the ground and extremely hard to shoot them down cost-effectively without causing collateral damage. The problem is mostly that drones tend to be flown at low altitudes which means that there isn’t a lot of time for an individual surface based defense system to spot them and be able to attack them even though they are comparatively slow moving. Drones also tend to be small and relatively poor radar reflectors so the standard aircraft and missile defense systems are going to struggle to shoot them down and, as noted above, when they do they use munitions that cost an order of magnitude more than the drone. If, as I suspect, flocks of drones become common place the ability of such SAM systems to successfully defend their targets is going to be pushed to the limit and probably beyond it. I’m not sure exactly how many missiles a standard patriot battery has but I think it’s in the range of a dozen before it needs to be reloaded. If 50 drones were sent sufficiently wide apart that one missile could only take out a drone but sufficiently close together than there’s no time to reload, then a significant number will manage to survive to attack the final target(s). I suspect that Iranian proxies will test the anti-drone defenses of Israel, Saudi Arabia and other Sunni countries using swarms of drones fairly soon, if they haven’t already done so. So a cost effective defense against drones and flocks of drones is going to be needed real soon now

As I mentioned above, the obvious solution is to develop defensive drone swarms that can detect incoming drones and destroy them. I’m not certain what the exact strategy will be but I expect it will include multiple observation drones to detect incoming drones and a number of killer drones/missiles that can be be launched to go after drones that have been spotted. Since airborne drones are relatively fragile, proximity kill weapons are ideal – basically get a fragmentation grenade to explode in the middle of the swarm and you can take out quite a few of them. Anti-drone defenses will be run in parallel with anti-aircraft/missile defenses and rich nations will have multiple layers of drone defenses because the obvious first counter to loitering recon drones is to develop attack drones that shoot them down. So we’ll need decoy drones, dummy recon drones, recon drones with their own defensive mini missiles and so on. The good news for humans is that drone on drone warfare will tend to not kill humans.

Drones and the Battlefield

As I hope I’ve made clear, the key point of drones is that they are cheap, slow, potentially reusable, cruise missiles or spotter aircraft. They can deliver a comparatively small punch accurately with little danger to the operator. Cleverly directed drones can utterly disrupt an enemy by taking out key targets and they are generally hard to defend against without a huge expenditure in defensive equipment. This huge expenditure is almost certainly not replicable everywhere so drone tactics will tend to involve identifying weak points and exploiting them in ways that leave the well defended parts untouched and short on resupply. Drones probably will not deliver the coup de grace, that will be down to more traditional infantry, tanks and artillery. But those forces will only be deployed once the drones have prepared the battlefield and the numbers of them required are likely to be a lot less because the drones will have done much of the hard work.

Drones are going to massively extend the battlefield. Since drones work best against relatively lightly defended targets and they are cheap, the ideal drone target is going to be an ammo dump, fuel dump or critical bridge/junction well away from the perceived front lines. The next most ideal target are convoys and after that large fairly static bits of military hardware: radars, missile launchers, artillery and the like. Observation drones will be like the blimps and early aircraft of the first world war only smaller, harder to detect or hit, and able to give more precise observations. This will allow very long distance artillery, air and missile strikes to hit targets accurately without the need for a human doing the spotting. It is likely that until effective anti-spotter drone tactics are developed that can take out loitering drones several miles away, all battlefield defense and attack will have to assume that the enemy has accurate visual intelligence of your maneuvers and has had that since well before you moved into engagement range.

The military doctrine for combat defense is the survivability onion. Drones radically change the equation for being identified and acquired. A number of loitering observation drones say 1500′ above the ground and several miles away will be able to spot maneuvers and yet be hard to acquire as targets. During cloudless times high altitude drones will be able to monitor a very large area and also be hard to detect, let alone shoot down. Yes a SAM can reach up to 40,000′ (see the Malaysian airline shot down a few years ago), but detecting it well enough to launch a SAM is going to need the sorts of radar that a radar detecting kamikaze done will love to attack and there seems to be no reason why a high altitude drone would not be designed to be a poor radar target (as opposed to a large Boeing airliner) and one that could both detect the missile attacking it and deploy some kind of counter measures like chaff or an anti-missile missile – as well as reporting the location the missile was launched from. A swarm of very high altitude drones (say c. 80,000 ft altitude) would be a great way to not just spot the radars and missile batteries but also cause said batteries to shoot themselves dry. Sure many (most) modern SAMs can reach out to the altitude, and can home in on even a small drone, but as discussed earlier, you’re using some of your limited supply of multi million dollar missiles to destroy $10,000 or so drones. If the attacker can make and deploy dozens of these drones every week and you can’t get dozens of multi million dollar missiles in the same timeframe then you cannot shoot them down because then you have no anti-air capability at all.

Moving to more serious levels of the onion. Drones make being engaged and hit much more likely. For softer targets they have enough power to also raise the likelihood of penetration and being affected by their own munitions. For armored targets a single drone may not be sufficient but an observation drone in conjunction with artillery or top attack anti-tank weapons, whether drones or not, radically increases the danger. Plus armor is, as Russia has discovered, vulnerable to running out of supplies. A tank with no fuel is of limited use. A tank or artillery piece with no munitions is basically a lump of steel. If drones can (as Ukraine has shown to a degree) block supply chains and interdict both air and ground resupply then armor can never move more than half a fuel tank away from its base. The point of tanks is mobility. Without that their use on the battlefield is distinctly limited.

I think battlefields are going to radically change. Since drones are relatively slow and unarmored, they are individually very soft targets if you have the right weapons to take them out. But if you don’t have (enough of) the right weapons they expose you to attacks that are devastating to you and yet you cannot retaliate against your attacker. You can shoot the drone down, but you can’t shoot the drone pilot or the launching crew. Moreover you have to protect not just the front line but also all the rear echelons. Either that or your front line needs to bring a couple of weeks of logistics with it along with the drones and anti-drone weapons to protect those logistics (and the front line equipment as well) from being attacked.

There’s a video posted on twitter of Russian artillery shelling Mariupol (allegedly) which has been the subject of some interesting analysis. One thing that jumped out to me was that if the Ukrainians had drones near that artillery then the (surviving) Russians would be in a world of hurt. Perhaps that battery had anti-air assets that were not shown, but as it was depicted it would be incredibly vulnerable to drones taking out the ammo dumps and reloads. In the future such batteries will have to be either very well dug in with protection from overhead strikes or they will have to have anti-drone defenses that can keep drones far enough away that they cannot launch missiles (or kamikazi themselves). And those defenses are going to have to run all the time and be able to detect and react to drones in all directions, altitudes and distances. They are also likely going to need to be much more mobile when not under cover. Set up, fire some rounds, pack up and move on before a spotter drone locates you well enough to call in an attack or an attack will come that takes out your ready munitions and anyone near them. This Ukrainian artillery video shows that are already following this advice which makes sense since (see above) the Russians have been using a lot of reconnaissance drones

This is not just me

I had written about half of this post and then I found (via Samizdata) a youtube guy who put out a video a few weeks ago that makes very similar points. Well worth watching, and I think complementary to this piece in that he is very much talking about current drone performance and I’m looking a year or two into the future.

Although in the interests of fairness I should also link to this video which is a response to another video by the same person and which makes a fair case for the continued presence on the battlefield of modern main battle tanks.

I don’t disagree with a lot of that. As the presenter points out, tanks allow you to attack defended positions and to respond to (or create) opportunities at speeds that are effectively impossible to react to. In the 1-2 mile range that you can expect a modern tank to be able to accurately target something the time from firing to impact is on the order of a second. Missiles, bombs and (distant) artillery take considerably longer, and therefore there’s a far greater chance that the target will have moved away by the time the munition arrives. More to the point a tank is a mobile platform that can take that shot while moving and can thus evade the counter battery response. Good doctrine and good support (neither of which we have seen in most of the Ukraine conflict) makes tanks highly effective tools to attack an enemy.

I do question whether they need to be manned though. An unmanned drone tank could use the space for the crew for something else like more fuel, more munitions or improved defensive capabilities.

Plus tanks (whether drones or manned) need lots of logistical support and they grind to a halt without it. The real defense against an invader with MBTs appears to be to hit them in the supply lines, and (then) when they are static, hit them with top attack ATGMs that may also be drones/drone launched. Drones can infiltrate the front and then run wild attacking the soft logistics underbelly. There are of course counters to that, but most of the counter is adding defensive forces to the logistics convoys and depots and that comes at the cost of fewer resources available to push the front.