A Look Back at the Nagorno-Karabakh Drone War

Four years ago today, Azerbaijani and Armenian forces were locked in an apparent stalemate in southern Nagorno-Karabakh. Azerbaijan’s attempt to cut off the region’s major artery at Lachin had recently failed, and they were redirecting their efforts toward Shusha—where they won a decisive victory a week later.

Southern front of the Second Nagorno-Karabakh War. Lachin, the target of a failed Azerbaijani offensive, is circled in red. The front as of the end of October is marked in purple.

Most international attention was focused on the role of Azerbaijani drones. Turkish TB-2s and Israeli Harops were shown demolishing Armenian armor and air defenses in a stream of well-publicized videos. Within just days, speculation was rife that these new weapons spelled the end of the tank, and of conventional ground combat in general.

Needless to say, this was premature. A smallish war between two armies using older equipment, yet which also featured brand new weapons, did not seem like a good model for future combat. Reports of the ground fighting were few and far between, but painted a very different picture. When Shusha fell on 8 November, effectively ending the war, it was to a very conventional ground assault.

Israeli-made K1 overflies another drone strike, observed by a third. Screenshot from Azerbaijani Ministry of Defense video.

The excited commentary around drones is what prompted me to start The Bazaar of War as a place to discuss the evolution of arms and the economics shaping it. My first post, published on 3 October 2020, is reproduced below in full with a few additional remarks at the end.

---

DRONE APOCALYPSE: The End of Warfare As We Know It?!

The spate of impressive drone strikes over the past year has revived talk of the extinction of armored vehicles. Since late 2019, Turkish UAVs in Syria and Libya and now Azerbaijani drones in Nagorno-Karabakh have destroyed a sizable number of tanks, APCs, artillery pieces, and SAMs. Protective armor can only do so much and brings exponential costs with it; against that, swarms of drones are cheap. Is the age of armored military vehicles over?

Before we get carried away, it is worth remembering a few things. Each one of these conflicts is a propaganda war above all else, and it is not always clear what we are actually seeing. Many videos of drone strikes are clipped just after the explosion, making it impossible to assess the actual damage. Moreover, we have only seen attritive strikes on individual targets, not decisive attacks in support of a larger offensive—picking off stray vehicles is very different from collapsing a defensive line. Finally, armored vehicles have not yet in fact disappeared from any of the battlefields in question.

Advances in UAVs will gradually alter the dynamics of warfare, but it is ridiculous to suggest that they are on the cusp of completely upending things. There are three principal reasons for this: the nature of combined arms, the economics of drone warfare, and the operational difficulties of drone strikes.

The combined arms race

UAVs are no different in the essentials from any other air or missile platform. Anti-air systems have already destroyed a number of them in Syria, and there are plenty of existing ways to blunt their effectiveness: active protection systems, reactive armor, and better tactics, to name a few.

Recent strike footage, by contrast, shows 40-plus-year-old Soviet equipment in isolated, unentrenched, and unsupported positions. This has concentrated a lot of attention on the vulnerability of individual vehicles while ignoring how armies fight as a whole. Individual arms do not fight separately: it is unthinkable to send armor into urban environments without an infantry screen, or infantry into action without artillery support. Soon, it will be unthinkable to deploy armored vehicles unsupported by counter-UAS systems.

Pantsir-S1 firing

Modern militaries are only now beginning to field cUAS systems as part of doctrine. The Russian-built Pantsir gives us a good idea of what these will look like: a rapid-firing radar-guided cannons and missiles mounted on an armored chassis that can maneuver alongside other ground elements. Pantsirs have already successfully defended against drone swarms in Libya and Syria, adding yet another element to the thicket of modern combined arms.

Economics

It is interesting, then, that some of the highest-profile drone kills have been Pantsirs. In all cases the attacking UAV is thought to be the Israeli-produced Harop, a kamikaze drone employed by the Turks in Libya and the Israelis themselves in Syria. The Harop likely exploits specific gaps in the Pantsir’s capabilities: it is relatively fast, has a small radar cross-section, and flies directly into the top of its target (it is also possible that Assad’s and Haftar’s crews are too inexperienced). These gaps can and undoubtedly will be closed, and may not even exist with newer models—only older S1 models have been destroyed, not the S2 with an upgraded radar.

Which brings us to the race between the offense and defense. As weapons get better, both drones and cUAS systems will get more expensive. But costs rise faster in the air, as every upgrade brings new weight, balance, and power issues in addition to the cost of the improved avionics and munitions themselves. This undercuts the primary advantage of drones, which is that they are cheap and easy to manufacture.

As it is, the Harop is not especially cheap—about $10 million a piece, against $15 million for the Pantsir. As weapons evolve, militaries will face a tradeoff between low cost and higher kill probability. Their adversaries will meanwhile use anything they can to reduce the latter: decoys, radar spoofing, baited ambushes, complex defenses, etc. War is not a simple cost problem, but economics do affect inventories (more on this in a bit). Simply put, commanders will have to be discriminating in their use of drones, and will not be able to wipe out enemy ground forces at will.

What about at the lower end, where small, cheap drones can sneak past radars undetected? ISIS has used off-the-shelf drones armed with grenades for years, and new variants will get better while remaining low cost. This is an old problem, however: a short-range, small-yield explosive is no different in effect from mortar-fire. They will face many of the same difficulties of employment and limited capabilities, meaning they will hardly be groundbreaking.

The employment of UAVs

Another point of note about recent footage is that we only see the final part of the strike. We do not see the rest of the kill chain: the identification of the target, the communication of targeting data, the UAV moving to engagement range. We also do not see the enabling infrastructure: the airfields, the technicians, or the logistics of getting them in theater.

From a supply chain point of view, commanders will have to consider the number of UAVs they have on hand and how long it will take to be resupplied. The more advanced the weapon, the fewer are stockpiled—the era of precision-guided munitions raises the real possibility of running out of ammo in the opening phases of a conflict. This will force commanders to reserve drones for deliberate offensives, where they can achieve combined effects in conjunction with other forces. The old rules of warfare will still prevail, in other words.

Finally, there are major tactical vulnerabilities in the current crop of drones. Most of their recent success has come from their ability to loiter as they seek out targets or wait to be fed targeting data. The skies will not always be uncontested, however, and the longer that drones loiter the more easily they can be targeted by ground fires or counter-drone swarms.

Revolution vs. Incrementalism

In 1513, awed by the cannons which had recently brought down castle walls across Italy, Machiavelli argued that fortresses were obsolete. Just seven years later, in The Art of War, he detailed the kind of fortifications a prince should build—fortress design had caught up to the offensive power of cannons. So it always is in war: new weapons do indeed alter the battlefield, but the change is far more gradual than early encounters suggest. Drone maximalists tend to imagine buzzing drones filling the sky and vaporizing any ground forces, but don’t think through what that would entail. Just as artillery in the First World War and bombers in the Second failed to simply wipe out the enemy from afar, UAVs will take their place in modern armies as one piece of a complex, slowly-evolving system.

---

The fighting in Ukraine has largely borne out the three points outlined above, but adds some interesting nuance:

1. The Combined-Arms Fight

As soon as the front became settled, both sides took the obvious steps to protect their ground forces: digging in, keeping concentrations of armor back from the front line, and employing layered air defenses. Perversely, the biggest factor limiting the use of armor has been a very old technology: the extensive minefields which bog down assaults, allowing other weapon systems to go to work.

Drones have had their part in this, of course. But their direct impact lies more with top-end systems used against high-value targets like air defenses and infrastructure, and FPVs which harass infantry in the open. Arguably drones’ most important role has been indirect, providing the eyes to call in fires. But overall it has been remarkable how quickly different systems have been integrated at all levels in the combined-arms array.

2. Economics

The economics of UAVs still largely favor the offense, but not wholly. Russia and Ukraine are losing huge numbers of drones every month, and the odds of an FPV drone striking a target are well below even.

This will probably improve in the near future as AI terminal guidance gets around the problem of jamming, but soon after a lot of hard-kill measures will become available: ground-based guns, counter-drone swarms, directed energy, etc.

3. Employment

This overlaps substantially with the first two points. As far as getting UAVs airborne and operating them, I was mostly thinking of larger systems which require dedicated ground stations (group 3 and above). But Ukraine shows that limitations also apply to much smaller drones. Teams must move on foot just to get in range, then launch spotter and relay drones in support, while being careful not to crowd the radio spectrum with too many systems at once—taken together, this restricts their offensive power to an attritional dribble.

Which speaks to a final point. Ever since the frontlines have become settled, we have seen few attempts to create a rapid breakthrough, on the scale of weeks if not days—most offensives have been a grinding shoving match along a fairly wide front. This is partly because of the inherent difficulties imposed by modern weapons, but it is also a command-and-control problem. Neither Russia nor Ukraine has the staff capacity to manage the complex coordination of attacks by units at all levels. As such, drones are mostly used for attritional strikes against targets of opportunity.

But eventually, when someone attempts something more ambitious—whenever and wherever that is—it will be a very different question. Operations will lay bare the tradeoff between massed effects and steady attrition. Any commander who accomplishes the former will have brought drone warfare a whole new level.

---

Thank you for reading the Bazaar of War. Please like and share if you enjoyed, it helps more people find this. You can also support with a paid subscription, supporters receive the critical edition of the classic The Art of War in Italy: 1494-1529 and get access to exclusive pieces.

You can also support by purchasing Saladin the Strategist, in paperback or Kindle format.