In 2020, a war in the Caucasus, invisible to most Americans, killed some 6,000 people, wounded tens of thousands, and displaced many others. From September to November, an intense series of battles between invading Azerbaijani forces and Armenia, in the mountains of the long-disputed region of Nagorno-Karabakh, pulled in fighter jets from Turkey and missile defense systems from Russia. The fighting also veered dangerously close to a critical new oil pipeline between Russia and Europe. When it was over, several thousand Russian troops were keeping a complicated truce, and governments around the world started learning difficult lessons. This May, Israel and Hamas engaged in a widely scrutinized and intense 20-day campaign. Hamas launched thousands of missiles into Israel, subjecting the Iron Dome missile defense architecture to a severe stress test, while Israel used precision airstrikes, guided by, the Israel Defense Forces (IDF) says, significant advances in data science, imagery and geographical intelligence, and technical collection.
These short conflicts offer some provisional lessons about the future of combat, particularly in the use of air power—F-16 fighters on loan to Azerbaijan from Turkey, a new drone fleet against an Armenian military more constrained by gravity, and Israel’s claim to have used artificial intelligence to drive targeting.
Armenia’s reliance on Russian ballistic missiles, rocket launchers, defense systems, and intelligence, surveillance, and reconnaissance platforms is probably top of mind at NATO headquarters because of what it implies for Russian use in a conflict with Europe. Notably, although both sides invested heavily in information operations and there were some indications of cyberwarfare, the war was as old-fashioned as war gets these days—there is little public evidence (as of yet) that cyberattacks or information operations made any significant difference in determining the outcome.
Israel went old-school: Its biggest influence operation became attributable instantly. On Twitter, the IDF announced—or seemed to announce—that it was prepared to invade Gaza from the ground. Hamas military officers scrambled from their hideouts, moving to others, all under the watchful eyes and open ears of Israeli signals and reconnaissance platforms. Moving targets thus identified, Israel proceeded to strike relentlessly. Hamas’s system of underground tunnels was damaged, although it is too soon to evaluate how resilient the Hamas defense architecture is.
Since the last spasm along the disputed line of conflict in 2016, Armenia fortified its military presence considerably. But Azerbaijan’s shaping strategy significantly reduced the reach and scope of Armenian fire. Drone warfare appears to have played an outsized role by reducing Armenia’s quantitative advantage. Fairly inexpensive unmanned aerial vehicle swarms were sent in advance of attacks and drew anti-aircraft fire. Armenia did not field an advanced counterdrone system, which then allowed Azerbaijan to identify the location of hidden Armenian air defense artillery.
One analyst, using open-source imagery, estimates that as many as 175 pieces of military equipment, mostly Russian-built medium-range rocket launchers, were lost simply because their locations were identified by the drone swarms. Equally importantly, Azerbaijan used a variant of Turkey’s Bayraktar TB2 drone to detect the surface-to-air missile regiments that were spread throughout the mountainous regions, followed by Israeli-made munition drones to destroy them, effectively achieving the suppression of enemy air defenses objective early on. Never before in warfare has this dual-drone search-and-destroy gambit been used so successfully.
NATO won’t put too much emphasis on this lesson, as Armenia’s air defense strategy and its lack of a common air defense architecture made Azerbaijan’s task relatively painless. Achieving suppression of enemy air defenses in Russian territory would be much less decisive. But Taiwan, among other countries, will study the asymmetric drone warfare in the context of defending the island against a littoral invasion from China and, in particular, preventing China from establishing an literal beachhead. But China, too, no doubt will find a justification to enhance its investment in counterdrone applications, embedding them with its marine and littoral forces.
In the Nagorno-Karabakh war, the drones shaped the battlefield, but light and some mechanized infantry followed through. Azerbaijan, according to several analysts, effectively used small, special-forces-type detachments, quickly calling in airpower when necessary. That was a tactic borrowed from the early U.S. invasion of Afghanistan, when the success of a particular engagement depended on the ability of combat air controllers to call for air fires near-instantaneously. Azerbaijan made headway along the line of conflict using targeted raids, often at night. When Armenian air suppression techniques were too onerous, Azerbaijan’s forces would call in strikes from its Israeli-made Lora ballistic missiles, safely protected in Azerbaijani territory but accurate to within 120 feet of their target hundreds of miles away. Armenia, prepared for a war of attrition and extended conflict as in the 1990s, was not prepared for an integrated multidomain attack, and it had postured its defenses to ward off a massive invasion. Instead, Azerbaijan neutralized the size of the force, controlled the air, and moved surgically to recapture the territory.
Notably missing from all this was any real attempt at that perennial buzzword: cyberwarfare. In the middle of conflict, the most conspicuous attempt at cyber-exploitation was revealed in terms of an Armenian intrusion into the emails of Azerbaijani political officials. It had little, if any, effect on the fighting and served no deterrent purpose. (Azerbaijan retaliated by shutting down some Armenian websites.) More importantly, because Azerbaijan planned to initiate the conflict, it throttled access to the internet in areas near the conflict zone on both sides, depriving civilians of access to critical information while silencing any retributive or hastily drawn Armenian information operations.
The Armenian diaspora is highly active on social media and organized global protests against the invasion. But the effects of their activism were probably muted by a combination of the global pandemic and the U.S. election. Both sides used their official government Twitter account to accuse the other of war crimes or of targeting civilians. Once again, the lack of a global media focus on the conflict made it difficult to separate fact from claims, or even to draw attention. The BBC noted several misleading and exaggerated claims on both sides. More dramatic footage, usually of missile strikes or from drones, spread through conventional social networks, even TikTok.
And that might be the biggest lesson of all this: Timing matters when it comes to revanchist ambitions. A pandemic or other global crisis occupying people’s minds is probably a good time to launch an invasion. Russia failed to supply—or failed to obtain—intelligence about Azerbaijan’s embrace of hybrid warfare, even though much of this doctrinal development was not hidden. Controlling the narrative about a war when powerful geopolitical entities have maxed out their ability to focus can be very difficult, which is one reason why the Azerbaijanis were able to claim a victory without Armenia finding enough bandwidth to counter the claim.
In Israel, the political situation has grown even more complicated since the cease-fire, with no clear sense of a winner or a loser; it is hard to independently assess claims of casualties, damage, and completed objectives. That said, Israel’s sophisticated state media made it known that it had used data science to try to minimize casualties and maximize effects.
From an underground coordination center, algorithms ran through reams of all-source intelligence data collected on target locations, flagging unusual patterns of behavior to combat controllers, who then transferred those anomalies to fighter pilots. (Israel’s version of the U.S. National Security Agency, IDF Unit 8200, codenamed this program “Gospel.”) Israel also says that damage assessments and sophisticated real-time moving-target tracking of artillery launching sites were fed into Iron Dome, allowing Israel precious minutes to orient its system to intercept Hamas-fired rockets. Israel also says that it discovered, through intelligence, that Hamas was using an office tower to try to jam Iron Dome, without success. These claims, of course, will be refined and vetted as we learn more about the conflict. (Israel’s proactive policing, which relies on AI and big data, has drawn criticism for invading the privacy rights of innocent Palestinians.)
A few lessons seem clear, however. Israel’s vaunted intelligence apparatus did not anticipate the amount of ordnance that Hamas managed to smuggle in or manufacture in the seven years since their last conflict. Iron Dome, while not perfect, proved incredibly effective at protecting the Israeli population from Hamas artillery. A short-range missile defense system has proved itself.
