On July 18, reports emerged that Ukrainian long-range drones struck two Wildberries warehouses in Russia. Wildberries is Russia’s largest online retailer, roughly equivalent to Amazon. The attacks killed eight warehouse employees and injured many others. Given that both of the warehouses were located far from the front (one only 30 miles from Moscow), the strikes brought the war home to Russian citizens in a very real sense. So too did the fact that they were civilian facilities with no apparent connection to the conflict. It seemed plausible, at least based on the initial information available, that Ukraine had adopted Russia’s tactic of direct attacks on the civilian population in order to sow terror. If so, it had violated the law of armed conflict (LOAC) by attacking civilian infrastructure, and those involved may have committed war crimes.
But Ukrainian President Volodymyr Zelensky quickly countered that the Russian warehouses were important logistics facilities used to “supply sanctioned components for drone production and navigation equipment.” That assertion complicated legal assessments of the attacks. If true (and as explained below), the warehouses were so-called “dual-use objects,” and therefore lawful targets under LOAC.
As this war has progressed over the four years since Russia’s full-scale invasion of Ukraine in 2022, attacks on targets other than those that are clearly military in nature have become ever more commonplace. Russia appears to have abandoned any pretense of limiting its attacks to military targets. For its part, Ukraine has been targeting Russian oil facilities, ports, merchant vessels, and other ostensibly civilian targets in a strategy that appears designed to impose costs on Russia beyond the battlefield. U.S. government officials, at least in the last administration, were concerned, and rightly so, that it would no longer be only the Russian side attacking civilian objects. For instance, in 2024 testimony before the House Committee on Armed Services, the Assistant Secretary of Defense for International Security Affairs Celeste Wallander acknowledged that the United States had raised the issue of targeting critical infrastructure with Ukraine. In some circumstances, such attacks can be lawful. But Ukraine is pushing the envelope.
The trend is unfortunately not unique to the conflict between Russia and Ukraine. In the ongoing international armed conflicts between Iran, on one side, and the United States, Israel, Jordan, and Gulf Arab States on the other, infrastructure attacks occur regularly. For instance, since the collapse of the interim ceasefire and the resumption of hostilities, U.S. strikes have hit Iranian bridges, railways, energy sites, telecommunications networks, and port facilities. Iran has answered in kind, for example, by attacking power and water desalination plants in Kuwait, a country that draws roughly 90 percent of its drinking water from desalination. It is also striking petrochemical and petroleum facilities throughout the Gulf States. And, of course, Iranian strikes against merchant vessels transiting the Strait of Hormuz off Oman’s waters ignited this round of exchanges. Although at least some of these targets had military significance, each also served civilian functions.
One challenge in discussing these strikes in public discourse is that characterizations of their legality tend to be overbroad. On the one hand, some believe any strike on a facility that continues to serve the civilian population is unlawful. That is not the law. But on the other hand, some claim that any nexus to military action, no matter how attenuated, renders a civilian entity subject to attack. It does not.
Proper assessments of strikes on facilities used by civilians are highly contextual. Complicating matters is the fact that the boundaries between lawful and unlawful targets can be ambiguous. In this primer, I review how LOAC applies to the targeting of civilian infrastructure in an effort to dissuade those in both camps from jumping to flawed conclusions. It is a subject about which I have written before (see, e.g., here and here), including with my Just Security colleagues Tess Bridgeman and Ryan Goodman. But in light of the most recent hostilities in both theatres, and the growing tendency to strike such targets, it is a topic to which a return is merited.
The Baseline Rule
Article 48 of Additional Protocol I to the 1949 Geneva Conventions (AP I) sets forth the principle of distinction:
“the Parties to the conflict shall at all times distinguish between the civilian population and combatants and between civilian objects and military objectives and accordingly shall direct their operations only against military objectives.”
Although the treaty binds only its Parties, like Russia and Ukraine, there is universal agreement with the International Court of Justice’s conclusion that distinction is a “cardinal principle” of LOAC that binds all States (Nuclear Weapons Advisory Opinion, ¶ 78).
This principle has been operationalized in LOAC’s so-called “conduct of hostilities” rules, the most important for our purposes being Article 52(1) of AP I: “Civilian objects shall not be the object of attack or of reprisals. Civilian objects are all objects which are not military objectives.” The rule, including the definition of military objectives, reflects customary law in both international and non-international armed conflict (ICRC Customary IHL study, rules 7, 8, 9, and 10; DoD Law of War Manual, § 5.6.3).
Two points regarding the definition of civilian objects merit emphasis. First, civilian objects are defined negatively. There is no agreed-upon list of protected objects to consult, nor is there a threshold of civilian significance that triggers the rule. Thus, terms such as “critical,” “essential,” “only,” and the like do no legal work despite their prominence in public commentary. For purposes of the rule of distinction, a refinery, airport, rail line, port, warehouse, desalination plant, and so forth is either a targetable military objective or protected from attack under LOAC. The options are binary.
Second, the entity to be assessed is the individual target, not sectors, systems, industries, or networks. Ukraine may not, for instance, conclude that Russia’s energy “sector” or Iran’s transportation “network” is a military objective. The conclusion that one bridge qualifies as a military objective does not relieve an attacker of the obligation to conduct a military objective assessment for the next.
To be clear, and as discussed below, whether an object is military or civilian in character is only one of several assessments that must be made to determine whether an attack on that object is lawful. Rules like proportionality (which generally prohibits attacking even a military objective when the expected civilian harm is excessive compared to the military advantage anticipated to be gained), and the requirement to take all feasible precautions to spare the civilian population must also be considered in every attack. But unless the object at issue is a “military objective,” none of the other rules come into play – do not pass go, do not conduct further analysis, the attack is unlawful.
All of which means that the controversy over targeting what appears to be civilian infrastructure in both the Russia-Ukraine and U.S.-Iran conflicts often boils down to a single question: has the enemy done something with an indisputably civilian object to make it lawfully targetable, that is, a “military objective?”
My concern about “pushing the envelope” in this context is that the question sometimes becomes: where can flexibility in the rules be exploited to treat civilian infrastructure as a military objective even when a good faith assessment would likely come out the other way? In short, can we make that case based on ambiguous or speculative versions of the facts or stretch the definition of ‘military objective’ far enough to make it so?
The Military Objective Definition
Additional Protocol I defines military objectives in Article 52(2), text that is universally accepted as also setting forth the customary law definition (ICRC Customary IHL study, rule 8; DoD Law of War Manual, § 5.6.5).
In so far as objects are concerned, military objectives are limited to those objects which by their nature, location, purpose or use make an effective contribution to military action and whose total or partial destruction, capture or neutralization, in the circumstances ruling at the time, offers a definite military advantage.
As the International Law Association’s Study Group on the conduct of hostilities (of which I was part) concluded, there is “no reason to believe that States not party to API have greater latitude of discretion” than that set forth in Article 52 (p. 342).
Two features of the definition structure the analysis of whether a civilian object has become a targetable military objective. First, it sets out a two-pronged, cumulative test. An object is a military objective only if it both contributes to military action in an “effective” manner and attacking it will yield a “definite” military advantage. Both prongs must be satisfied; neither may be presumed from the other.
Second, the definition is temporal. Status is assessed “in the circumstances ruling at the time” of the attack, not by reference to what an object was or might someday be in the abstract. An object that is civilian today can become a military objective tomorrow, and one that is a military objective now may revert to civilian status once the circumstances that made it targetable no longer hold (see below).
The effective-contribution prong: The first prong requires that the contribution be “effective.” There has to be an actual nexus between the object and the enemy’s military action. The contribution may not be remote or speculative, although it need not be indispensable or direct. Importantly, an object that sustains or supports military operations qualifies even if there is no direct tie to combat. Second, the contribution must be to “military action” – the war-fighting and war-supporting activity of the enemy’s forces, including the logistics and materiel that enable it. Political or economic ends, whether domestic or international, do not suffice. The focus is on the military capability of each party to a conflict. (This distinction raises the so-called “war-sustaining” debate, to which I return below.)
The effective contribution requirement can be satisfied by an object’s nature, location, purpose, or use. An object makes an effective contribution by its nature when its inherent character supports military action (AP I Commentary, ¶ 2020; DoD Law of War Manual (§ 5.6.6.1). Paradigmatic examples include military aircraft, integrated air defense systems, and weapons systems used to attack vessels. The category is defined by what an object is, even when not in use. Accordingly, civilian infrastructure cannot qualify as a military object by nature – a warehouse, refinery, or power plant has no inherent military character.
This point is sometimes muddled. For instance, it has been suggested that oil installations and main transportation arteries qualify as military objectives by nature (AMW Manual, rule 23). However, the better view, reflected in the ILA Study Group’s conclusions, is that such facilities become military objectives, if at all, through use or purpose (more on each of those below). This is a critical distinction because qualification by nature is permanent – it is inherent in the object – while use and purpose are contingent and can lapse (pp. 330–31).
An object can also make an effective contribution when its location is of military importance (AP I Commentary, ¶ 2021). An example would be a bridge, the destruction of which would, because of its location, bar resupply of enemy forces. However, such a bridge could also qualify as a military objective by use or purpose.
That leaves use and purpose – the primary criteria through which civilian objects become military. “Use” refers to an object’s present function; “purpose” concerns its intended future use. An object makes an effective contribution by use when supporting military action now (AP I Commentary, ¶ 2023; DoD Law of War Manual § 5.6.6.1). Examples include a bridge over which the enemy’s military supplies are passing or a storage tank currently holding POL (petroleum, oil, and lubricants) for the enemy forces. President Zelensky invoked this as the justification for the Wildberries warehouse attacks, stating they were being used to supply components for drone production and navigation equipment. If accurate, that use makes an effective contribution to Russia’s military action, and the warehouses were military objectives when struck (subject to satisfaction of the second “definite military advantage” criterion discussed below). Whether the factual predicate was accurate is a separate question, one depending on the quality of the intelligence and the reasonableness of relying on it. But assuming it was accurate, the first prong was clearly satisfied.
Importantly, the extent of civilian use is legally irrelevant to whether an object qualifies as a military objective (though it may come into play in a proportionality or precautions analysis, discussed below). Once an object makes an effective contribution to military action, the first prong is satisfied; that it also performs substantial civilian functions has no bearing on its qualification as a military objective for distinction purposes. A power plant supplying a command and control node contributes effectively to military action even if an overwhelming percentage of its output is for civilian consumption (Study Group, pp. 335–36; DoD Law of War Manual § 5.6.1.2). This is the sense in which the term “dual-use” is often used, although it is not a legal term of art.
A common concern is partial use. To illustrate, consider (1) a POL storage facility in which only certain tanks are used to store military fuel, and (2) a large warehouse where military equipment ready for shipment is stored in a particular area. In the case of the POL tanks, only those serving the military make the requisite effective contribution. Whether a collective noun exists for a grouping – compound, facility, center, etc. – is legally irrelevant. Where a facility comprises several distinct structures, only those put to military use are candidates for military objective status; the rest remain civilian objects. They may not be attacked, and any incidental harm to them must be considered when assessing the proportionality of the strike and whether the attacker can take precautions to minimize harm to them (see below). Applied to the Wildberries strikes, each warehouse had to be assessed individually; a finding that one of them forwards drone components had no determinative legal bearing on whether other Wildberries facilities might qualify as military objectives.
As to the hypothetical warehouse, by the majority approach, the entire structure is a military objective, one defined by its physical integrity. Once any part contributes effectively to military action, the entire object is the objective; LOAC does not carve it into targetable and non-targetable parts (DoD Law of War Manual § 5.6.4.2; Study Group, p. 334). A minority position, one I endorse, is that where an attacker can identify and surgically strike only the portion of a structure that is in military use, the remainder retains civilian character (pp. 334–35; my article). In both approaches, civilians inside the facility retain their protected status and harm to them is subject to the proportionality rule and the precautions requirement.
Whereas use looks to what an object is doing now, purpose looks to its intended future use (AP I Commentary, ¶ 2022; DoD Law of War Manual § 5.6.6.1). From a military perspective, the criterion is indispensable. The key idea is that one should not have to wait for an object to be used for military purposes when there is a high degree of confidence it will be. But this criterion is easily abused, because the future is a matter of inference rather than observation. If the mere possibility that a civilian object might someday serve a military end sufficed, the number of targetable objects would be astronomical.
Accordingly, the law provides that qualification as a military objective by purpose necessitates evidence of intended enemy military use that is reliable enough for a reasonable commander, based on the information reasonably available at the time, to conclude that the object will in fact be so used – not speculation, worst-case assumptions, or mere capacity for conversion to military use (Study Group, pp. 332–333).
The Wildberries warehouses attacks illustrate the distinction. If the warehouse held commercially available, off-the-shelf electronic components that could be used for drone assembly but had not yet been purchased by the military or otherwise definitively committed to military production, those items and the warehouse retained their civilian object status. The mere possibility that stock might later enter the military supply chain amounts to speculative future use that the purpose criterion excludes. But if the facilities were functioning as a node in that supply chain, regularly channeling components to military end-users, the warehouse’s current use would qualify it as making an effective contribution, and any items in it destined for use by the military but not yet delivered would satisfy the first prong of the military objective test.
The Definite Military Advantage Prong: Satisfaction of the effective contribution prong is a necessary condition for transforming a civilian object into a military objective, but not a sufficient one. An object constitutes a lawful target only if its total or partial destruction, capture, or neutralization “offers a definite military advantage in the circumstances ruling at the time.” The adjective “definite” denotes “a concrete and perceptible military advantage, rather than one that is merely hypothetical or speculative” (DoD Law of War Manual § 5.6.7.3).
And the advantage gained must be a “military advantage”; an advantage that is merely political, economic, financial, or psychological does not suffice, nor would the generalized aim of pressuring an adversary to abandon the fight (Study Group, pp. 342–44). Unfortunately, in the context of the current conflicts, the Eritrea-Ethiopia Claims Commission somewhat blurred this line when it reasoned that “there can be few military advantages more evident than effective pressure to end an armed conflict” (Partial Award, Western Front, ¶ 121). But to be clear, the fact that attacking a civilian target might coerce the enemy to the bargaining table reads the term “military” out of “military advantage.”
In paradigmatic cases, the second prong of the military objective test will be readily satisfied once the first is fulfilled, as when destroying a warehouse will interrupt the imminent delivery of military drone components or where disabling a refinery will disrupt fuel supplies to enemy forces. But the second prong remains indispensable because it is a mechanism by which military objectives return to civilian status. Because status is assessed “in the circumstances ruling at the time,” a civilian object that becomes a military objective reverts to civilian status when that use ends, and there is no reasonable basis for concluding it will be used for military purposes in the future. In a protracted conflict fought with intelligence that can quickly become stale, this is an important limitation on finding otherwise civilian infrastructure targetable.
The War-Sustaining Debate
The so-called “war-sustaining” controversy mentioned above is at its heart a dispute about whether revenue makes an “effective contribution” to military action (the first prong), although it implicates the “definite military advantage” requirement (the second) as well (on the subject, see the careful treatment by Casey Biggerstaff). The United States interprets the definition of military objectives as extending to objects that make an effective contribution to the enemy’s “war-fighting or war-sustaining capability,” a claim the DoD Law of War Manual defends on the ground that “it is not necessary that the object provide immediate tactical or operational gains or that the object make an effective contribution to a specific military operation.” (§ 5.6.6.2; see also the analysis by Ryan Goodman). By this view, infrastructure whose economic output finances the war (oil destined for export being the paradigmatic example) may be attacked because the revenue it generates “sustains” the enemy’s capacity to fight.
Most States and commentators reject the interpretation, and rightly so (Study Group, pp. 340–41; AMW Manual, rule 24 commentary). In their estimation, the connection between sources of government revenue and military action is too attenuated to constitute an “effective contribution.” Indeed, nearly every productive activity in a belligerent’s economy could, on the same logic, be said to sustain the war effort. Moreover, the military advantage anticipated from destroying a revenue-generating facility is not only remote, but sometimes far from “definite,” as it is subject to markets, national budgets, procurement decisions, and a myriad of other factors that may never materialize during the conflict at hand.
But the absence of a war-sustaining justification does not foreclose attacks on revenue-generating infrastructure that independently satisfies the definition. A refinery that provides fuel for the enemy armed forces is a military objective by use, regardless of any revenue it generates (or anything else it does).
Protections That Survive Conversion to “Military Objective” Status
Two categories of objects of particular relevance to the ongoing conflicts remain protected even if they satisfy the requirements for a military objective. Objects indispensable to the survival of the civilian population receive special protection under Article 54 of Additional Protocol I. They may not be attacked to deny their sustenance value to civilians, although there are limited exceptions for objects used solely by enemy forces or in direct support of military action. Even then, an attack is prohibited if it may be expected to leave the civilian population with food or water so inadequate as to cause starvation or forced movement.
The ICRC views this prohibition, which is particularly relevant to the desalination strikes against Kuwait, as generally customary in character, an assertion with which I agree (Customary IHL study, rule 54). The United States accepts the underlying prohibition on the starvation of civilians but does not regard every element of Article 54 as customary in character (DoD Law of War Manual § 5.20.4).
Works containing dangerous forces – dams, dykes, and nuclear electrical generating stations – enjoy heightened protection against attacks that may release those forces and cause severe civilian losses; the protection ceases only in very narrow circumstances (AP I, art. 56). Note that the treaty prohibition reaches generating stations, but not enrichment or research facilities. Additionally, the rule binds Protocol parties like Russia and Ukraine; non-Parties need only comply with the less rigorous customary law requirement of “particular care” captured in Rule 42 of the ICRC’s Customary IHL study (on the subject see the pieces by Tom Dannenbaum here and here).
Can Reprisals be Lawful Attacks Against Civilian Objects?
Reprisals are actions taken by a belligerent against the enemy that would otherwise be unlawful, such as attacking civilian objects, but for the fact that they are intended to coerce the enemy into desisting from its own unlawful conduct. Additional Protocol I expressly bars such reprisals against civilian objects for Parties to the instrument, such as Russia and Ukraine (art. 52(1)). The United States disputes the customary status of the prohibition as applied to civilian objects (DoD Law of War Manual § 18.18), and the ICRC has not asserted that such attacks are altogether banned as a matter of customary law (Customary IHL study, rule 147).
However, even under customary law, reprisals against civilian objects are generally not lawful. To pass muster, they must be preceded by a demand that the enemy cease its unlawful action; may be taken only after other means of convincing the enemy to comply with LOAC have been exhausted; the enemy has been warned about reprisals and afforded an opportunity to comply; must be proportionate; and can be authorized only at a very senior level – conditions that the tit-for-tat exchanges we are seeing between Iran and its enemies have not met. Moreover, reprisals apply only to the State responsible for the original violations. For instance, Iran cannot claim to be engaging in a reprisal against alleged U.S. LOAC violations by attacking targets in Kuwait, such as the desalination facility.
And reprisals can only be taken by the victim State. Yesterday, President Trump threatened that “From this point forward, any time the Islamic Republic of Iran shoots at a ship in the Strait of Hormuz, whether it be by Missile, Rocket, Drone, or any other device or weapon, the United States will bomb and destroy ONE BRIDGE OR POWER PLANT, including those located next to, or in, the Capital City of Tehran.” This appears to be a threat of reprisal. But taking the action would be unlawful unless the bridges or power plants first qualified as military objectives. This is because the U.S. may not launch reprisals unless it were the victim of Iran’s unlawful action – it is not – and all of the other requirements listed above had been met.
Conversion to Military Objective Status Does Not Alone Justify Attack: Proportionality and Precautions
Establishing that an object is a military objective is only the first step in the lawfulness assessment. The attacker must still determine whether any expected incidental injury to civilians and any collateral damage to civilian objects are “excessive” relative to the concrete and direct military advantage anticipated from the strike (AP I, art. 51(5)(b); Customary IHL study, rule 14). If so, the strike is prohibited under the rule of proportionality.
The proportionality assessment must also consider so-called “reverberating effects.” Infrastructure attacks can cause civilian harm well beyond the immediate blast radius. Disabling electricity, for example, may interrupt hospital care, water purification, sewage treatment, and emergency response. To the extent that resulting civilian death, injury, illness, or damage to civilian objects is reasonably foreseeable, those effects must be factored into proportionality and precautions in attack assessments. The outer limits of the obligation remain unclear, but it is beyond dispute that an attacker may not confine its analysis to the immediate kinetic effects of an attack.
An attacker must also take “all feasible precautions” to avoid harming civilians and civilian objects, in particular by using available means to verify the target’s status and by considering alternative targets, weapons, and tactics that can produce the desired military effect with less risk to protected persons and objects (AP I, art. 57; Customary IHL study, rules 15–21). For instance, in the case of the Wildberries warehouse strikes, the presence of a night shift did not shield the warehouses from attack if they were in fact military objectives and striking them would yield a sufficient military advantage. But the decision to strike in the early morning hours may well have been made, and rightly so, because fewer civilians would be present.
Concluding Thoughts
The bottom line is crystal clear. Civilian infrastructure is not targetable merely because it is economically significant, broadly associated with the war effort, or politically useful to strike. It becomes a military objective only when, in the circumstances at the time, it makes an effective contribution to military action and attacking it offers a definite military advantage. That judgment must be made object by object, not sector by sector, and may change as the facts on the ground change. Even then, qualification as a military objective is only the first step; the rule of proportionality and the requirement to take precautions in attack to minimize civilian harm, among others, must independently be satisfied.
In the present conflicts, the parties seem tempted to stretch the definition of a military objective; some have simply discarded it. This is a dangerous slippery slope, not only because doing so forfeits often valuable international support, but also because the law is designed to provide a moral compass for what is the most brutal of human endeavors. Abandoning that compass will benefit no one.
