Assessing Violence in Wartime

The return of large-scale war has become one of the defining features of the 2020s. Russia’s invasion of Ukraine, the wars in Gaza and Sudan, and interstate confrontations elsewhere have effectively overturned post–Cold War assumptions about the decline of high-intensity armed conflict. In 2025, the Uppsala Conflict Data Program (UCDP) recorded 65 state-based armed conflicts — the highest number since its series began in 1946 — and approximately 245,000 deaths from organised violence, seven times the number recorded in 2010 and the second-highest annual toll since the Rwandan genocide in 1994. But these aggregate figures tell only part of the story. Equally important are changes in the character, targeting, and spatial distribution of violence.

The Return of Interstate War

The resurgence of interstate conflict is the clearest break with the recent past. UCDP recorded two interstate conflicts in 2023, four in 2024, and eight in 2025, the highest number in the post-1945 period. While most armed conflicts — rom Sudan and Myanmar to the Democratic Republic of Congo and the Sahel — remain intrastate, struggles over political power, territory and exclusion, often expressed through ethnic or nationalist claims, continue to sustain civil violence. Even so, the return of interstate war should not obscure the continuing violence inflicted on populations by states and armed groups within their own borders. These dynamics also cross the interstate–intrastate divide: nationalism and territorial claims can fuel interstate war, while civil wars increasingly draw in outside states and armed groups operate across borders. Even so, the return of interstate war should not obscure the continuing violence inflicted on populations by states and armed groups within their own borders.

The Selectivity of Violence

A second dimension concerns how violence is targeted: is force directed selectively at identifiable individuals or more broadly against populations? New military technologies have greatly increased the capacity for selectivity. Drones, surveillance, precision-guided munitions, and AI-assisted targeting permit targets to be identified and struck with unprecedented accuracy. But precision and selectivity are not synonymous. A munition can strike precisely where intended while producing wide-area effects, particularly when used in densely populated environments. The technological capacity to discriminate among targets has therefore increased considerably, without necessarily making wartime violence itself more selective.

Casualties in Wartime Violence

The technological capacity to discriminate among targets has therefore increased considerably, without necessarily making wartime violence itself more selective. The composition of casualties provides a different measure of contemporary violence. In the Russia–Ukraine war — an exceptionally lethal conventional conflict in which combatants constitute the lion’s share of direct deaths — Ukraine accounted for roughly three-fifths of state-based battle deaths worldwide in 2025. Sudan presents a very different pattern. UCDP recorded a fivefold increase in deliberate one-sided violence against civilians between 2024 and 2025 — from around 14,300 to 76,500 deaths — driven largely by mass killing in Sudan. Action on Armed Violence (AOAV) similarly recorded nearly 50,000 deaths and injuries from explosive weapons in 2025; 91% were civilians, rising to 96% when these weapons were used in populated areas.

Territory and the Geography of Violence

A final dimension concerns the drivers and geography of violence. Territorial conquest has returned to prominence, most visibly in Ukraine. With the exception of 2023, every year since 2020 has seen at least one new or reactivated interstate territorial dispute. But violence is also used to control populations, punish collaborators, extract resources, and displace communities with such processes often driven by local dynamics. Our earlier work on Rwanda and Burundi shows how the scale and spatial extent of ethnic violence can vary markedly across otherwise comparable episodes, as patterns of trust, individual behaviour, and social norms interact to determine whether violence remains localised or escalates.

These local dynamics also shape the geography of contemporary violence, concentrating it in particular hotspots: front lines, border regions, transport corridors, cities, and neighbourhoods. Cities have become especially consequential sites of high-intensity warfare. Gaza, Khartoum, and Ukrainian cities illustrate the effects of combining heavy explosive weapons with dense civilian settlement and vulnerable infrastructure. However, this does not constitute a simple shift from rural to urban warfare. Rural insurgency remains central in contexts like the Sahel, eastern Congo and Myanmar.

In sum, no single trend defines violence in the post-2020 period.  In sum, no single trend defines violence in the post-2020 period. Interstate war has become more prominent without displacing civil conflict, while selective targeting coexists with large-scale civilian harm. Territorial conquest has returned, but much violence remains localised, driven by struggles over control, identity, coercion and displacement. Its geography is equally uneven, concentrated in particular conflicts and local hotspots. Yet, the category of war itself does not capture all the organised violence characteristic of the period. In Mexico alone, organised crime has been associated with roughly 20,000 homicides annually during much of the 2020s. Assessing contemporary violence therefore requires looking beyond conventional categories and aggregate casualty counts to examine whom violence affects, why it is used, and where it is concentrated.