The Myanmar
Conflict Monitor

1. Armed Conflict Fatalities Fell Sharply While the Civilian Share of Total Deaths Rose

Fatalities resulting directly from armed conflict events declined by 25.4 percent during the quarter, yet civilian fatalities decreased by only 7.0 percent. Consequently, the proportion of civilians among total fatalities climbed steadily each month, reaching 23.6 percent by the final month of the quarter.

2. Myanmar Military Ground Offensives Were the Primary Driver of Mass Civilian Displacement

Military counter-offensives across the dry zone (Kani, Myitche), Chin State (Tonzang, Thantlang), and Hpapun in Kayin State severely impacted non-combatants, with half of all recorded displacement concentrated in just 17 events. The two events with the largest recorded displacement both stemmed from the Kani offensive, each recording 30,000 times people were displaced.

3. Conflict Entrenchment in Central Strongholds Accompanied by Geographic Expansion to Peripheries

Sagaing and Magway regions accounted for 35.5 percent of all conflict-related fatalities, remaining the conflict's central lethal axis, even as conflict events in Ayeyarwady Region surged by 53.3 percent. The Arakan Army's (AA) thrust into Yegyi Township, which resulted in the single deadliest clash of the quarter, underscored the widening of the conflict's geography.

In early April (April 2–13), the Kachin Independence Army (KIA) and allied forces recaptured three outposts around Maw Wun in Hpakant Township that had been lost in 2025, after which the Myanmar military retook a position. On April 13, the military launched a human-wave assault against KIA positions on the hilltop using newly conscripted personnel and young militia members; the KIA held the position. This engagement underscored that recruits raised under the People's Military Service (conscription) Law, activated in February 2024, were being directly deployed in frontline assault operations.

In the first full quarter following the election, on April 17, military authorities announced an amnesty for 4,335 prisoners; President U Win Myint was among those released. However, a count by a prisoner assistance network (recorded by MCARI) found only 1,686 people actually released, highlighting a significant divergence between official pronouncements and ground realities. Toward the end of the month, on April 28, the military commenced a major corridor offensive in Kani Township to control the Monywa–Kalewa road, where mines along the route began to kill soldiers.

Following the corridor push initiated in late April, the military carried out systematic clearing operations across southern Kani recorded under the date of May 5, accompanied by airstrikes, arson, and the blockade of Chindwin River ferry crossings. Recorded displacement from these village clearances reached 30,000 (times people were displaced), followed by another 30,000 across Yinmarbin, Kani, and Mingin townships toward the end of May. Centered on transit corridor control, these operations made May the peak displacement month of the quarter.

Concurrently, the Arakan Army (AA) and allied forces engaged military forces near villages at the foot of the Rakhine Yoma in Yegyi Township, Ayeyarwady Region, on May 31, resulting in the quarter's single deadliest clash. This confrontation demonstrated the AA’s ability to project force beyond the Rakhine theater and push strategically into Ayeyarwady Region.

In June, Pyu Saw Htee-held villages, upon which the military relies for its ground operations, increasingly became frontline targets. Between June 3 and 5, a military aircraft struck the Pyu Saw Htee-controlled village of Kukko Su in Pauk Township, killing at least 20 local residents. People's Defense Forces (PDF) and allied units subsequently overran the village and police station, liberating more than 200 civilians held as human shields and capturing roughly 70 fleeing militia personnel. Similarly, on June 12 in Myaing Township, a military column of more than 400 was attacked and forced to retreat to Pyu Saw Htee bases; separate military columns advancing into southern Myaing targeted places where displaced people had gathered with airstrikes and drones, and recorded displacement reached 20,000 (times people were displaced).

At the close of the quarter, between June 22 and 28, an entire military column numbering 117 personnel, including officers, surrendered to the AA in Rakhine State. This event marked the largest single surrender of military personnel recorded during the quarter.

Total recorded conflict events rose by just 0.9 percent quarter-on-quarter. However, armed conflict events fell by 11.9 percent, whereas violence against civilians rose by 7.8 percent and coercion events increased by 12.7 percent. This empirical shift indicates a transition away from force-on-force engagements between opposing armed formations toward operations that directly exert coercive pressure and violence on civilian populations.

This compositional realignment was directly tied to the conduct of military ground offensives. When military columns advanced, operations were routinely accompanied by preparatory artillery shelling into villages, airstrikes, arson, and the detention of civilians, as documented in Kani, Myitche, and Shwebo. Consequently, "fire on civilian targets" was the second most frequent event type at 930 events—trailing only armed clashes (1,012 events)—while 584 incidents of property destruction and arson were recorded across the country.

Within coercion events, forced conscription and porter duty accounted for 302 recorded events. Meanwhile, incidents involving the use of civilians as human shields escalated monthly, rising to 28 events in June. These trends align with ground observations, such as the deployment of newly conscripted recruits in frontline assaults in Hpakant and the confinement of villagers as human shields in Kukko Su.

Certain sub-categories registered marked monthly shifts. Peaceful protests rose sharply in June to 92 events. In contrast, drone attacks within armed conflict events declined steadily across the quarter, falling to 18 events in June.

A total of 3,902 fatalities were recorded in Q2 2026, representing a 21.3 percent decline from the previous quarter and reducing the daily average from 55.1 to 42.9 deaths. This decrease was primarily driven by a 25.4 percent drop in fatalities within armed conflict events. The military and pro-coup forces lost 2,685 personnel, representing 68.8 percent of all recorded deaths.

Civilian fatalities fell by only 7.0 percent to 876 deaths. This marginal decline reflects the reality that civilians predominantly lost their lives not in crossfire during clashes, but as a result of artillery shelling, airstrikes, and ground clearance operations—tactics that did not abate. Within the primary event category "violence against civilians," 727 civilian deaths were documented. Among its sub-event types, "fire on civilian targets" events recorded 471 fatalities in total, while 201 "attacks on civilians" resulted in 293 deaths. Prominent incidents included an April 29 airstrike in Kanpetlet Township that killed 9 civilians repairing a bridge in an area without active clashes, and a June 17 airstrike by four military aircraft on residential quarters in Kyauktaw Township that killed 9 civilians.

Mines and improvised explosive devices (IEDs) were among the deadliest weapons used against advancing military columns. Although mine-related events fell by 19.0 percent to 119 incidents, fatalities rose by 2.8 percent to 408 deaths, of which 377 were military personnel. Two of the quarter's ten deadliest events occurred on strategic roads in Sagaing Region that the military was trying to control: the April 28 event in Kani, where mines struck soldiers along the road, and the June 4 mine ambush on a column in Yinmarbin. In the Yinmarbin engagement, a PDF official claimed that 58 military personnel were killed.

Fatalities were documented in only 875 events (14.6 percent of total conflict events). However, the 100 deadliest events accounted for 51.5 percent of all fatalities, demonstrating that aggregate death tolls were driven heavily by a small number of mass-casualty incidents. As an empirical caveat, in 189 events a death was reported without a number and recorded as 3; those events carry 633 deaths in all.

Sagaing and Magway regions accounted for 35.5 percent of all fatalities nationwide, remaining the primary focal point of the conflict. Sagaing recorded the highest event count at 1,146 incidents (+7.1 percent), but Magway suffered the highest death toll with 714 fatalities, as its event volume rose by 14.9 percent.

Magway's emergence as the deadliest region was driven by June battles surrounding Pyu Saw Htee militia strongholds (over the quarter, 74 events and 101 fatalities in Myaing; 86 events and 100 fatalities in Pauk), alongside the military occupation of Myitche in Pakokku Township from May 10 to June 2, which was marked by arson and killings. In Sagaing, the death toll was driven by offensives in Kani, Yinmarbin, and Indaw. In Indaw Township (124 events, 140 fatalities), military columns advancing from Katha pushed toward revolutionary-held Indaw town beginning April 30, recapturing surrounding villages in the Mawlu area in early May.

Several border and peripheral regions experienced substantial escalations in conflict events, marked by key operational developments:

• Ayeyarwady Region (Events +53.3%): The region recorded 274 fatalities, comprising 266 military personnel and 7 civilians. Yegyi Township suffered 237 fatalities—the second deadliest township nationwide—stemming from clashes as the Arakan Army pushed across the Rakhine Yoma into Ayeyarwady Region.
• Chin State (Events +44.2%): The military recaptured Tonzang and roughly half of Thantlang town, reopening the Hakha–Thantlang road. In Tonzang Township, despite only 15 recorded events, 88 fatalities occurred, making it the township with the highest fatality-to-event ratio in the country.
• Kachin State (Events +41.3%): Fierce battles over jade mining tracts in Hpakant Township generated 256 events and 271 fatalities, making it both the most active and the deadliest township in Myanmar. Elsewhere, in Kyaukpyu Township, Rakhine State, an AA ambush on May 30 targeting military columns emerging from the Taung Maw Oo naval base contributed to a township total of 163 fatalities. Mon State also registered a 51.3 percent rise in conflict events.

Several regions registered notable declines in conflict events: Eastern Bago (−38.1%), Kayah State (−34.3%), Northern Shan State (−31.9%), Tanintharyi Region (−30.2%), Yangon Region (−22.8%), and Mandalay Region (−22.4%). These reductions reflect a drop in recorded operational engagements rather than an improvement in security or stability.

Northern Shan State recorded 56 fatalities, of which 55 were civilians—the vast majority resulting from the Namhkan explosion. The 2025 China-brokered ceasefire serves as an important background factor in the region. Concurrently, several townships recorded high event volumes but minimal fatalities, including Sittwe (105 events, 5 fatalities), Kutkai (24 events, 3 fatalities), and Taunggyi (21 events, no recorded deaths).

Shifts in the geographic distribution of conflict produced varied local protection challenges:

• Dry zone: Repeated displacements affected families previously uprooted by war, while the blockade of Chindwin River ferry crossings prevented fleeing civilians from crossing the river.
• Chin State: Ground counter-offensives into urban areas previously held by revolutionary forces brought renewed airstrikes and drone attacks.
• Ayeyarwady Region: Rural populations in an area that had experienced relatively little warfare compared to the dry zone were suddenly exposed to frontline combat.

According to reporting sources, the Myanmar military was the acting party in 3,768 conflict events and was linked to 715 civilian fatalities, representing 81.6 percent of all civilian deaths (down from 87.9 percent in Q1). (Note: "Acting party" refers solely to the organization named by reporting sources and does not constitute a formal legal finding of criminal responsibility). Unidentified armed groups were linked to 7.9 percent of civilian deaths (up from 5.7 percent in Q1), while the Ta'ang National Liberation Army (TNLA) was associated with 43 civilian deaths stemming entirely from the single explosion in Namhkan.

The decrease in the military’s relative share of civilian fatalities was driven both by a net decline in deaths attributed directly to the military (falling from 828 to 715) and by an increased share resulting from other entities and unidentified actors.

Disaggregating the data into two primary analytical layers reveals where civilian vulnerability was concentrated:

• Protest and Crackdown Layer: Encompassing violence against civilians, coercion, and protests, this layer accounted for 748 civilian fatalities.
• Armed Conflict Layer: Direct force-on-force clashes between opposing armed combatants resulted in only 65 civilian fatalities. This stark disparity demonstrates that civilians are overwhelmingly killed not within the crossfire of clashes between combatants, but along the paths of advancing military columns through village clearances, detentions and killings, and aerial bombardments.

The military conducted ground operations alongside local armed groups and militias. In Tonzang, the Zomi Revolutionary Army (ZRA) fought alongside the military. In Myitche, police and Pyu Saw Htee militia accompanied the advancing column, while Kukko Su was directly garrisoned by Pyu Saw Htee forces. Because of this integration, Pyu Saw Htee strongholds became primary targets for resistance counter-attacks, as seen in Pauk and Myaing.

Territorial control data documented 98 instances of territory gained by the military compared to 61 by armed opposition groups, marking the quarter as an active period of military counter-offensives. While opposition forces led in territorial gains during April, the trajectory reversed decisively in favor of the military from May onward.

Several shifts occurred within the organizational landscape:

• Newly Recorded Groups: CDF (Thantlang) was documented in 10 events amid the battle for Thantlang. The military’s National Solidarity and Peacemaking Negotiation Committee (NSPNC) appeared in 6 events as ceasefire and peace discussion activities trended upward toward June.
• Absences: The Burma National Revolutionary Army (BNRA), which recorded 28 events in Q1, recorded zero events in Q2.
• Unidentified Armed Groups: Unidentified armed groups were recorded as one of the sides in 23.0 percent of all armed conflict events (up from 22.0 percent in Q1), underscoring that a significant portion of ground engagements cannot be attributed to a recognized organization.

A total of 37 targeted killing events resulted in 58 fatalities, including 15 civilians. Local administrators were the most frequently targeted category; 11 attacks directed at administrators resulted in 12 civilian fatalities, with victims targeted due to their administrative functions.

Recorded displacement totalled 708,424 times people were displaced (−17.4 percent) during the quarter, peaking in May at 310,373 as clearance operations intensified. Displacements were heavily concentrated, with just 17 events accounting for 50.5 percent (358,100 times) of the quarterly total.

The five events with the largest recorded displacement were in Kani Township on May 5 (30,000 times); across Yinmarbin, Kani, and Mingin townships on May 31 (30,000); in Myingyan Township on April 11 (20,000); in Tanintharyi Township on April 17 (20,000); and in Shwebo Township on April 22 (20,000). In Myingyan, a military column torched villages northeast of the town, and 8 civilians were killed in a village raid on April 2. Across the 355 events where displacement was recorded, the primary triggers were artillery and missile strikes (109 events) and forced displacement (104 events).

These patterns reveal that advancing columns forced entire village tracts to flee simultaneously. In the dry zone, ferry closures along the Chindwin River left displaced civilians trapped and unable to cross, while repeatedly displaced families experienced cyclical displacement, with single individuals displaced multiple times.

Nationally, 7,332 individuals (−21.0 percent) were detained, peaking in May at 3,230 during active clearing operations. Of these, 6,860 detentions occurred within coercion events. At the township level, Paletwa recorded the highest number of detentions at 540, followed by Monywa, Shwebo, Sagaing, and Pakokku. (An additional 438 individuals were detained in events where the township could not be confirmed).

Conversely, 5,978 people were released (−61.9 percent); 66 of the quarter's 94 prisoner-release events took place in April, when 4,176 people were released. While military authorities announced an amnesty of 4,335 prisoners in April, a count by a prisoner network found only 1,686 people actually released (MCARI record). This wide disparity between official announcements and actual releases heightened distress and uncertainty among families seeking information on detained relatives.

A total of 796 people defected or surrendered (−43.7 percent), though monthly figures rose steadily from 163 in April to 309 in May and 324 in June. The most notable event occurred in Rakhine State in late June, when an entire column of 117 personnel surrendered to the AA.

As combatants surrender to armed opposition groups, we assess that their treatment and conditions of detention are emerging as an important humanitarian and protection consideration.

Overall, non-combatants faced three compounding and interconnected pressures:

• Ground clearance operations that left villages torched and deserted;
• Forced conscription and portering; and
• The use of civilians as human shields during military maneuvers. These simultaneous pressures continue to create acute challenges for civilian survival and physical protection.

Based on operational trends observed during the quarter, the following dynamics warrant continued monitoring:

1. Sustained Nature of Military Counter-Offensives

Recorded territorial gains favored the military (98 to 61), reversing the earlier trajectory from May onward across Tonzang, Thantlang, the Hakha road, Indaw, and Hpapun. Given the three-month observation window, it remains to be seen whether these operations mark a sustained counter-offensive campaign or temporary operational advances.

2. Trajectory of the Ayeyarwady Front

Yegyi Township recorded 237 fatalities, with 200 occurring in a single clash on May 31. Because this development rests heavily on an isolated mass-casualty engagement, monitoring is required to determine whether the Arakan Army establishes a persistent operational presence in Ayeyarwady Region or whether this represented an isolated cross-Yoma raid.

3. Proportion of Civilian Fatalities

The civilian share of total fatalities climbed monthly (20.8% in April, 22.7% in May, and 23.6% in June). However, because this proportion is observed over a single quarter and remains sensitive to isolated mass-casualty events, further observation is necessary to assess whether it represents a durable shift in conflict lethality.

4. Mine Warfare Along Contested Dry-Zone Transit Routes

Despite low event numbers (119 events), mine warfare caused heavy casualties (408 deaths, primarily military personnel). This depends on which transit corridors military columns attempt to contest next.

5. Reliance on Civilians as Human Shields

Incidents involving the use of civilians as human shields rose monthly (from 9 in April to 21 in May and 28 in June). Given that this trend is seen over a three-month window only, continued monitoring is required to evaluate whether this practice becomes an entrenched operational tactic.