Deaths of Servicemen in Kazakhstan

The Caspian tragedy and the state’s responsibility for soldiers’ lives

Analytical article • 31 October 2026

On 24 September 2026, 14 Kazakh servicemen died in the Caspian Sea while taking part in the “Khazar Qorgany — 2026” military exercises. According to the Ministry of Defence’s final figures, 17 servicemen ended up in the water. Three survived. The bodies of the last two missing servicemen were recovered the following morning. [1]

The tragedy raises a broader question about what happens to a person once the state places them under military command. Who decides whether an order is too dangerous? Who has the authority to stop an unsafe exercise? And who is responsible when the measures meant to protect servicemen fail?

Individual criminal responsibility must be established by investigators and, ultimately, by the courts. But the broader question of how military service is organised cannot wait for a final verdict. Other servicemen are still serving today.

By the end of October, the case had moved far beyond the initial explanation of sudden bad weather. The authorities had acknowledged serious shortcomings in planning and safety, six officials had been remanded in custody, and a military reform plan had been approved. [8–10] The real test of the state’s response will be whether these events lead to changes capable of protecting the next generation of conscripts.

The last exercise before going home

One of the most painful details emerged the day after the disaster. Deputy Commander-in-Chief of the Naval Forces Arman Satkanov said that the conscripts who died had only eight days left before completing their service and that their tickets home had already been bought. His statement should not be taken to mean that all 14 victims were conscripts: the dead also included a senior lieutenant and a sergeant. [1–3]

Sailor Omirbek Omirzak was expected home on 5 October. According to his father, he planned to join the police after completing his military service. Behind the impersonal term “personnel” were people who had already begun planning their return home and the lives they would lead afterwards. [3]

The fact that some of them were only days away from discharge does not make their lives more valuable than those of new conscripts or career servicemen. But it does show how misleading the feeling can be that the greatest danger is already over. The duty of command to protect a serviceman does not diminish as the end of service approaches. A family’s final week of waiting should not end with a ticket home that will never be used.

Military exercises are intended to prepare people for real danger, not to create unnecessary danger in the name of training. Commanders must therefore always retain the ability to stop an exercise. If the risk becomes disproportionate to the purpose of the training, the exercise should be halted. That is not weakness or excessive caution. It is basic responsibility for human life.

What happened on the coast

The “Khazar Qorgany — 2026” exercises began on 22 September. According to Kazinform, more than 500 servicemen took part, along with naval ships and boats, hydrographic vessels, Coast Guard units of the Border Service of the National Security Committee, and aviation. Two days later, disaster struck during one of the landing exercises. [4]

A Ministry of Internal Affairs statement issued on 25 September said the incident occurred on the Caspian coast, 43 kilometres from Aktau. Servicemen were swept into the sea while disembarking from the “Qaisar” boat in poor weather. The figure of 43 kilometres refers to the distance from Aktau, not from the shoreline. [5]

Satkanov said that the men were hit by a wave during the landing. But that explanation alone does not tell us how the disaster unfolded. Investigators need to establish where the boat was positioned, the order in which the servicemen disembarked, the state of the sea and how quickly rescue teams could respond. These questions should be answered through documents, witness testimony and technical data, not assumptions based on a few public statements. [3, 5]

Initial reports said that nine people had died. By the evening, the confirmed death toll had risen to twelve. At around 09:15 on 25 September, the bodies of Sergeant Mansur Gumarov and sailor Miras Turdybekuly were found. When the search operation ended, the final toll stood at 14 dead and three survivors. [1, 4]

Three people were rescued. But that fact does not explain why the others died. The investigation must establish that separately. A general reference to bad weather or a large wave is not enough.

The names of those who died

On 24 September, the Ministry of Defence published the names of the first twelve victims. They were Senior Lieutenant Olzhas Mukhtaruly and sailors Serzhan Talğatuly Ayapbergen, Ömirbek Ömirbolatuly Ömirzaq, Almas Jüsipuly Köbenov, Aqylbek Qabdenuly Tölegen, Jeñis Tleuuly Temirğaliev, Meirambek Ermukhanuly Berdikhanov, Äset Samatuly Jañaba, Azamat Bolatbekuly Mädikhan, Dastan Nūrboluly Begei, Aqjaiyq Bolatbekuly Gilazh and Olzhas Täjibaiuly Dikhanbaev. The following day, the deaths of Mansur Gumarov and Miras Turdybekuly were confirmed. The names are reproduced here as they appeared in the relevant official statements; other publications use different transliterations. [1, 2]

It would be inaccurate to describe all of the dead as conscripts. But the fact that conscript sailors were among them matters. The state drafts a person into military service, decides where they will serve and requires them to follow orders. A conscript cannot simply choose which risks to accept and which to refuse. That gives the chain of command a particularly serious responsibility for their safety.

It also matters that the dead are named. Behind the number “14”, it is easy to lose sight of individual lives and families and to reduce the discussion to institutional failures, reputational damage or personnel decisions. Each of those who died had a life of his own and people waiting for him at home. Establishing how and why he died is not a bureaucratic formality. It is an obligation to him and to those close to him.

What “sudden bad weather” actually explains

In its first statement, the Ministry of Defence attributed the tragedy to a sudden and sharp deterioration in the weather. Yet Kazhydromet had already published a forecast on 23 September for the period 24–26 September, warning of unstable weather and stronger winds across much of the country. That general forecast does not, by itself, establish what conditions were like at the exact landing site. [1, 6]

More specific information was later provided by Kazhydromet spokesperson Aruzhan Muratova. According to a statement published by Informburo.kz, south-easterly gusts of 15–20 metres per second were forecast for 24 September in western, northern and central parts of Mangystau Region. Weather stations later recorded easterly winds of up to 15–16 metres per second. But these figures cannot simply be applied to the site of the disaster. Exact coordinates, timing and information about sea conditions are still needed. [7]

The important question is therefore not only what forecast existed, but what the organisers did with it. When did they receive it? How was the risk assessed for the chosen location? Was the landing plan changed in response? Equally important is what happened immediately before the operation. Even if conditions appeared acceptable in the morning, commanders still had to monitor how they changed during the day.

A particular wave may indeed have been unexpected. The broader risk of deteriorating weather was not. Safety procedures exist precisely because events do not always unfold according to plan. Bad weather can only become a meaningful explanation if it is also shown why the warning, monitoring and protective measures in place did not prevent people from dying.

Rescue must be planned in advance

The Ministry of Internal Affairs said that investigators were examining compliance with safety requirements, the technical condition of the boat and the actions of officials. That does not mean that a technical fault or a lack of lifesaving equipment has already been established. Such conclusions can only follow from the investigation. [5]

It is important to reconstruct not only the moment when the servicemen entered the water, but everything that led up to it. What equipment were they issued? Was it suitable for this type of landing? Was it checked beforehand? Did the participants know what to do if the exercise went wrong? The mere fact that a safety briefing took place says little unless its content and the practical preparation of the participants are also examined.

Investigators also need to establish what rescue capacity was already in place before the landing began. The large-scale search that followed shows that considerable resources were later mobilised. But that does not answer the more immediate question: could people have been pulled from the water in the first minutes after danger arose?

A precise timeline is needed — when the danger was first noticed, when the landing was stopped, when an alarm was raised, when rescuers reached the area, and when each serviceman was found.

It would be wrong to claim in advance that different equipment or a different deployment of rescuers would necessarily have saved everyone. That is a matter for professional assessment. But an investigation must answer not only what happened, but whether the deaths could have been prevented.

From the first detentions to six suspects

On the evening of 24 September, a government commission chaired by Deputy Prime Minister Kanat Bozumbayev heard a report on the investigation. According to an official statement by the Mangystau Region administration, criminal proceedings were opened under Part 2 of Article 453 of Kazakhstan’s Criminal Code, concerning negligent performance of official duties resulting in serious consequences, and Article 465, concerning violations of navigation rules. [11]

On 25 September, the Ministry of Internal Affairs announced the detention of five officials: the Chief of Staff and First Deputy Commander-in-Chief of the Naval Forces, two military unit commanders, the commander of a marine company and the commander of the boat. The ministry said investigators were examining decision-making, safety arrangements, the handling of hydrometeorological information and the vessel’s technical condition. [5]

The number of suspects later increased. On 6 October, the Ministry of Internal Affairs confirmed the detention and remand in custody of the former Commander-in-Chief of the Naval Forces. On 14 October, in a response to Informburo.kz, the ministry said that six officials responsible for organising and conducting the exercises were being held in custody and that the possible involvement of other individuals was still under examination. [8, 12]

These developments show the direction of the investigation. They do not establish the guilt of any particular suspect. Pretrial detention is a preventive measure, not a conviction. Final conclusions about criminal responsibility must rest on evidence and judicial proceedings.

At the same time, the scope of the investigation matters. Decisions about military exercises are made at several levels. It is not enough to examine only the actions of the person physically present at the landing site if the original plan, logistical support and systems of control were determined higher up the chain of command. Investigators need to establish what each level was responsible for and how much influence it had over what happened.

The authorities acknowledged failures in the preparation of the exercises

On 27 September, President Kassym-Jomart Tokayev ordered a review of the concept behind the exercises, their relevance to contemporary threats and the role of the General Staff. In his published remarks, he criticised irresponsibility, a showpiece approach and deficiencies in crisis-response planning. These were political assessments by the head of state, not judicial findings concerning individual defendants. [13]

On 5 October, Akorda published the preliminary findings of a comprehensive inspection of the Ministry of Defence. The official statement referred directly to serious shortcomings in the preparation and planning of the exercises, risk assessment, organisation of the landing and measures intended to protect servicemen. It also referred to accumulated systemic problems and inadequate oversight by the previous leadership. [9]

As part of the wider review, 13 criminal cases were opened concerning possible embezzlement, abuse of office and other violations. It would be misleading to present this figure as the number of criminal cases directly connected with the deaths in the Caspian Sea. The statement concerned the work of defence institutions and units more broadly. [9]

These preliminary findings change the nature of the discussion. Questions about how safety was organised are no longer being raised only by journalists or relatives. The authorities themselves have acknowledged that there were serious failures.

Replacing commanders may create an opportunity for change. But personnel changes alone do not prove that the underlying causes have been addressed. If the next commander inherits the same incentives, the same reporting culture and the same barriers to raising safety concerns, the risk may survive any reshuffle.

The Caspian has already claimed the lives of conscripts

On 18 September 2015, Kazakhstan’s Ministry of Defence reported another fatal incident near Aktau during the “Tsentr-2015” exercises. Six BTR-80 armoured personnel carriers and 52 servicemen were involved in an amphibious landing. According to the initial official report, four armoured vehicles sank after a powerful gust of wind and high waves. Four conscripts were initially reported dead. [14]

The incidents of 2015 and 2026 should not be treated as technically identical. The earlier case involved armoured personnel carriers entering the water, while the 2026 disaster occurred during a landing from a boat. But both involved the sea, amphibious training, deteriorating weather and loss of life. That makes it reasonable to ask whether lessons from the earlier tragedy were carried forward.

After a serious accident, an institution should retain what it has learned. Those lessons should be reflected in procedures, commander training, equipment requirements and rescue preparedness. Eleven years later, it is legitimate to ask what recommendations were made after the 2015 deaths, which of them were implemented and whether they still shaped planning in 2026.

This does not prove that the new disaster was caused by a failure to implement earlier recommendations. It is simply one of the questions that should be examined if future prevention is to be meaningful. An accident report has practical value only if its conclusions are used by those planning the next operation.

Why the statistics need to be handled carefully

Concerns about servicemen’s deaths in Kazakhstan long predate September 2026. On 18 April 2023, First Deputy Chief Military Prosecutor Maksat Kaziev told the Senate that 270 servicemen had died since the beginning of 2020: 78 in 2020, 96 in 2021, 85 in 2022 and 11 from the beginning of 2023 up to the date of his statement. He cited illness, road accidents and suicide among the causes. [15]

These figures should not be described as the number of conscripts killed. They include different causes of death, and the published data do not allow every case to be assigned to a particular type of service or institution. Nor can meaningful mortality rates be calculated without data on the number of people serving, their age, service conditions and length of service.

But careful interpretation does not make these deaths less important. It helps identify the right response. Preventing deaths from illness requires one set of measures. Transport safety requires another. Preventing violence and suicide requires something different again.

Meaningful public oversight requires regular and comparable statistics: the circumstances of deaths, type of service, outcome of investigations and measures taken afterwards. Such data should protect personal information while still making recurring patterns visible. Hiding important distinctions undermines prevention just as much as deliberately mixing unlike cases together.

Conscription increases the state’s responsibility

A conscript and the organiser of a military exercise do not have equal control over the conditions in which an operation takes place. Commanders choose the location, timetable, equipment and procedures. The serviceman acts within decisions that have already been made for him. That gives commanders a duty to justify the risks they impose and to provide adequate protection.

Military training can never be entirely free of danger. But that fact should not become a blanket justification for avoidable risk. Training always requires a balance between its purpose and the actual conditions in which it takes place. A level of danger that may be unavoidable in a real emergency is not automatically acceptable during a scheduled exercise.

A person who reports unsafe conditions is giving the chain of command information it needs to make a better decision. If every such warning is treated as weakness or unwillingness to perform a task, the system loses the opportunity to correct mistakes before they become fatal. Professional command must therefore be able to distinguish between evading a duty and raising a genuine safety concern in good faith.

There is no basis, at this stage, for claiming that anyone warned of danger before the Caspian incident and was ignored. But investigators should examine this as well: whether concerns were raised from lower levels, how they moved through the chain of command and whether the responsible officer or specialist had a real ability to stop the landing. For future servicemen, that practical ability matters as much as the wording of any written instruction.

Support for families and access to the truth

Kazakhstan declared a national day of mourning on 25 September. The authorities also announced financial and psychological assistance for the families of those who died. The government commission specifically instructed that aid, including donations from private benefactors, should be distributed transparently and according to need. [4, 11]

Such support is essential. After losing someone close, a family faces not only grief, but expenses, paperwork and a sudden change in everyday life. But financial assistance cannot replace answers about how and why a person died. These are separate responsibilities.

Relatives should be able to understand how the investigation is progressing, whom they can contact and what questions investigators are examining. They may also need legal assistance and an opportunity to raise questions with experts.

An investigation does not have to make all of its evidence public while proceedings are still under way. But confidentiality should not become a reason for sending families formulaic replies that tell them nothing about what is actually being examined.

What the October military reform promises

On 21 October, Kazakhstan’s Security Council approved a military reform plan. According to Akorda, compulsory service is expected to be divided into two stages: six months of training at specialised centres followed by six months of service in military units. The authorities also announced plans to create a system of military advisers and an office of Commissioner for Servicemen’s Affairs. Organisational changes are planned by the end of 2026, while the main conceptual measures are scheduled for 2027. At this stage, these are announced plans rather than completed reforms. [10]

The value of a future commissioner will depend on the powers attached to the office. Servicemen should be able to raise concerns confidentially, expect an independent review and be protected from retaliation simply for making a complaint. If a complaint ultimately goes straight back to the same command structure that is being complained about, a new institutional title will change little.

For safety during military exercises, several practical things matter: commanders must be able to stop a dangerous phase in time, rescue teams must be ready to intervene immediately and the process for assessing risk must be clear. Simply changing the duration or sequence of training guarantees nothing. The important question is whether servicemen become better prepared and whether commanders make safer decisions.

Trust in the armed forces cannot be rebuilt through public messaging alone. For servicemen and their families, practical experience matters more: whether injured people receive proper care, whether complaints are taken seriously and whether identified failures are actually corrected. Good results are more credible when failures can also be examined independently.

What would count as real change

The investigation must reconstruct the full chain of events. But there is no reason to wait for a final criminal judgment before correcting dangerous practices that have already been identified.

Military reform should be judged by what changes in practice. New instructions must reach units and be used. Reporting a danger should not lead to punishment for the person who raised the alarm. Families of the dead should receive information without having to turn every request into a public campaign. And inspection reports should explain not only what went wrong, but what was changed afterwards.

The Caspian disaster points to a simple principle: the security of the state cannot be separated from the safety of the people who serve it. A serviceman does not lose his rights when he receives an order. The more dependent he is on decisions made by commanders, the greater the responsibility of those making those decisions.

Fourteen people did not return from these exercises. The investigation will matter if it changes the conditions under which others serve. The most important result will not be another report or another sentence added to a regulation, but a system capable of recognising danger in time, stopping an unsafe task and bringing people home alive.

Sources and note on the article

This article was prepared on the basis of the monitoring material provided and a review of publicly available sources as of 31 October 2026. The factual section was supplemented with reports published on 5, 6, 14 and 21 October. The assessments and proposed measures are the author’s analysis. In the materials reviewed, the investigation is described as ongoing; no final judicial findings concerning the guilt of specific individuals are cited.

[1] Ministry of Defence of Kazakhstan. Bodies of two servicemen missing in the Caspian Sea found; search operation completed. 25 September 2026.
Open publication

[2] Kazinform, citing the Ministry of Defence. Names of 12 servicemen killed during exercises in the Caspian Sea announced. 24 September 2026.
Open publication

[3] Kazinform. Eight days remained before they were due to return home. Statement by Arman Satkanov and the account of Omirbek Omirzak’s family. 25 September 2026.
Open publication

[4] Kazinform. Deaths of servicemen in the Caspian and information available one day after the incident. 25 September 2026.
Open publication

[5] Kazinform, citing the Ministry of Internal Affairs. Five Defence Ministry officials detained in connection with servicemen’s deaths in the Caspian. 25 September 2026.
Open publication

[6] Kazhydromet. Weather forecast for Kazakhstan for 24–26 September 2026. Published 23 September 2026.
Open publication

[7] Informburo.kz. What was the wind speed in Mangystau on 24 September? Statement by a Kazhydromet spokesperson. 24 September 2026.
Open publication

[8] Informburo.kz. Six suspects in the case concerning the deaths of sailors in the Caspian remain in custody. Ministry of Internal Affairs response to a media request. 14 October 2026.
Open publication

[9] Akorda. Head of State hears the first results of the Commission reviewing the Ministry of Defence. 5 October 2026.
Open publication

[10] Akorda. Head of State chairs Security Council meeting on military reform. 21 October 2026.
Open publication

[11] Mangystau Region administration. Government commission investigates circumstances of servicemen’s deaths in Mangystau Region. Statement concerning the meeting of 24 September 2026.
Open publication

[12] Interfax, citing Kazakhstan’s Ministry of Internal Affairs. Former Commander-in-Chief of Kazakhstan’s Naval Forces detained after the deaths of 14 servicemen in the Caspian. 6 October 2026.
Open publication

[13] Akorda. Head of State instructs officials to prepare a military reform plan. 27 September 2026.
Open publication

[14] Kazinform, citing the Ministry of Defence. Four conscript soldiers drowned during exercises in the Caspian. Initial report. 18 September 2015.
Open publication

[15] Kazinform. 270 servicemen have died in Kazakhstan since 2020. Remarks by First Deputy Chief Military Prosecutor Maksat Kaziev in the Senate. 18 April 2023.
Open publication