CODE 9.2, Ukraine’s laboratory of contemporary warfare

By integrating pilots, technicians and electronic-warfare specialists into its assault units, Ukraine’s 475th Regiment shortens the chain between the front line’s need and the technical solution.

Its organisation sheds light on a major transformation: the speed of adaptation now matters almost as much as the performance of the weapons themselves.

From improvised prototype to mass production, Ukraine is inventing an « innovation loop » in which the fighter becomes user, evaluator and co-producer of the weapon.

In an interview granted in Ukraine to Michael Kofman, a researcher specialising in the Russian army and host of the podcast The Russia Contingency for War on the Rocks, Oleksandr « Flint » Nastenko describes the organisation of the 475th Separate Assault Regiment, which he commands. CODE 9.2 is not simply an infantry unit equipped with drones. The regiment rests on a different architecture, combining a relatively small assault force with an air and robotic component far larger than in a conventional formation. Drones there are not an external support requested ad hoc before an operation; their operators and technicians belong organically to the regiment and take part both in the manoeuvre and in the very design of the systems. The interview thus helps us understand how this close integration, tested in particular from Kupiansk to Zaporizhzhia, is simultaneously transforming reconnaissance, strikes, the protection of infantrymen and the conduct of the assault.

The regiment also has its own artillery, armoured vehicles, reconnaissance and attack drones, robotic ground platforms and capabilities designed to strike in depth. When Nastenko announced the creation of this formation in early 2025, he already explained that he wanted « sufficient » infantry to accomplish the mission, but also a very powerful company of unmanned aerial systems and a large detachment of ground robots. The objective was explicit: to reduce the number of men exposed on the line of contact without giving up the ability to advance and retake ground. Nastenko’s project also included tanks, a fire-support company and observation or strike assets able to act up to 120 kilometres away.

This organisation was built up gradually. CODE 9.2 was born in 2022 as a small reconnaissance element of the 92nd Mechanised Brigade. The drones first served to observe, then to correct fire, and finally to attack directly. The detachment became an attack-drone company in 2024, an assault battalion in early 2025, then a regiment a few months later. By August 2025, according to the Kyiv Post, its drone component had reached a size almost comparable to that of the brigade from which it had sprung. This growth therefore reflects more than an increase in numbers. It corresponds to the emergence of a new type of formation, specialised in combining ground operations and unmanned systems. The trajectory shows how an initially marginal innovation can end up transforming the structure of the unit itself.

The technicians are no longer in the rear

The most important aspect of the model, however, does not lie in the number of drones alone. CODE 9.2 organically integrates technical specialists: electronics engineers, mechanics, programmers, communications and electronic-warfare specialists. These men and women wear the uniform, belong to the regiment and work directly with the operators and the infantrymen. This proximity profoundly changes the circulation of information.

In a conventional organisation, the fighter reports a difficulty to his command. It travels up the chain of command, then is passed on to the departments in charge of programmes and procurement. The manufacturer then receives a set of specifications, designs a solution, has it tested, certified and produced. Between the problem discovered in the field and the delivery of the corrected equipment, months — sometimes years — can go by. In the meantime, on the ground, technical work-arounds are improvised and procedures are modified to get round the problem.

In CODE 9.2, the distance between the user and the technician can shrink to a few metres. The pilot who loses control of a drone because of jamming can immediately explain what he saw on his screen. The signals specialist can examine the frequencies used, recover the aircraft if it has come back, analyse the components and modify the antenna, the software or the flight parameters. The new setting can sometimes be tested on the very next mission.

The technician therefore no longer intervenes only after the battle, in a distant workshop. Fully integrated into the unit, he takes part in the operational cycle and knows the terrain, the fighting methods, the pilots’ real performance and the Russian countermeasures encountered in the sector. Conversely, the operators better understand the technical possibilities and limits of their equipment. The technicians can also talk more effectively with the engineers in the rear, whose language and professional references they share. The military need and the technical response thus cease to belong to two separate worlds.

« The military need and the technical response cease to belong to two separate worlds. »

This integration is particularly important in electronic warfare. A drone that is effective in one sector can become almost unusable a few dozen kilometres away, because the Russian units there do not jam the same frequencies or do not use the same detection systems. A solution certified six months earlier can be outdated before it even reaches the front in quantity. Jamming prompts a change in the data links, which triggers an adaptation of the enemy’s jamming, then a change of frequency, a better antenna or fibre-optic guidance, which in turn forces the design of a new countermeasure. The cycle is measured in weeks, sometimes in days.

The front becomes a laboratory

CODE 9.2 pushes to its limit an evolution observable in many Ukrainian formations. Some brigades now have research-and-development cells close to the front. They dismantle captured equipment, repair damaged drones, print parts, adapt explosive charges, modify navigation software and test new data links.

These workshops obviously cannot manufacture millions of systems on their own. Their function is different: to produce a solution quickly, confront it with combat and pass on to industry what deserves to be manufactured on a larger scale. The long-range Pavuk drone, today produced with the company Omnitech, thus came out of a laboratory of the 3rd Assault Brigade. According to a study devoted to Ukraine’s battlefield innovation loop, several items of equipment now in use throughout the army were born as prototypes cobbled together by a unit to solve a local problem.

The fighter thus becomes at once user, evaluator and, in some cases, co-producer of the weapon system. A section commander can send the manufacturer, via Signal or WhatsApp, a video showing the breakdown of a ground robot. Some companies have teams able to travel to the units, while others put their engineers directly in touch with the operators. Certain software or hardware fixes can be delivered within days. It is Darty — France’s household-appliance retailer — in uniform.

« It is after-sales service in uniform. »

Ukrainian manufacturers even affix QR codes to their drones. In case of a problem, the operator can get in touch directly with the manufacturer’s technical team. The reviews posted by units on the procurement platforms then make it possible to identify the reliable systems and those that do not withstand real conditions. Effective equipment receives more orders; the rest must evolve or disappear.

This permanent selection is merciless, but consistent with the speed of the competition. According to an American officer who spent seven months working with Ukrainian institutions, units and industry, local manufacturers routinely modify their software and hardware three or four times a year, and the brigades’ innovation cells make far more frequent adaptations. By comparison, the development of a major American military programme, even under so-called accelerated procedures, can take years. A study published by the Modern War Institute sums up the problem: in drone warfare, it is no longer enough to produce at scale — one must also innovate at the edge of the system and close the feedback loop very quickly.

Starting from the problem, not from the programme

The essential change therefore concerns the way the military need is defined. The traditional system generally starts from a forecast: the general staff imagines the threats of the next ten or twenty years, defines the necessary capabilities, then launches a programme intended to supply standardised equipment to the entire force. This approach remains indispensable for building a combat aircraft, a submarine, an air-defence system or a tank. But it becomes too slow when it comes to digital equipment, small drones, on-board computers, communications or electronic warfare. In these fields, the threat can evolve several times during the course of a single acquisition procedure.

The Ukrainian experience suggests partially inverting the logic. The starting point is no longer a central programme decided long in advance, but a concrete problem observed by the fighter: how to fly despite jamming? How to resupply a position without exposing a driver? How to evacuate a wounded man in an area watched by FPV drones? How to detect a fibre-optic-guided drone? How to attack an armoured vehicle hidden under a net or inside a building?

The unit formulates the problem, its technicians design or adapt a first response, the prototype is tried out in operation and the manufacturer then steps in to industrialise the solution. The sequence is no longer strictly linear. It becomes a loop:

• observation of a difficulty in combat;
• formulation of the need with the operators;
• manufacture or modification of a prototype;
• immediate trial in real conditions;
• technical correction and tactical adaptation;
• limited series production;
• distribution to other units and renewed evaluation.

The Ukrainian DOT-Chain Defence system already translates this logic into public procurement: units choose the equipment suited to their situation and the state organises its financing and delivery. The Brave1 Market, a digital platform created in 2023 by the Ukrainian state to bring together the armed forces, defence companies, investors and government agencies, also enables formations to acquire drones, ground robots or electronic-warfare systems from among thousands of listed products. This mechanism does not replace centralised procurement, but complements it with shorter circuits better suited to the front’s urgent needs.

Quantity matters as much as perfection

Speed, however, is not enough. An excellent isolated prototype does not change the balance of forces. One must be able to produce it by the thousands, define the doctrine and the operating procedures, train the operators, supply the spare parts and guarantee the provision of batteries, motors, antennas and electronic components.

Ukraine thus hoped to manufacture more than seven million drones in 2026 and to commit around 9,000 per day across the whole front. These figures give the measure of an industrial transformation: many small drones are now consumed like ammunition. The pursuit of the perfect system, costly and usable for twenty years, is being replaced in certain segments by a succession of inexpensive models, produced quickly and regularly modified.

This forces a distinction between two industrial rhythms. Heavy platforms — tanks, artillery pieces, missiles or engineering vehicles — require stable production lines, common standards and long-term maintenance capacities. They can be stored for years, then modernised. Digital equipment and drones, by contrast, must be able to evolve without every change of camera, frequency or software triggering a new procedure lasting several years. It is often pointless to stockpile them in large quantities, since they risk becoming obsolete before they are even used.

The challenge is therefore to produce at scale while retaining the ability to modify the product quickly. An army that only builds prototypes cannot sustain a war of attrition. An industry that produces millions of units of a model that has become vulnerable to jamming wastes its resources. Interfaces, batteries, payloads and command systems must be standardised, while allowing the components whose period of relevance is short to evolve rapidly.

Tanks are not disappearing, but their use is changing

The structure of CODE 9.2 also sheds light on the debate over the future of armour. The proliferation of drones has considerably increased its vulnerability. Concentrated columns, movements carried out without electronic cover and vehicles immobilised in open terrain become easy targets. The battlefield is observed almost permanently, sometimes to a depth of several dozen kilometres.

As early as April 2024, Nastenko described to the Washington Post a « death zone » in which any movement could be detected and struck. The concentration of drones, mines, artillery and sensors makes a mechanised breakthrough comparable to those of 2022 far more difficult. For him, however, this revolution did not mean that armour had become useless.

« A death zone in which any movement can be detected and struck. »

CODE 9.2 retains tanks and protected vehicles. Armour remains necessary to transport men, support the infantry, evacuate the wounded, open a passage or provide direct fire. But the tank can no longer be regarded as the autonomous centre of the manoeuvre. It must operate within an ensemble bringing together reconnaissance drones, electronic warfare, short-range air defence, artillery, engineers and dispersed infantry.

The question, then, is not one of choosing between the drone and the tank. In Nastenko’s model, the drones watch, isolate and weaken the position; the artillery and strike systems destroy the strongpoints; the ground robots can carry equipment, lay mines or evacuate the wounded; finally, the infantry and the armour occupy the ground. Technology reduces human exposure and prepares the assault, but it cannot yet hold a trench on its own, search a house or durably control a locality.

Rethinking the structure of armies

The lesson of CODE 9.2 ultimately goes beyond drone warfare. A contemporary army can no longer draw a sharp line between those who fight, those who design the equipment and those who buy it. Software, communications and electronic warfare are now part of the weapon itself. Being able to modify it quickly is becoming as important as knowing how to use it.

This does not mean turning every brigade into an industrial enterprise, nor sacrificing controls to speed. Mass production, certification, cybersecurity and compatibility with the rest of the forces still require central organisation. The recent affair of the Royal Navy’s K3 Scout naval drones was a reminder of why. A cybersecurity assessment revealed that Chinese-made components built into their cameras were automatically sending technical signals to an IP address located in China, apparently without the knowledge of the boats’ British manufacturer, Kraken Technology Group. The British Ministry of Defence states that no sensitive data or military system was compromised, but the incident shows the risks associated with complex and insufficiently controlled supply chains. Accelerating innovation therefore does not dispense with knowing the origin of components or auditing hardware and software. Subject to that proviso, engaged units must have organic technicians, decentralised budgets, prototyping resources and direct access to manufacturers. They must also be able to discard ineffective equipment quickly, without being forced to use it simply because it was ordered several years earlier.

On a battlefield where the adversary can identify a frequency, copy a drone or develop a countermeasure within weeks, time becomes a component of military power in its own right. A slightly less capable system, but one delivered immediately, repaired locally and improved after each engagement, can produce more effect than a technically perfect weapon that arrives too late. Risks must therefore be accepted — but identified, assessed and kept under control.

« Time becomes a component of military power in its own right. »

The main innovation of CODE 9.2 thus perhaps lies neither in its latest drone nor in the unusual distribution of its personnel between infantrymen, pilots and technicians. It lies in the gathering, within a single formation, of those who identify the problem, those who fight and those who know how to adapt the tool. In contemporary warfare, superiority increasingly belongs to the army capable of turning a lesson learned at the cost of combat into equipment available before the adversary has had time to adapt.