Ukraine Military History Institute
One change sits under the whole list
Andrii Hrytseniuk, who runs Brave1, frames the remaining gaps as a list of problems the market will not solve on its own. He walks through it domain by domain, starting with the air.
It reads like four separate shopping items: deep and middle strike, counter-UAS, radio links, artificial intelligence. Read in order, though, most of the list follows from the first item. Ukraine has pushed the distance at which it fights much deeper, and everything after that is the bill for doing so.
His headline figure is that over the last half year Ukraine extended the gray zone from 10 to almost 70 kilometres on Russian territory.
The depth was bought on purpose
The starting point he gives is modest. At the end of 2025 Ukraine had just a couple of hundred hits at distances of more than 20 kilometres.
What changed was a decision to focus on that band. The ePoints system was used to prioritise targets beyond 20 kilometres. Then, about one and a half months before the recording, a new segment was added: 70-plus kilometres.
The sequence is the point. Range was not left to arrive by itself as drones improved. It was rewarded and steered. Hrytseniuk ties this to "a military strategy on how to stop and win in the war" and says the technology is growing according to that plan.
The defensive problem is speed, and the answer is cost
The other half of the air domain is counter-UAS, and the problem he names is specific. The share of high-speed kamikaze drones coming from Russia is increasing.
He names the Shahed and the Banderol alongside other jet kamikaze types. His conclusion is that the drone interceptors themselves need rebuilding: jet and other high-speed drone interceptors, plus missiles and other countermeasures, and new products in the counter-UAS segment.
The economic argument is the one he leans on. Drone interceptors cost much less than the Shahed or the jet kamikaze they are meant to stop. He gives no figure, only the direction of the ratio, so the claim is about the relationship between costs rather than a price.
He reports that the development is funded, that the programme has been running for about four months, and that there are already working solutions with what he calls very, very good results. That is a progress report from the funder, not an independent evaluation, and it should be read as one.
Depth breaks the radio link
This is the most concrete consequence of the deeper fight, and Hrytseniuk states the causation himself. Because the gray zone is extending, data links and radio technologies have to improve.
Two years ago, a radio link that worked at 30 kilometres was fine. Now, he says, it absolutely cannot be used on the front line.
Set that against his own numbers. If targets are being prioritised at 70-plus kilometres, a link that reaches 30 does not cover the band that is now being rewarded. His answer is high-distance radio links and mesh networks, which he names as the category for the coming period.
It is the least glamorous item on a list that also includes jet interceptors and AI. It is also the one most directly tied to the depth figure. Range gained in the airframe is of little use if the control signal cannot follow it.
The pilot is becoming a supervisor
The last item is artificial intelligence, and his rationale is that this war is increasingly drone versus drone, not human against human or human against drone. Where technology fights technology, he argues, automation is needed.
He describes the path in stages. Ukraine moved from manually operated drones to human-in-the-loop, where routine tasks are automated and the AI advises on better decisions. The next stage, already started, is human-on-the-loop.
His definition of that stage is the striking part. It means there is no longer a need to be an extremely effective and skilled pilot. The skills, he says, are being transferred to the AI, because computer vision sees better than a human, AI decides faster, and a computer handles large amounts of data better.
What remains for the human is to monitor, help, and confirm or cancel wrong decisions. He keeps the reason in plain view: the human stays because AI is not perfect yet.
His forecast, and the discomfort he admits
Hrytseniuk's prediction is that next year, on both sides and across the defence industry, fully automated drones will be in wide use. He says it is frightening when drones take decisions themselves, but calls it the new reality.
His prescription is regulatory rather than technical. Legislation, he says, needs to be adopted that acknowledges this cannot be avoided, with drones fighting drones in fully automatic mode.
Read alongside his description of human-on-the-loop, the direction is consistent. Each stage removes a little more of the human from the decision, and the law is framed as something that has to catch up rather than something that decides whether the next stage happens.
What to watch
- Reported hits beyond 70 kilometres. He gave a baseline of a couple of hundred hits beyond 20 kilometres at the end of 2025. Whether the 70-plus segment produces a comparable reported volume is the test of whether the gray zone keeps moving.
- Jet interceptor results against Shahed and Banderol class targets. The programme was four months old with strong results claimed by the funder. Independent confirmation, or its absence, will settle how far that claim goes.
- Fully automated drones in wide use next year, as he predicts, and whether Ukrainian legislation on autonomy is adopted before or after those systems are fielded.