Ukraine is celebrated as a bottom-up laboratory for military innovation. The popular narrative casts its defence-tech scene as a landscape of decentralised, but inspired, chaos, where open-source coders develop battlefield analytics and those who like to tinker assemble FPV drones in makeshift frontline workshops. It’s fast, flat, and improvisational.
But within the Ministry of Defence, a new model is emerging for AI development. As the war with Russia evolves, Kyiv is adding centralised mechanisms to improve and scale the country’s grassroots innovations. The goal is to turn experimentation into enduring operational capability — and position Ukraine as a global leader in military AI.
At the centre of this shift, Ukraine has established a new unit to bring the activity together: the A1 Defence AI Centre.
First announced in March with £500,000 in backing from the UK, A1 aims to create and join up AI systems across Ukraine’s armed forces. The centre would act as the government’s “in-house developer of technical products for the defence sector,” CEO Danylo Tsvok told Resilience Media in an interview.
By bringing AI development inside the Ministry of Defence, A1 wants to create common data standards, coordinated testing pipelines, and faster deployment cycles. Its ultimate ambition is to build the “most effective defence system in Europe.”
You can see the logic of it.
Flat, fast and improvisational can also mean disorganisation and bad decisions going unchecked, so A1 could give the country a better shot of taking the systems, applications and hardware getting the best results and using them more widely. Then, if A1 wants to build more AI and other defence products, bringing everything together would give it a valuable amount of data to do that. That also makes Ukraine a more lucrative partner to companies and other countries.
But the big question is will A1 fly as the government hopes.
Given the extent of the country’s hacker-style ethos, getting A1 off the ground — even with the endorsement of the country’s defence ministry — may be easier said than done.
At worst, will adding bureaucratic oversight run the risk of slowing down the pace that’s given Ukraine advantages on the battlefield up to now?
For the moment, the belief is that A1 will ultimately give it more speed, not less.
“In a war of intelligent systems, the advantage goes to the side whose technology can respond faster than the enemy’s,” Tsvok said. “Ukraine is investing in artificial intelligence to anticipate developments and stay several steps ahead of the adversary.”
One of A1’s aims is to turn fragmented experimentation into a more cohesive national AI capability.
That will include bridging two pillars of Ukrainian military innovation: Brave1, a government-backed defence-tech accelerator; and DELTA, a battlefield management system.
“A1 appears to sit between these two layers,” said Federico Borsari, a defence-tech expert at the Centre for European Policy Analysis (CEPA), “by trying to turn AI into an organised defence capability via an institutionalised avenue.”
He sees the initiative as part of an effort to bring a more systematic approach to innovation in the country’s defence ecosystem.
“To me, A1 appears as an attempt to formalise, centralise, and scale AI workflows that were already emerging through Brave1, individual developers, drone units,” he said.
Tsvok, meanwhile, described A1, Brave1 and DELTA as part of a “unified state defence innovation ecosystem,” with A1 serving as “a hub of AI expertise.”
“The shift A1 introduces is a transition to a full-cycle innovation model, where the path from idea to frontline deployment is minimised,” Tsvok said. “To achieve this, the centre will develop simulation environments that allow algorithms to be tested on real battlefield data before being deployed in the field.”
The international connection
A1 is shaped by international collaboration, specifically with backing from the UK. Tsvok described the partnership as “technology driven and strategically aligned.”
At the heart of Ukraine’s cooperation with others is data. Over more than four years under Russia’s full-scale invasion, the country has become a live testing ground for military AI systems, generating large-scale operational datasets under real combat conditions. International companies and organisations seek inroads into Ukraine not just to generate business, but also to gather their own live data.
Kyiv knows the strategic value of that position. That value is now pushing the government to formalise how AI systems are developed inside the state.
“The goal is to enable Ukrainian companies and strategic partners to effectively train models and develop solutions that are ready to meet the challenges of modern warfare,” Tsvok says.
Fragmented, existing AI initiatives
Borsari believes the initiative could also strengthen interoperability across Ukraine’s growing military AI ecosystem. “Where A1 can bring novel services is data governance. One of their goals is to establish unified data standards,” he said.
Indeed, A1 is not the first time that the government has tried to streamline and centralise its various military AI efforts.
Some of those other AI efforts are being rolled into the new operation. In March, Kyiv launched a new platform for allies to train autonomous systems on real combat data: A1 takes that model a step further by bringing AI development directly into the Ministry of Defence.
Yet other AI initiatives launched by the government are not necessarily being rolled in, and some appear to have stalled.
For example, in July 2025, the government announced K4, an accelerator for military AI startups. Supported by the German government, the program offered $250,000 grants to four winning teams. The winners were never announced, and the official K4 site does not indicate if the project is still active or if it may align with A1 going forward.
It’s also unclear if other national AI initiatives will fit into A1. For instance, in December 2025, Ukraine’s Ministry of Digital Transformation and telecoms carrier Kyivstar said they were developing a new “National Large Language Model” built on Google’s Gemma framework. The model would be hosted initially by Google and then moved to Ukraine-based infrastructure. The government said the first beta of the LLM would be released for testing in the spring of this year.
Thirty-five year-old Tsvok has spent most of his career in the public sector, working with both government and commercial end users. He also worked for several years in blockchain technology that then evolved into an AI focus (as have many others in crypto). A1 was mapped out initially at the WINWIN AI Centre of Excellence, which Tsvok previously ran at the same time that he was Chief AI Officer at Ukraine’s Ministry of Digital Transformation, working to get more of the industry to use AI and other new technologies.
That latter background supports A1’s remit beyond frontline systems. Tsvok says the centre aims to promote the use of AI to streamline bureaucracy, functioning as a “copilot” in workflows, automating audits, and improving procurement.
In effect, A1 is attempting to institutionalise AI in both combat systems and the civilian machinery that supports them. The aim is to evolve Ukraine’s wartime innovation model into a more durable state structure.
In the end, A1 doesn’t intend to completely replace Ukraine’s decentralised innovation model. At its most successful, it will add a layer designed to standardise, validate, and scale what emerges from the frontline.
If A1 can pull it off, the frontline workshop will merge with the machinery of the state.
