Key Developments:
Ukraine opens the Avengers Labs platform to eligible defense companies, providing access to a 5-million-frame battlefield dataset for training AI models
Ukraine’s Defense Forces deploy more than 70 AI- and computer vision-enabled systems, while over 200 Ukrainian companies develop AI-powered drones
Ukraine plans a Sovereign AI Factory, national AI infrastructure, and expanded defense-tech financing
Kyivstar and the Ministry of Digital Transformation test a Ukrainian AI language model designed to support national digital services, data sovereignty, and security-sensitive applications
DEFENSE & BATTLEFIELD AI
Ukraine Opens 5M-Frame Battlefield AI Dataset to Defense Companies
Ukraine's Defense Ministry opened access to Avengers Labs, a platform that lets vetted Ukrainian defense companies train AI models for unmanned systems on an annotated dataset of 5 million battlefield frames, most of it drawn from the DELTA battlefield management system. The dataset covers tanks, artillery, air defense systems, infantry, Shahed drones, and reconnaissance UAVs. A target-detection system trained on that data already runs inside DELTA, processing more than 100,000 drone video streams a month; in Vezha, a DELTA module, it identifies 70% of enemy equipment in video, taking 2.2 seconds per item, according to the ministry. Models trained on the dataset also support partially autonomous missions, including trajectory correction after an operator marks a target and independent target detection within a designated area. The first companies have signed licensing agreements, and partner countries can apply subject to security vetting.
Read more
Ukraine Deploys More Than 70 AI-Enabled Weapons Systems Across Defense Forces
Ukraine’s Defense Forces are using more than 70 AI- and computer vision-enabled systems to strike enemy targets, while more than 200 Ukrainian companies are developing AI-powered drones, according to the Defense Ministry. Brave1 Market currently lists 46 AI-related defense products, including target-recognition, optical stabilization, and terminal guidance modules, and more than 100 companies have gained access to Brave1 Dataroom to train and validate models on real-world defense datasets. The ministry said its long-term objective is for 100% of frontline drones to incorporate computer vision and AI, while keeping the final decision on target selection and engagement under human control.
Read more
GOVERNMENT & DIGITAL INFRASTRUCTURE
Ukraine Plans Sovereign AI Infrastructure and Expanded Defense Tech Investment by 2027
Ukraine’s government plans to establish a Sovereign AI Factory, a state-controlled computing infrastructure intended to reduce reliance on foreign cloud providers and AI models, along with a Ukrainian-language data corpus to train a national LLM. By 2027, government digital services are expected to migrate to domestic infrastructure, while AI is also planned for proactive public services and executive agencies. In defense technology, the government plans to direct half of €60 billion in EU-backed weapons procurement loans to Ukraine’s defense industry, while Brave1 aims to increase investment in defense companies from $220 million to $350 million, having financed more than 400 military technology projects.
Read more
Kyivstar Tests Ukrainian AI Model with Government, Targets Completion by End-2026
Kyivstar is testing a small language model developed with Ukraine's Ministry of Digital Transformation, intended as the foundation for Siaivo, the country's national LLM, which CEO Oleksandr Komarov said he hopes to finish by the end of 2026. The model, built on Google's Gemma 3, entered closed testing in June; the ministry expects a full test version by late December and public availability in January 2027. The model is being trained on a dedicated Ukrainian dataset prepared by the ministry and other state institutions, focusing on Ukrainian-language context, data sovereignty, and security-sensitive applications, including defense technologies. Kyivstar CEO Oleksandr Komarov also said a domestic model could reduce token-processing costs for large-scale public digital services such as Diia compared with relying on foreign AI systems.
Read more
