Key Developments:
Ukraine’s national LLM enters closed beta testing under the Ministry of Digital Transformation and Kyivstar development
Lapathoniia adds Model Context Protocol integrations with Prozorro and the State Statistics Service
De Novo reports stable beta testing for Lapathoniia, with peak loads reaching 61 million tokens per day
More than 100 Ukrainian companies train AI models on combat data through Brave1 Dataroom
Ukraine plans a unified AI-based battlefield operating system for decision support across the front line
YARTURA develops DANCER 4.5.0, an AI-enabled drone interceptor targeting UAVs at speeds up to 450 km/h
IT Ukraine Association publishes Ukraine AI Industry Map, forecasting market growth to $3.22B by 2031
Serpstat launches Stus.ai — an AI platform automating SEO research and reporting workflows
GOVERNMENT & DIGITAL SERVICES
Ukraine’s National LLM Enters Closed Beta Testing
Ukraine’s national large language model, developed by the Ministry of Digital Transformation and Kyivstar, has entered closed beta testing after pre-training and supervised fine-tuning. The current version is a smaller experimental prototype to test the architecture, training methods, and Ukrainian-language performance before scaling to a full national LLM. Scientific committees in history, linguistics, ethics, law, and technology will evaluate the model’s quality, safety, stability, and response accuracy. The project name, Siaivo, was selected by a public vote on Diia; the model is being developed to support future AI products for government and business, and using Ukrainian data and cultural context is key to relevance and applicability.
Read more
Lapathoniia Adds Prozorro and State Statistics Service Integrations
Ukrainian AI infrastructure platform Lapathoniia released a major update adding Model Context Protocol connectors to Lapathoniia Chat. This enables language models to access external data tools without separate integration code. The first two connectors give models direct access to Prozorro, Ukraine’s public procurement system, for tender search and analysis by organization name or company registration code. They also connect to the State Statistics Service of Ukraine for public datasets and machine-readable statistical data. The platform introduced Rada, a multi-model workflow in which LapaLLM serves as an operational agent for search and tool use, while MamayLM performs deeper analysis and produces the final response. Lapathoniia is in closed beta and runs on cloud infrastructure from Ukrainian provider De Novo.
Read more
De Novo Reports Stable Beta Testing for Ukrainian LLM Platform Lapathoniia
Ukrainian cloud provider De Novo reported stable beta-test performance for Lapathoniia, a platform that hosts and integrates the domestic language models LapaLLM and MamayLM. The platform handled heavy user demand after Lapathoniia Chat’s launch, reaching peak loads of 61 million tokens per day and over 11,000 requests. GPU utilization approached 80% without major infrastructure failures. Testing evaluated Ukrainian-language comprehension, Ukrainian-English translation, summarization, instruction following, and school-level exam tasks. De Novo noted that further work is needed to address complex edge cases. For enterprise users and developers, the results position Lapathoniia as an emerging domestic LLM infrastructure layer, though it remains in beta and heavily depends on continued improvements in model quality.
Read more
DEFENSE & SECURITY
Over 100 Ukrainian Companies Train AI Models on Combat Data
More than 100 Ukrainian companies are training AI models using real combat data through the Brave1 Dataroom platform, according to Ukraine’s Ministry of Defense. The platform provides access to standard and thermal imagery of aerial targets collected during combat operations, including data on various drone types across day, night, and varied weather conditions. The main focus is on improving systems for detecting and intercepting Shahed-type drones. Real-world data is expected to improve target recognition accuracy, reduce reaction time, and support the development of autonomous interceptor drones. Ukraine has also opened access for foreign partners to train AI models for unmanned systems using battlefield data, expanding the international role of its defense AI ecosystem.
Read more
Ukraine Plans Unified AI Network for Battlefield Decision Support
Ukraine’s Ministry of Defense plans to develop a unified AI-based operating system to connect weapons, intelligence, and analytics tools across the battlefield, according to Defense AI Center head Danylo Tsvok. The system will analyze data along the 1,200-kilometer front line and provide commanders with decision-support recommendations. This reflects a shift toward what officials call a future “war of operating systems.” The Defense AI Center, launched in March, is part of a broader defense digitalization effort that also includes AI tools for personnel management and recruitment. Ukraine says it still requires human involvement in final battlefield decisions, while using projects like Brave1 Dataroom to share combat data expertise with allied countries and technology partners.
Read more
YARTURA Develops AI-Enabled Drone Interceptor for High-Speed Aerial Targets
Ukrainian company YARTURA developed DANCER 4.5.0, a fixed-wing drone interceptor designed to target enemy UAVs at speeds up to 450 km/h. The system includes an AI-based automatic target tracking module that can re-engage a target if the first approach fails. This addresses the growing speed of kamikaze drones used against Ukrainian cities and critical infrastructure. The interceptor can operate against reconnaissance, strike, relay, and small tactical drones, with engagement parameters up to 4.8 km in altitude and 30 km in range. For defense investors and partners, the product reflects rising demand in Ukraine’s air-defense market, where interceptor drones have reportedly destroyed more than 33,000 enemy UAVs.
Read more
MARKET INTELLIGENCE
IT Ukraine Association Publishes Ukraine AI Industry Map
IT Ukraine Association presented the Ukraine AI Industry Map, the first comprehensive study that structures Ukraine’s AI market for an international audience; the Map was developed in partnership with analytical partner Mind.ua and compliance partner Velstadt. The report estimates Ukraine’s AI market at $573.7 million in 2025, with potential growth to $3.22 billion by 2031 — implying a 5.6x expansion and a 33.3% CAGR. The study also includes a catalog of 64 verified Ukrainian AI companies across product startups, AI/ML/Data Science service firms, corporate R&D units, and international R&D centers in Ukraine. For investors, the report highlights potential ROI of up to 400% in DefenceTech and 310% in HealthTech, and identifies DefenceTech, AgriTech, EdTech, and HealthTech as key AI verticals.
Read more
PRODUCT & PLATFORMS
Serpstat Launches Stus.ai to Automate SEO Research Workflows
Ukrainian SEO platform Serpstat launched Stus.ai, an AI product that turns SEO research tasks into ready-to-use files, reports, audits, and technical briefs. The platform uses Serpstat’s own data layer, including search positions, traffic, keywords, backlinks, and competitor dynamics, rather than relying only on general model knowledge. Stus.ai automates fixed SEO workflows, such as keyword clustering, site audits, quick-win detection, content gap analysis, keyword cannibalization checks, CSV reports, and HTML audits. It integrates with Google Search Console, Google Sheets, and Lighthouse. According to early user data cited by the company, the product can reduce routine SEO work by 4–6x, thus targeting SEO specialists, agencies, SaaS founders, and marketing teams without dedicated SEO departments.
Read more
