Ukraine is developing a significant language model using the Google Gemma family. The project will initially employ Google’s computational infrastructure for training before fully transitioning to a local setup, ensuring Ukraine maintains full control over AI systems accessed daily by 23 million citizens. “The Ukrainian large language model is expected to become the foundation for a new generation of AI-based services in both the public and private sectors of Ukraine,” states a Kyivstar announcement.
Ukraine aims to create its own artificial intelligence model instead of paying millions of dollars to foreign companies, as was previously noted by Deputy Minister of Digital Transformation, Alexander Bornyakov in a Reuters report.
Bornyakov revealed that the decision to avoid proprietary systems like OpenAI’s ChatGPT was intentional. He added that Chinese systems DeepSeek and Qwen were also rejected. Training will occur on secure graphic processors outside Ukraine provided by Google, with the completed models to be deployed in local data centers.
In recent discussions, officials highlighted how this development not only underscores innovation but aligns with Ukraine’s broader technological aspirations amid current geopolitical challenges. The completed AI system will initially support government applications and Kyivstar platforms, and later extend to the private sector, with particular emphasis on enhancing local industries including agriculture and logistics. Although Kyivstar wasn’t able to confirm the expected launch date, anticipation builds as the AI community eagerly awaits a transformation in technological infrastructure across Ukraine.
