VEON Ltd. (NASDAQ:VEON), a global digital operator, announces today that Kyivstar, VEON's digital operator in Ukraine, has joined forces with Ukraine's Ministry of Digital Transformation to develop Ukraine's first national large language model ("LLM") trained on Ukrainian-language data. The cooperation memorandum builds on VEON's and Kyivstar's pledge to invest USD 1 billion to rebuild Ukraine's digital infrastructure over 2023-2027.
The parties plan to develop Ukrainian LLM using open-source architectures and trained exclusively on Ukrainian data sources. In addition to capturing the full range of Ukrainian dialects, terminology, history, and cultural context, the project is designed to ensure that sensitive national data is securely stored and processed within Ukraine, a critical requirement for sectors such as government, defense, healthcare, and financial services.
The LLM will also serve as the foundation for AI-powered products and services, including legal and regulatory analysis tools, as well as specialized AI applications across both public and private sectors including in fields like education, finance, and health. Trained on Ukrainian data, the model is expected to deliver more accurate, relevant, and actionable insights for Ukrainian users than general-purpose global models.
Kyivstar will develop the LLM in collaboration with Ukraine's Ministry of Digital Transformation and the WINWIN AI Centre of Excellence. The first version of the Ukrainian LLM is scheduled for release by December 2025.