Kiefer develops Greek-language LLM to enhance national AI capabilities

In recent months, Kiefer, a company focused on artificial intelligence, has initiated a significant project aimed at developing a Greek-language large language model (LLM) to address the specific linguistic and cultural needs of Greece. This initiative comes at a time when countries around the world are prioritizing AI autonomy, particularly as Nvidia has reached a substantial market value of $4 trillion, highlighting the urgency for nations like Greece to strengthen their AI infrastructures. Kiefer's project is expected to facilitate more effective communication, information retrieval, and data analysis in Greek, thus catering particularly to the local population of approximately 10.5 million people. The project team, led by technologist Christos, is motivated by the desire to establish a stronghold in the AI market by tailoring solutions to local requirements. One of the prominent challenges identified by Christos is the overwhelming volume of data available in various formats, which necessitates extensive curation to extract usable information for training the model. The current estimated buy-in for AI adoption in Greece is roughly 9%, with even lower engagement observed in government and public sectors, indicating a need for increasing awareness and investment in AI technologies. Underpinning the development of the Greek-language LLM will be a combination of various technologies, including models from DeepSeek which were chosen for their accuracy and open-source nature. Other models under consideration include Nemotron and a demo from Google, reflecting Kiefer’s commitment to leveraging the best technologies available. Although services are currently offered free of charge, a subscription pricing model is anticipated in the future to support costs associated with Nvidia hardware utilization and other operational expenses. This national initiative to develop AI capabilities represents a shift towards greater control over data and intellectual property, particularly as the landscape for AI training data continues to evolve. Kiefer plans to build its own data centers and utilize a range of CPUs and GPUs to establish a sustainable infrastructure. By integrating this advanced model with local universities and participants, Kiefer aims to enhance the academic and technological landscape in Greece, ultimately contributing to the establishment of sovereign AI capabilities that equip the nation for future technological advancements.