Farang Secures €1.5 Million Funding to Launch New AI Architecture

Farang, a Swedish AI Research Lab developing the next-generation of foundational Large Language Models, led by Emil Romanus, has secured €1.5 million in funding led by Voima Ventures and the Amadeus APEX Technology Fund, with participation from notable angel investors including Tero Ojanpera (Co-founder, Silo AI), Nilay Oza, and Niraj Aswani (Former Founders, Klevu).

Farang is developing a new AI architecture that departs from the Transformer model used by ChatGPT, Claude, and Gemini. Inspired by how the human brain processes information, it forms complete responses conceptually before generating them, delivering greater coherence while using 25x fewer computational resources and reducing training costs significantly.

The funding will support scaling proof-of-concept models and securing compute resources for training and fine-tuning in specialized areas, starting with AI assistants for specific programming languages like React, specialized medical applications, and privacy-focused internal company tools. The technology enables industries such as healthcare, legal, and financial services to deploy AI models entirely on-premise, ensuring complete data sovereignty and eliminating the need to share sensitive information with external providers.