Decoding DNA with AI: Living Models emerges from stealth with $7M

Paris-Berkeley startup Living Models has emerged from stealth, securing $7 million in seed funding to develop foundation models for biology trained on extensive DNA, RNA, and multi-omics datasets. This significant investment underscores the burgeoning potential of artificial intelligence in revolutionizing biological research and agricultural innovation by enhancing our understanding of complex living systems. The company's ambition is to decode the fundamental "language of life" through advanced AI techniques, mirroring the success of large language models in human communication. To fuel its ambitious development, Living Models has also gained access to a substantial computing cluster featuring 120 NVIDIA B200 GPUs, crucial for training its next generation of sophisticated biological AI models. These transformer models are designed to analyze intricate patterns within living organisms by processing vast amounts of genomic and transcriptomic data. While AI has already transformed numerous sectors, its application in agriculture, particularly in developing climate-resilient crops, remains a frontier with immense growth potential. Living Models' initial focus is on plant science, aiming to accelerate the development of new crop varieties that can withstand pressing climate challenges. Their flagship offering, BOTANIC, is a family of transformer models trained on diverse plant genomic sequences, capable of identifying genetic markers linked to traits like climate resilience and disease resistance. This technology promises to significantly shorten the traditional, multi-year crop breeding cycles by enabling researchers and agricultural companies to computationally predict and prioritize promising genetic variants for testing, thus fostering European digital sovereignty in a critical sector.