Ex-DeepMind and Owkin scientists launch Bioptimus for universal AI foundation in biology

French startup Bioptimus has secured $35 million in seed funding to develop the first universal AI foundation model for biology. Led by Professor Jean-Philippe Vert and comprising a team of scientists with backgrounds in Google DeepMind and Owkin, Bioptimus aims to revolutionise biology by leveraging cutting-edge AI foundation model technologies. The funding round, led by Sofinnova Partners and supported by Bpifrance Large Venture and global technology venture capitalists, positions Bioptimus to connect various biological scales, from molecules to organisms, using generative AI. The goal is to drive scientific breakthroughs and accelerate innovation in biomedicine and beyond.

Bioptimus's world-class team, with expertise from Google DeepMind and Owkin, underlines the significance of applying foundation models and generative AI to biology. Professor Jean-Philippe Vert, the co-founder and CEO of Bioptimus, emphasised the potential impact of capturing the laws of biology through advanced algorithms trained on extensive biological and multimodal data across scales. The startup plans to create a holistic understanding of biology, critical for advancing biomedical and environmental science.

The company will benefit from Owkin's data generation capabilities and global access to patient data, sourced from leading academic hospitals worldwide. Additionally, Bioptimus will utilise a best-in-class, scalable, and secure computing environment provided by Amazon Web Services. By harnessing vast amounts of data from various scales and modalities, Bioptimus aims to establish strong differentiation from models trained solely on public datasets and a single data modality, enabling a comprehensive representation of biology's full diversity.