French AI startup Bioptimus, which seeks to develop a large language model (LLM) for biotechnology, has secured a $35m seed round. This comes just a few months after the company’s inception. Bioptimus sprung from the French unicorn Owkin, which leverages AI to better understand and elevate the impacts of medical treatments on various patient clusters. Owkin’s current Chief R&D Officer and former Google researcher, Jean-Philippe Vert, is one of Bioptimus’ founders. However, Vert’s role at Bioptimus will not be operational, and he will maintain his position at Owkin.

Sofinnova Partners, a French VC specializing in life sciences, and the public bank Bpifrance Large Ventures led the seed round. French VCs Frst and Cathay Innovation, US investor Headline, Owkin, and renowned entrepreneur Xavier Niel also took part in the funding initiative.

The technology behind language model underpins ChatGPT, which has the capability to produce human language, images, sounds, and software. It involves training AI models on extensive datasets. Bioptimus aims to train its model on data crucial for decoding biological procedures at various scales. The model Bioptimus is developing is termed a “multiscale model” as it encompasses diverse scales of biology, such as organs, tissues, cells, molecules, and atoms.

AI models have traditionally fallen short in capturing the full intricacy of organisms as they haven’t been trained on data representing different biological levels. As a result, simulations of biological processes, like an organism’s response to a medicine, have not been as precise as they could potentially be.

By utilizing foundation models and sophisticated algorithms trained on an immense volume of biological and multimodal data over scales, Bioptimus aims to decipher the laws of biology that have otherwise proven too complicated to be properly grasped, as stated by Vert. He believes that Bioptimus has a significant edge in the field, seeing that it will be able to train its model using Owkin’s vast pool of high-quality patient data in an industry where data access is stringently controlled.

Although a $35m seed round for a relatively new startup may appear considerable, such funding has become typical for startups devising LLMs, which require substantial computational power and top-tier talents. Last year, another French AI startup, Mistral, raised a $105m seed round shortly after its launch. Meanwhile, Poolside, working on an LLM capable of writing code, secured a $126m initial round.

Source: An LLM for biotech: French startup Bioptimus raises $35m seed.