A new French AI startup has raised $41 million to develop fundamental AI models in biology.
Just as OpenAI’s ChatGPT took the world by storm as a super-smart generative AI tool capable of textual natural language conversations, Bioptimus takes that concept but transforms its models into downstream biological We provide application-specific training. It comes with its own set. There are many challenges given that the required clinical training data is sensitive and may not be publicly available.
Jean-Philippe Vert, co-founder and CEO of Bioptimus, said the company seeks to deepen our understanding of biology by learning how biology works from raw data, from molecules to whole organisms. says. This will allow scientists and researchers to simulate the biological world to “predict disease outcomes” and develop more effective treatments, he says. And Vert says the simulations made his technology look a little like the model underlying ChatGPT.
“Essentially, this is like GPT for biology, but instead of generating text, it simulates biology,” he said in a statement.
hotbed
France has emerged as something of a hotbed for AI startups, with generative AI companies across the country securing the lion’s share of funding as of last year. Huge funding rounds include Mistral AI’s $640 million tranche, Company H’s $220 million secured, and Hugging Face’s $235 million completed investment. All took place in the past 18 months.
Bioptimus, on the other hand, was only founded last year but quickly secured a $35 million seed round. In less than a year after its founding, the company has raised a total of $76 million, a testament not only to the current AI hype but also to the backgrounds of Bioptimus’ six co-founders. For example, Chief Technology Officer Rodolphe Genatton previously served as a senior researcher at Amazon and Google. Meanwhile, Vert is co-founder and CEO of Bioptimus and chief R&D officer of Owkin, a French unicorn and another AI-powered biotech startup with backers such as GV. there is.
This dual role suggests a bioptimus origin. Owkin leverages AI and machine learning to accelerate drug discovery and builds extensive partnerships with top biopharmaceutical companies. As part of this effort, Owkin has also amassed a large amount of multimodal patient data. This is what Bioptimus uses to train the underlying model.
Rather than creating a tangential unit within Owkin that focuses on the underlying model, it made more sense to create a separate entity.
“While building biology (fundamental models) is not part of Owkin’s roadmap, Owkin is keen to support and partner with companies like Bioptimus,” Vert explained in an interview with TechCrunch last year. did. “Training a very large (base model) requires significant resources in terms of data volume, computing power, and a wide range of data modalities that are easy to unlock as specific entities. As a ‘pure player’, Bioptimus is better set up to do this. ”
In the previous months, Bioptimus launched H-Optimus-0, an open source foundational model for pathology trained on millions of images, to help research and diagnose diseases such as cancer. I raised it. However, with $41 million in new cash in the bank, the company will strengthen its AI platform with more diverse data sources covering a broader range of therapeutic areas, as well as seek to forge further partnerships with the pharmaceutical and biotech sectors. It’s planned.
As part of this, the company is preparing to release new multimodal foundational models across the biological spectrum later this year that can drive development in areas such as medicine, biotechnology, and even cosmetics.
“Beyond medicine, this model will unlock limitless potential in many other industries and drive biological discovery in ways we are only beginning to imagine,” Burt said.
Bioptimus’ latest funding injection was led by US venture capital firm Cathay Innovation, with participation from Sofinnova Partners, Bpifrance, Andera Partners, Hitachi Ventures, Boom Capital Ventures, Pomifer Capital, Sunrise, and several angel investors .
