Ex-Google, self-driving engineers’ Atlas Bio raises $12.5M run drug trials as AI simulations
Tech Funding News Sofia Chesnokova ● Covered by 2 sources
Atlas Bio just raised $12.5M to use AI to simulate drug trials. It wants to spot failures before Phase II burns years and cash.
Based on reporting by Tech Funding News, Sofia Chesnokova — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Atlas Bio has raised $12.5 million in a round led by Creandum to build AI simulations of drug trials. The Swiss startup says the goal is simple: help drugmakers see which patients respond before they lock in a Phase II design, where roughly 70% of candidates fail.
The company was founded in 2024 by Simon Romanski, who was among the first engineers working on autonomous vehicles in 2014. He says the move from self-driving cars to drug development wasn’t a retreat from one field to another so much as a chance to apply the same engineering mindset to biology. Atlas Bio is headquartered in Basel, incorporated in Delaware, and has a Swiss subsidiary.
Romanski is joined by Sebastian Bayer, who worked on perception models for autonomous driving and later led Google research on language models for Search, YouTube, and Google Play, and by chief scientific officer Dominik Hartl, who spent more than 10 years in drug development after a doctoral fellowship at the Friedrich Miescher Institute for Biomedical Research. Atlas Bio says its foundation model is trained on about 2 million human RNA readouts, which lets it model gene activity in human tissue and simulate responses across diseases and treatments.
The company’s pitch is that this can make Phase II less of a blind gamble. Atlas Bio says a Phase II trial can cost at least $100 million and take two to three years, and it argues that the industry still leans too heavily on mouse data and huge patient readouts that arrive late. Its virtual trials let sponsors test multiple treatments on the same patient profile at the same disease stage, which is impossible in a real trial.
Atlas Bio says it’s already working with a top-10 pharma company and is talking to several others. It also says a Basel-based biotech client skipped animal studies for one programme and went straight to human trials using its simulations. The company expects its first trial results about six weeks after the October 27 announcement, and that matters more than the fundraising headline. If the results hold up, the business gets real. If regulators won’t accept simulations in place of a Phase II readout, it stays a very clever advisor.
My take — AI-written commentary, not fact-checked reporting
TLDRocket usually likes software that attacks the mess before it becomes a billion-dollar disaster, and this is exactly that kind of bet. But the real hurdle isn’t model quality; it’s whether regulators let a simulation stand where a trial used to sit. Biology doesn’t care about startup ambition, which is rude but useful.
Read more about this at: Tech Funding News