Zurich-based Atlas Bio has emerged from stealth with $12.5 million to take on one of drug development's most expensive bottlenecks: Phase II clinical trials. Instead of recruiting patients and waiting years, the company wants to run trials on virtual patients simulated by an AI foundation model.

The funding, announced on October 27, 2026, is a $10 million seed round led by Creandum, the Swedish venture firm that backed Spotify and Trade Republic early. A previously unannounced pre-seed led by First Momentum Ventures and LEA Partners brings the total raised to $12.5 million. Atlas Bio was founded in 2024 by Simon Romanski and Sebastian Bayer, who previously led large language model research at Google, and the company says it is already working with a top-10 global pharmaceutical firm and in talks with several others.

How a virtual trial works

Atlas Bio's model is trained on about two million human RNA readouts from human tissue — the molecular signals showing how genes switch on and off. By learning those patterns, it predicts how a drug behaves across a virtual cohort of patients, simulating efficacy and safety before a real Phase II trial begins.

The pitch is speed and breadth. A virtual trial takes 12 weeks, the company says, and it can evaluate a drug across multiple diseases at the same time — something not feasible in a conventional clinical trial. Results from a first trial are expected within six weeks, which will be the first public test of whether the simulations hold up.

A researcher examining a glowing test tube in a laboratory
Atlas Bio's model is trained on about two million human RNA readouts. Illustration: AI Frontier Post.

Why Phase II is the target

Phase II is where most drug candidates die. A trial at this stage costs more than $100 million and takes two to three years, yet about 70% of drugs fail there, according to the company. Cutting that stage down to a 12-week simulation would change the economics of drug development — if regulators and pharma partners trust the results.

The regulatory wind is at Atlas Bio's back, at least in principle. The FDA Modernization Act 2.0, passed in 2022, removed the legal requirement for animal testing before human trials, and in March 2026 the US Food and Drug Administration issued draft guidance on how developers can validate computer models, organoids, and organ-on-chip systems for drug development. Acceptance, however, remains evidence-dependent, and animal testing has not been broadly abolished — so virtual trials still have to earn their credibility case by case.

Flowing ribbons of glowing RNA sequencing data
RNA readouts show how genes switch on and off — the signals Atlas Bio's model learns from. Illustration: AI Frontier Post.

What to watch

The six-week countdown is the story now. If the first trial's results land on time and convince a top-10 pharma partner, virtual trials move from pitch to precedent. If they slip or stay vague, Atlas Bio joins a long line of AI drug-discovery startups whose simulations looked better than their clinical record. Either way, the $12.5 million buys a front-row test of whether AI can simulate patients well enough to skip the most expensive step in medicine. Reuters reported the funding on October 27, 2026.