Atlas Bio, a Swiss biotech, left stealth on October 27 with $12.5M to test experimental medicines on simulated patients before any of them reach a hospital. The money splits into a $10M seed round led by Creandum and a pre-seed of undisclosed size from First Momentum Ventures and LEA Partners.
What it sells is a model for running clinical trials inside a computer. Fed roughly two million RNA readouts lifted from human tissue, the genetic switches that decide which proteins a cell makes, it is meant to forecast how a compound will act in one specific person rather than in a statistical average, so sponsors see trouble early.
Why sponsors are listening
Mid-stage human trials cost over $100M and take two to three years, and about 70% of candidates still collapse there. Atlas contends that a large share of those losses trace back to how a study was drawn up, not to chemistry that never worked.
Its argument runs this way: only 30% of drugs move the needle on the condition they were tested for, yet among the survivors that win approval, some 60% turn out to help most against an entirely different illness. Screening a molecule across many diseases at once, the founders say, catches those signals while a narrow protocol can still be changed.
What would have to be proven
A second promise is a path around animal studies, which regulators on both sides of the Atlantic are revisiting. The obstacle is evidence: showing a synthetic cohort tracks what happens in real people, and no sponsor has yet published a prospective study built on one.
Atlas has not named partners, released validation results or set a date for a first regulatory filing. Creandum partner Katharina Wilhelm took a board seat.
