The Pulse
Rivercell Raises $25 Million to Build an AI Model of Cell Responses
Paris-based Rivercell has raised a $25 million seed round led by HV to expand its biological data platform and laboratory. The company aims to train an AI model to predict how human cells respond to drugs and genetic changes, but says the d
AI.info Team · Written with AI ·
Rivercell says it wants to predict how human cells respond to drugs, while acknowledging that the data needed to train such a model do not yet exist at the required scale. The Paris-based biotechnology company announced a $25 million seed round on October 7, 2026, to build that data foundation and develop what it calls an AI Virtual Cell.
HV led the round; HCVC, Alven and Bpifrance Digital Venture also invested. Rivercell plans to use the money to expand its data-generation platform, enlarge its Paris wet lab and launch its AI Virtual Cell program. The release does not report model benchmarks or a date for a working system.
Rivercell’s model depends on data it says is missing
The company describes its proposed model as a “world model” designed to predict how cells change after a drug or genetic intervention. Its aim is to learn responses across different biological settings rather than build a system for one disease alone. Rivercell says existing laboratory instruments were not designed to gather the breadth of measurements needed, so it is developing its own experimental platform.
“A world model of the cell is the opportunity of the century in medicine. The missing piece is data: how cells change over time and under treatment, seen through multiple lenses and at multiple layers, captured at scale,” said co-founder and chief executive Yann Fleureau. He added that building the data engine in Europe is strategic for both Rivercell and the continent’s role in AI and biology.
The approach Rivercell describes combines interventions with repeated observations of the same cells and multiple kinds of measurements. The company’s manifesto says its platform is intended to link live-cell imaging with gene-expression measurements, follow cells over time, and test interventions across cell types, genetic backgrounds and environments. The model is also intended to help choose which experiment to run next.
A research ambition, not a demonstrated predictor
That plan sets a high bar: the model must learn cellular responses from experiments and make useful predictions beyond the conditions it has already seen. Rivercell’s own account says current datasets often capture a single measurement or time point and cover too few biological contexts. Its manifesto also says that measurements leave parts of a cell’s underlying state, including aspects of protein activity and metabolism, unobserved.
The company presents the resulting model as a way to reduce the need for physical laboratory experiments and eventually change how drug discovery works. Those are goals, not demonstrated outcomes in the launch announcement: it provides no performance results showing that Rivercell’s model predicts cellular responses or improves drug development.
Founders bring experience in diagnostics and drug research
Fleureau started Rivercell in summer 2025 after co-founding AI cardiac-diagnostics company Cardiologs, which Philips acquired in 2021. Eric Durand joined in 2026 as co-founder and chief scientific officer. The release says Durand previously led oncology data science at Novartis, served as chief data science officer at Owkin and co-founded biology-model company Bioptimus.
Rivercell says Dr. Charlotte Bunne of EPFL and Prof. Fabian Theis of Helmholtz Munich have joined its scientific advisory board. Its immediate task is more concrete than the promise of a virtual cell: build the laboratory and data system that could give such a model its training material.