Rivercell’s seed round underwrites a clear thesis: in AI-enabled drug discovery, defensibility is shifting from algorithms to proprietary biological data and the infrastructure that produces it.
Paris-based Rivercell has raised EUR 25 million in seed funding from HV, with participation from HCVC, Alven and Bpifrance Digital Venture. The funding was recently announced. Terms beyond the headline amount were not disclosed.
Why this round matters
A growing share of early-stage “AI biotech” financings are converging on the same constraint: models improve only as fast as the training data, and public or broadly licensed datasets tend to be non-exclusive and uneven in quality. Rivercell is explicitly betting on owning the data creation stack.
The company is building a proprietary data engine and an automated wet lab in Paris designed to generate time-resolved, multimodal single-cell data at scale. That dataset is intended to train Rivercell’s AI Virtual Cell models, which aim to predict cellular responses in silico and reduce the need for physical experiments.
In other words, the core asset is not only software. It is a data-flywheel that links experimentation, proprietary datasets and model performance.
Investor syndicate signals cross-border appetite
The syndicate combines French venture capital and public innovation capital with international participation. HV led the round, joined by HCVC, Alven and Bpifrance Digital Venture. For the French ecosystem, this is another data point that local and international capital continues to back platform-style biotech infrastructure plays, not just single-asset therapeutics.
Team pedigree and the “institutional bridge”
Rivercell’s founding team includes former leadership from Novartis and Owkin AI research. That mix matters for execution risk. Building automated lab infrastructure, generating reproducible datasets and translating model outputs into drug discovery workflows requires both pharma-grade process discipline and modern ML research capability.
The background also reflects a broader European pattern: new biotech formation increasingly pulls from large pharma and established AI-first life sciences companies, rather than purely academic spin-outs.
Key execution questions
The strategy is coherent, but the operating plan is capital- and integration-heavy at company formation. Key questions for investors and future partners include:
- Data quality and standardisation: Can Rivercell consistently generate high-throughput single-cell data with the signal-to-noise ratio needed for robust model training across experiments and time?
- Throughput economics: Automated wet labs can become strategic assets, but only if utilisation, uptime and consumables costs support sustained iteration.
- Model-to-lab feedback loop: The company’s stated goal is to use models to reduce physical experiments. The test will be whether model predictions measurably lower cycle times and experimental burden, rather than simply adding another analytical layer.
- Partnerability and IP posture: If proprietary biodata is the moat, Rivercell will need a clear stance on what is shared with partners versus retained, and how that impacts commercial adoption.
Competitive backdrop
Rivercell’s approach aligns with a visible shift in AI drug discovery: the most credible platforms are increasingly those that control data generation rather than relying on publicly available corpora or third-party licensed datasets. In that framing, automated experimental infrastructure moves from being a cost centre to being part of the product.
The open question is how quickly the company can translate infrastructure into repeatable discovery outcomes and partner traction, especially as well-capitalised incumbents pursue similar data-advantaged strategies.
What to watch next
- Build-out milestones for the automated wet lab in Paris and initial throughput metrics
- Evidence that the Virtual Cell models reduce experimental cycles or improve hit identification
- Early pharma or biotech collaborations and the commercial structure (service, platform access, co-development)
- Hiring signals in lab automation, data engineering and translational biology
- Any follow-on financing plans tied to specific validation milestones