WARSAW, Poland, June 03 — Ingenix, the AI and biology company built around a single question – “What would it take for AI to truly understand biology?” – has today announced a €13m seed-extension funding round led by Sofinnova Partners, with participation from Inovo VC and OTB VC.
The funding will scale development of Ingenix’s Biological Reasoning Engine and broaden its work with pharma and biotech partners through the Qualified Access Program, launching today.
The dominant approach in AI for drug development, which trains larger models on larger datasets assuming biology will yield to scale, faces a structural problem. Biology is profoundly complex, fundamentally non-linguistic, and runs across modalities and biological scales. No single model, however large, can capture that on its own. Moreover, no single company holds all the data, across every modality and biological scale, to train one. Ingenix has built a different architecture, Modality Fusion, which integrates best-in-class models across modalities and biological scales, then reasons across their representations directly. The Biological Reasoning Engine is what Modality Fusion enables.
“This funding lets us extend the Biological Reasoning Engine to the partners and questions where it can do the most useful work,” said Piotr Surma, CEO and co-founder of Ingenix. “We built Ingenix on the conviction that biology needs an AI architecture designed for biology and not a general-purpose model retrofitted to it. The early results have been stronger than we forecast, and we’re excited to extend the Engine to a select number of partners through the Qualified Access Program.”
“It is no longer enough to just build models,” added Simon Turner, Partner at Sofinnova Partners. “Ingenix is building the reasoning layer, the part that actually connects the biology, the chemistry, and the clinical data into something a scientist can interrogate and act on. That’s the hard bit, and that’s where the value compounds. We’re thrilled to back a team that gets that.”
The Engine in Action: ADC Payload Prioritization
In a recent engagement, Ingenix applied its Biological Reasoning Engine to a dual-payload ADC prioritization problem for an oncology biotech which had thousands of possible payload configurations but lacked the experimental capacity to test them.
The Engine produced 15 candidate combinations. Under blind expert review by the biotech’s translational science team, the predictions broke down as follows:
– 5 were publicly known hypotheses.
– 2 were supported in existing literature but not widely cited.
– 3 had been confirmed by the biotech through internal experiments but never published and never disclosed to Ingenix.
– 5 were novel hypotheses not previously considered by the biotech team. Of these, the biotech flagged 3 as actionable candidates.
The double-payload ADC space is too new for any AI system to have seen meaningful training data on it. The engine reached its predictions by reasoning from first principles about the underlying biology, rather than by pattern-matching against prior examples.
In short, insights that had taken the biotech several years of research and millions of euros to develop were accomplished by the Engine in a matter of minutes.
Applications to the Qualified Access Program are now open at ingenix.ai/qap.