AI drug discovery deals move to data: 3billion, Cascade and Etcelera

3billion takes its genomic data pitch to Cologne

3billion said on 3 November it will showcase its AI drug discovery work at BIO-Europe 2026. The conference runs 9-11 November in Cologne. The Seoul-based company will hold partnering meetings there with pharmaceutical and biotechnology companies.

For pharma scouts and biotech founders, the value of an AI partner now hangs on data access rather than model quality. Patient genomes, immune atlases and living tissue are the assets being traded. What none of these alliances has produced yet is a clinical candidate.

Its pitch rests on what it calls a Genetics-First strategy. That approach draws on genomic data from more than 120,000 patients and more than 600,000 clinical case records, according to the company's announcement. 3billion says it looks for genetic modifiers that affect how severe a disease becomes, by studying people who carry disease-causing variants but stay asymptomatic or only mildly affected. The aim is to find protective mechanisms and design molecules that reproduce them. The company also says it uses AI to find previously unexplored drug-binding sites on proteins that have been hard to target. It is advancing oral small-molecule programs in oncology and immunology.

BIO-Europe is expected to attract more than 6,000 industry professionals.

The partnership is the dataset

Parse Biosciences and Nomic Bio announced Cascade, a baseline dataset of 100 million cells built to map the human immune response. The numbers are the argument. Cascade will combine 100 million single-cell RNA-seq profiles with more than 11,000 secretory protein profiles. It will aggregate how immune cells respond to more than 150,000 individual and collective cases of cytokines and drugs.

GSK, Pfizer, Sanofi, Graph Therapeutics and Turbine have made foundational commitments to the project. Nomic's Omini 1000 platform will measure more than 1,000 secreted proteins across all samples. Sequencing will run at Parse's GigaLab.

Parse co-founder and chief technology officer Charlie Roco said AI models in biology have faced "a real hurdle and a lack of data available". Nomic co-founder and CEO Milad Dagher made a related point. He said such models need to capture the protein signals cells use to communicate, not only changes in cell state.

Flagship merges three platforms into Etcelera

Flagship Pioneering took the consolidation route. It formed Etcelera by merging Cellarity, Etiome and Empress Therapeutics, BioXconomy reported. Etcelera says the combined platform rests on a foundation model trained on 150 million cells and more than 30 AI-powered prediction models. It claims the platform can trace cause and effect in disease at the molecular level by combining transcriptomic data, chemoproteomics and perturbation data.

CEO Ad Rawcliffe told BioXconomy the investors are not disclosing the terms. "Partnerships will be a big part of the equation," he said. He listed an internal pipeline, an R&D consortium where several pharma companies pay for platform access, and monetising individual components such as data or models. Etcelera launches with programs in myelofibrosis and essential thrombocythemia, plus work in metabolism including partnered MASH programs, and in immunology. Rawcliffe said the company expects to work with both returning and new investors in the future.

The wet lab still sets the pace

Data still needs tissue. Tech.eu reported on 3 November that Vitroscope had raised $7 million, made up of a $5 million seed round and $2 million in non-dilutive funding, taking its total raised to $8.8 million. Its vitro.alive benchtop platform runs six independent, parallel experiments with programmable perfusion and live-cell imaging, and vitro.scale, still in development, is designed for 96.

Vitroscope and Instant Bio kept vascularised human brain organoids under perfusion for seven days, tech.eu reported. Duration is the pitch. Co-founder and CEO Carlo Kriesi said living tissue "cannot be reduced to a snapshot".

In another study using the platform, kidney cells lacking AdipoR1 showed impaired responses to fluid flow. The receptor is linked to diabetic kidney disease.

Model vendors are hedging too. Eric Kauderer-Abrams, Anthropic's head of life sciences, told Pharmaphorum that Claude Science is a "front door" for scientists rather than an AI drug discovery engine. He described it as complementary to NVIDIA's BioNemo toolkit, which provides specialised models for specific tasks.

What to watch at BIO-Europe

BIO-Europe runs 9-11 November in Cologne. 3billion did not name any partners in its announcement. Etcelera's creation terms remain undisclosed. Cascade's pharma backers include GSK, Pfizer and Sanofi.

The comparison worth carrying into the conference is scale. Cascade's 100 million cells sit below the 150 million cells Etcelera says its foundation model was trained on, though the two are counting different things. Vitroscope has raised $8.8 million in total.

Whether any of this shortens the path to an approved drug is unknown. No one at the conference can answer it yet.