We’ve raised $30M to build the foundational dataset for Virtual Cell Models
We’ve raised $30M to build the foundational dataset for Virtual Cell Models: 1Bn single-cell datapoints, mapping 1M drug-patient interactions, to be shared with one partner. Our goal: Move the frontier - From models to precision medicines that help patients.
Building AI models of proteins was a pinnacle of human achievement. The next summit: modeling cells and systems, and their interactions with drugs. Protein models democratized design; virtual cell models will help predict impact in diverse patients.
We do not yet know what model ends up deserving of the “virtual cell” crown. What we know: the oversampled, observational data from human cells will not be enough to bring about this monumental inflection point.
Earlier this year, we released Tahoe-100M — the world’s first gigascale single-cell perturbation dataset. It turned a slow-burning quest for virtual cells into a blaze, becoming a scaffolding for every serious effort on modeling human cells, downloaded 100K times in months.
We are now going 10x larger: 1Bn Cells, 1M drug-patient interactions. Because that’s what it takes to cross the chasm between experimental models to models with clinical impact.
For clinical translation, we need depth, clarity and focus - to that end, we will select one partner to join us in this journey. We’ll combine our data with the right partner’s capabilities, whether in AI or clinical development, to move the frontier from models to the clinic.
We invite you, the serious, the ambitious, the optimist to join us in this next chapter.
It’s morning in biology. Let’s build what matters.