Polyphron raises $20M Seed

Quiet Capital led the round, with Gradient, Haystack and Compound participating. Valuation was not disclosed. The New York company pairs two systems: the Tissue Foundry, an automated lab that grows donor-diverse human micro-tissue from stem cells, and the Tissue World Model, an AI layer that predicts how tissue responds to genetic and chemical changes. Cardiac tissue is already in production across multiple donor lines, a broader donor panel is being onboarded, and liver tissue is due later this year to cover drug metabolism and liver toxicity. The team is led by CEO Matthew Osman and chief scientific officer Fabio Boniolo. Vinh Q. Tran recently joined as co-founder and chief AI scientist after post-training research on Gemini at Google DeepMind. COO George Pilitsis previously built the Datapoints product line at Ginkgo Bioworks. The pitch targets a bottleneck in AI drug discovery. Models now generate candidates and hypotheses faster than labs can check them against human biology. Polyphron is selling the verification step rather than the candidate generation. For AI founders, the company is a bet that physical data will be the scarce input in biology. Teams building models in this field may find that proprietary experimental loops matter more than model quality. One open question from outside observers is how well the tissue will capture donor variability and disease states.