Vivodyne, a biotech startup spun out of the University of Pennsylvania, argues that AI drug discovery is being held back by a lack of good data on human biology, and says it has built the fix. Its robotic lab system, called HIVE, grows 20 types of human tissue and autonomously doses and monitors them, generating "causal" biological data that current AI models lack because most existing data comes from animal testing or isolated cells and proteins rather than living human tissue. The claim pushes back on bold predictions from AI leaders, including Dario Amodei, Sam Altman and Demis Hassabis, that AI is close to curing cancer or disease more broadly, when tangible results so far remain limited.

Founded in 2021 by CEO Andrei Georgescu after his bioengineering PhD at Penn, Vivodyne says its tissue models closely mirror real human organs, citing 94% predictive accuracy for liver toxicity, 96% accuracy for airway tissue, and 100% concordance across 20 chemotherapy drugs tested on its bone marrow tissue. The company has raised almost $80 million across two funding rounds led by Khosla Ventures, and last week opened what it calls the world's largest "human data centre" near San Francisco, claiming to already achieve twice the throughput of all US animal trials combined. Vivodyne says it is working with several major, unnamed pharmaceutical companies, aiming to help drug candidates clear the costly, high-failure path to clinical trials more reliably.