Vivodyne: AI drug discovery needs better human tissue data
Original: AI isn’t close to curing cancer. This startup says it knows what it will take.
Why This Matters
Vivodyne's approach highlights a critical data gap limiting AI's real-world drug discovery potential.
Biotech startup Vivodyne, backed by $80M from Khosla Ventures, argues AI drug discovery is failing due to poor biological data. Its robotic HIVE labs grow 20 types of human tissue, autonomously dosing and monitoring them to generate causal data that current AI models lack.
Biotech startup Vivodyne contends that the AI drug-discovery industry faces a fundamental data problem: existing AI models are trained on animal test results, single-cell studies, or protein data — not living human tissue. CEO and co-founder Andrei Georgescu warns this gap means AI will 'cure cancer in mice,' not humans.
Vivodyne's solution is its HIVE system — modular robotic laboratories capable of growing 20 types of human tissue, then autonomously dosing and monitoring them to produce causal biological data. The company reports its liver tissue achieves 94% predictive accuracy compared to human toxicity trials, its airway tissue matches real human behavior 96% of the time, and its bone marrow shows 100% concordance across 20 chemotherapy drug tests.
Spun out of the University of Pennsylvania in 2021, Vivodyne has raised just under $80 million across two funding rounds led by Khosla Ventures. The company recently opened what it describes as the world's largest robotic tissue lab.
The broader AI-drug-discovery landscape remains cautious. Anthropic CEO Dario Amodei acknowledged over the weekend that AI curing cancer has become 'more cliche than credible.' Google DeepMind's Nobel Prize-winning AlphaFold, while a major scientific advance, has yet to produce an approved drug. Isomorphic Labs — built on AlphaFold — now expects its first human trials by end of 2026. Industry-wide, 90% of drugs effective in animal testing fail in human clinical trials.