Vivodyne's HIVE human data centers aim to close the data gap in AI drug discovery
Vivodyne, a biotech startup located in Philadelphia, is tackling the shortage of causal human biological data essential for AI-driven drug discovery through its innovative HIVE modular robotic laboratories. These labs are capable of cultivating 20 different types of human tissue and can perform experiments autonomously. Co-founder and CEO Andrei Georgescu asserts that the absence of this data means AI models may only succeed in 'curing cancer in mice.' This perspective is shared by Dario Amodei, CEO of Anthropic, who has recently voiced skepticism regarding AI's cancer treatment claims. Founded in 2021 as a spin-off from the University of Pennsylvania, Vivodyne has raised nearly $80 million and has established the largest human data center globally near San Francisco to enhance drug testing reliability prior to clinical trials.
Key facts
- Vivodyne is a biotech startup that uses HIVE robotic labs to grow 20 types of human tissue and generate causal biological data.
- Andrei Georgescu, CEO and co-founder, says AI models without human data will only cure cancer in mice.
- Anthropic CEO Dario Amodei recently called AI-curing-cancer claims clichéd; Sam Altman and Demis Hassabis have made similar statements.
- Only a few AI-designed drugs have entered human trials; one is in Phase III, and AlphaFold has not produced a new drug.
- Isomorphic Labs expects its first clinical trials by the end of this year, later than the original 2025 plan.
- Vivodyne was spun out of the University of Pennsylvania in 2021; it has raised nearly $80 million led by Khosla Ventures.
- The company's human tissue models show high predictive accuracy: 94% for liver toxicity, 96% for airway, and 100% for bone marrow.
- Vivodyne opened what it calls the world's largest 'human data center' near San Francisco, with double the throughput of all U.S. animal trials.
- About 90% of drugs that pass animal testing fail in human clinical trials, a challenge the startup aims to address.
- A Nature Methods study found no clear data scaling laws for generative AI trained on cellular data.
Entities
Artists
- Andrei Georgescu
- Dario Amodei
- Sam Altman
- Demis Hassabis
- Tim Fernholz
Institutions
- Vivodyne
- University of Pennsylvania
- Khosla Ventures
- Anthropic
- OpenAI
- Google DeepMind
- Isomorphic Labs
- Nature Methods
- FDA
- NHTSA
Locations
- San Francisco
- United States