Biohub commits $1.8B to Virtual Cell AI initiative
Original title: Controlling Biology - Fabian Theis, Helmholtz Munich
Biohub, NIH, DOE, Google and Meta commit $1.8B to large-scale biological data collection for AI Virtual Cell modeling led by Fabian Theis.
Biohub has announced a $1.8B investment commitment from Biohub, NIH, DOE, Google and Meta to advance Virtual Cell projects using new imaging and measurement capabilities. Fabian Theis, a key figure in Virtual Cell concepts, is involved through his work at Helmholtz Munich. The initiative aims to model multi-scale biological data from molecules to organisms. Existing models struggle with perturbation prediction compared to simple baselines as shown in the 2025 Arc Virtual Cell Challenge.
Unconfirmed: The effectiveness of large-scale models in solving Virtual Cell problems remains questioned.
- Biohub, NIH, DOE, Google and Meta each committed $1.8B? Wait, the text says $1.8B of commitments [1] from Biohub, NIH, DOE, Google and Meta. It's likely total $1.8B
- New lab-in-the-loop automation capabilities at scale
- Virtual Cell Challenge 2025 showed models failing to beat baselines for perturbation prediction
- Focus on multi-scale, co-registered biological data
From the original article
4 min read at latent.space
As systems biology efforts heat up to map biology at higher resolution, greater breadth and more modalities, Fabian Theis is pushing the frontier from mapping to predicting to controlling biology
By RJ Honicky · latent.space
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.