Fabian Theis maps the next phase of Virtual Cell research
In a Latent Space interview, Fabian Theis argues that Virtual Cell research should move beyond basic perturbation prediction toward modeling interactions across molecules, cells, tissues and organisms. The discussion also highlights Biohub's reported $1.8 billion in commitments to collecting large-scale biological datasets, alongside questions about evaluation metrics and the performance of complex models.
The interview frames a shift from AI-driven molecular design toward data-intensive, multi-scale biological modeling, while highlighting how dataset scale and better evaluation could affect Virtual Cell research.

Fabian Theis has been one of the key players in the various iterations of the Virtual Cell concept and initiatives. It is not hyperbole at all to say that his lab has been pushing the frontier of algorithms to understand multi-scale, multi-modal (”multiomics”), time-varying biological data.
Indeed, he has shaped the large scale vision and data collection through his leadership at Helmholtz, various startups and probably at Biohub itself.
Watch the full episode on YouTube…
Much of the recent attention in AI x Bio has been on protein and molecular design. AlphaFold, ESM and BioHub, Chai, Leash, Isomporphic, Boltz, Genesis, Octant, Baker Lab, and many many more. After a lot of initial excitement, Virtual Cell projects have received less attention, due in no small part to questions about the effectiveness of large-scale models in solving even basic Virtual Cell problems.