Insilico Medicine has launched the concept page and interactive preview of its Virtual Aging Cell (VAC) platform, a multi-agent swarm architecture that introduces biological age as the core conditional variable for dynamic cellular simulation across six biological scales: molecular, intracellular, intercellular, tissue, organ and whole-organism. Unlike static-snapshot virtual cell models, VAC deploys master agents for global task orchestration and specialist agents for scale-specific reasoning, enabling adaptive negotiation and joint inference of cellular differentiation, reprogramming, ageing trajectories, and intervention responses.

The platform builds on Insilico's 12-year trajectory from the 2014 NVIDIA GTC proposal through the PreciousGPT series (Precious1GPT for age prediction, Precious2GPT for conditional multi-omics generation, Precious3GPT for cross-species multimodal modelling) to this multi-agent convergence.

The VAC system is designed to answer the question of how to reverse and reshape cell fate, rather than merely describe static states. When simulating drug inhibition, gene knockout, or environmental stimuli, agents across scales predict microenvironment remodelling, downstream cascade effects, and cross-scale structural reorganisation in real time.