Insilico Medicine Launches AI-Powered Virtual Aging Cell

Hong Kong, August 14, 2026

Virtual Aging Cell Brings Biological Age Into AI Research

Virtual cells use AI and mathematical modeling to simulate cellular biology, enabling researchers to investigate drug responses, disease mechanisms, and potential therapeutic targets. Insilico says many existing virtual-cell approaches rely heavily on data captured at individual time points, which can limit their ability to represent dynamic biological processes such as cell differentiation, reprogramming, aging, and environmental responses. The VAC concept is designed to address this limitation by incorporating biological time and age into a unified computational framework. Its multi-agent architecture is intended to evaluate cellular behavior across six biological scales: molecular, intracellular, intercellular, tissue, organ, and organism/population. By connecting these levels, the platform aims to simulate how biological systems change over time and how interventions may influence cellular and systemic outcomes.

Multi-Agent AI Targets Dynamic Biological Processes

The new VAC platform introduces a Multi-Agent Swarms architecture that goes beyond static characterization of biological states. Insilico describes a system in which Master Agents coordinate tasks and Specialist Agents perform functions across different biological scales, using databases, omics tools, and literature knowledge graphs. Biological age serves as a core condition throughout the reasoning process, allowing the system to consider how young, mature, and aged biological environments may respond differently to interventions. The platform is also designed to explore cell-fate intervention, including scenarios involving drug-target inhibition, gene knockout, or environmental stimuli, with agents assessing downstream pathway changes and interactions across biological hierarchies. This approach is intended to move virtual-cell research from observing biological states toward computationally exploring how those states could potentially be changed. The VAC initiative builds on Insilico’s PreciousGPT series, which progressively expanded the company’s capabilities in aging biology and multi-omics modeling. Precious1GPT, released in 2023, used multimodal Transformers and transfer learning to predict biological age and distinguish disease from control samples. Precious2GPT, introduced in 2024, expanded into conditional generation of synthetic multi-omics data with tissue and age characteristics, while Precious3GPT integrated text, tabular information, knowledge graphs, and multiple omics modalities across human and animal species. Insilico says this research progression ultimately contributed to the development of a broader framework for modeling cellular life cycles and dynamic biological processes.

VAC Supports Future Drug Discovery Applications

Insilico expects the Virtual Aging Cell platform to provide computational support for identifying therapeutic targets, designing cell-reprogramming strategies, evaluating potential anti-aging interventions, and studying disease mechanisms. The company has also released an initial demonstration preview alongside the VAC webpage and plans to discuss the platform and related multimodal foundation-model research with academic and industry partners at ARDD 2026 in Boston. The company is positioning VAC within its broader AI-driven drug discovery ecosystem. Insilico reports that it has nominated 33 preclinical candidates since 2021, with 13 receiving IND approvals or clearances, while its AI-discovered and AI-designed programs have progressed into clinical development. Its most advanced program, rentosertib, completed a Phase IIa clinical trial, providing clinical validation for the company’s AI-driven drug discovery approach.

With the launch of VAC, Insilico is seeking to extend AI-based research from analyzing biological data toward dynamic simulation of cellular and organism-level biology. The company believes combining biological age, multi-omics data, and multi-agent reasoning could create a more comprehensive computational environment for studying healthy and diseased states and identifying new opportunities for pharmaceutical research and drug development.

Source: Insilico Medicine press relese