DeepCell raises ~$14.8M seed to fuse AI and physics for drug discovery
What's the deal? DeepCell, a Chinese AI drug-discovery and virtual-cell startup, has raised nearly 100 million yuan (about $14.8 million) in a seed round. CAS StarDealroom has a profile for this one. Try Dealroom →, Shunxi FundDealroom has a profile for this one. Try Dealroom →, and Qianji CapitalDealroom has a profile for this one. Try Dealroom → co-led, with Dingfeng Kechuang, Shuimu Venture CapitalDealroom has a profile for this one. Try Dealroom →, and Zero2IPO Holdings participating.
What's the endgame? DeepCell wants to move AI from fitting historical data toward understanding molecular interactions and disease mechanisms. Its plan pairs a high-precision molecular-design platform with a "virtual cell" world model that simulates how cells respond to drugs.
Who's behind it? The team is led by Chen Xindong, a researcher trained under Tsinghua University academician Feng Xiqiao, with a decade in the field. His work spans quantum-scale molecular interaction, molecular dynamics, and cell-scale physical modelling, published in journals including PNAS and ACS Nano.
Why now? Investors see AI drug discovery moving from proof of concept to real deployment. "AI is pushing drug discovery into an era of 'precision engineering,'" said Qianji Capital chief investment officer Wu Hai.
Why fuse AI with physics? Chen argues mainstream models lean too heavily on statistical patterns from structure and sequence data. That struggles with novel molecules, new targets, and systems lacking known structures. DeepCell instead tries to encode molecular physics — force, energy, and dynamics — as learnable, verifiable model constraints.
What could go wrong? The approach carries a high technical bar and must bridge AI, physical science, and drug development. As several investors noted, the payoff depends less on model size than on whether it improves research decisions and produces verifiable, druggable candidates.
The signal: The round underscores investor appetite for foundational research infrastructure over quick wins. Backers frame virtual-cell models as potential base-layer infrastructure for future biotech — and a marker of China's push for self-reliance in drug research.
Read more: Jiemian
Image credit: oakridgelabnews
