NVIDIA announced on September 24, 2026, that it has joined a coalition of research organizations—including Google DeepMind, EMBL-EBI, the Coalition for Epidemic Preparedness Innovations, Seoul National University, Sungkyunkwan University, the Swiss Institute of Bioinformatics, and the University of Glasgow—to release predicted 3D structures for the protein complexes of more than 2,800 viruses. The structures are openly available through the AlphaFold Database, which NVIDIA's blog states now holds more than 260 million protein and protein-complex predictions.
The structures were generated using AlphaFold2, Google DeepMind's protein-folding model, with inference optimized by NVIDIA BioNeMo Inference Runtime on NVIDIA GPUs. NVIDIA is also making the BioNeMo Structure Prediction Pipeline openly available, a GPU-accelerated workflow that lets researchers go from a protein sequence to a predicted 3D structure for their own targets.
According to NVIDIA, approximately 30% of the protein interactions in the new dataset represent structures not previously documented in the Protein Data Bank. The dataset covers viral families known to infect humans, ranging from common-cold viruses to emerging threats such as Mpox, and predictions are labeled by confidence level.
The release coincides with a United Nations General Assembly meeting convened by the World Economic Forum on pandemic prevention,