A new update to the AlphaFold Database has made available over 2,800 AI-predicted structures of protein complexes from viruses, which are now open to the public. These structures are intended to help scientists better understand viruses and prepare for future pandemics. The dataset focuses on viral families known to infect humans, offering insights into the proteins that viruses use to interact with human cells. This information can help identify key targets for developing diagnostics, vaccines, and treatments. The effort was a result of an international collaboration involving several research institutions and technology companies, including EMBL's European Bioinformatics Institute (EMBL-EBI), Google DeepMind, NVIDIA, Seoul National University, the University of Glasgow, the Swiss Institute of Bioinformatics (SIB), the Coalition for Epidemic Preparedness Innovations (CEPI), and Sungkyunkwan University in South Korea. Additionally, researchers from Lund University in Sweden contributed independently to the dataset. The release of this data coincides with the United Nations General Assembly High-level Meeting on Pandemic Prevention, Preparedness and Response, scheduled for September 25 in New York. The new dataset will be officially announced on September 24 at a roundtable hosted by CEPI and the World Economic Forum. The AlphaFold Database was launched five years ago through a partnership between Google DeepMind and EMBL-EBI. Since then, it has grown to include more than 260 million predicted protein and protein complex structures, covering nearly all known proteins in science. For this new release, the team developed a specialized workflow to predict the structures of viral proteins, using optimized tools like NVIDIA BioNeMo Inference Runtime. They also used a carefully selected set of viral proteomes to ensure high accuracy and efficiency in their predictions. While the dataset provides detailed information about the structure of viral proteins, it does not predict how genetic changes might affect a virus’s behavior or how it interacts with humans. Scientists still need to conduct experiments in laboratories to understand how these proteins function in real-world conditions. The dataset is a valuable resource for research and the development of countermeasures, but it does not explain why some viruses are more dangerous or transmissible than others, nor does it allow the creation of more dangerous pathogens. The data can be accessed through the Pandemic Preparedness Portal on the AlphaFold Database website. Scientists and researchers around the world can now use this information to advance their work in virology and pandemic preparedness.