Google DeepMind has introduced the Antigravity platform, designed to accelerate drug discovery by automating research workflows. The system integrates over 30 scientific databases with proprietary models, allowing researchers to complete complex analysis tasks in minutes rather than days.
During a demonstration on October 15, 2026, the platform identified a new antimicrobial peptide candidate targeting Streptococcus pseudopneumoniae. The process involved using the ProtNLM model to analyze millions of unannotated entries from the UniProt database.
The identified candidate underwent in silico verification via the AlphaFold model, which predicts protein structures. Currently, the project is in the validation phase, with the platform generating publication-ready artifacts, including experimental plans and 3D renderings.
Antigravity builds upon the Science Skills suite launched by Google in May 2026, which provides more than 11 tools for proteomics and cheminformatics. The new platform adds an agent layer that enables the autonomous chaining of research steps.
