Google antigravity launches AI chemist tool to accelerate cancer drug discovery

Google DeepMind announced the release of new Cheminformatics Skills for Google Antigravity, a tool designed to expedite various stages of drug discovery for researchers. This AI agent integrates large language model capabilities with deterministic chemical calculation tools like RDKit, connecting to public databases such as ChEMBL and PubChem. In a demonstration focused on lung cancer therapeutics, the system was shown to screen thousands of compounds and propose novel rigid macrocycles with viable synthetic routes in about one hour—a process that previously took several days. The tool can standardize molecular identification across different formats, including IUPAC names and SDF files. During the analysis of crizotinib and Lorlatinib, the AI processed 568 compounds, identified links between molecular rigidity and efficacy enhancement, and designed potential superior macrocycles, complete with reaction mechanisms. Google stressed that the AI output represents testable scientific hypotheses, requiring subsequent lab validation. The feature is currently available to Google Antigravity users via Google Cloud, with Google planning to offer six months of the highest usage tier free to up to 10,000 eligible scientists.