Google DeepMind has introduced new AI-powered drug discovery tools within its Antigravity platform, allowing researchers to use autonomous agents to analyze chemical data, predict molecular potency and propose new drug candidates. Available through the Science plug-in, the tools aim to accelerate early-stage pharmaceutical research by automating complex analytical tasks that traditionally require several days.
The platform enables researchers to instruct an AI agent to search scientific literature and patents, retrieve chemical compound data and identify relevant molecular properties. The agent can then analyze relationships between chemical structures and biological activity, develop predictive models and suggest new molecules for laboratory testing. This approach combines scientific data retrieval, computational modeling and molecular design within a single workflow.
Google DeepMind demonstrated the technology using crizotinib and lorlatinib, two drugs used to treat certain forms of non-small cell lung cancer. The AI agent analyzed thousands of chemical compounds and compared molecules with small structural differences to understand how those changes affected potency. The analysis suggested that adding bonds to create more rigid ring structures could improve molecular activity, helping guide the design of new compounds.
Based on these findings, the agent developed potency prediction models and proposed new ring-shaped molecules alongside potential synthesis methods. According to DeepMind, the demonstration took approximately one hour, compared with several days using conventional research workflows. However, the company emphasized that all AI-generated molecular designs still require laboratory synthesis and experimental validation before their potential can be confirmed.
The tools are available through the Google Antigravity application, with the company offering six months of free access to its highest usage tier for up to 10,000 eligible scientists. The launch reflects growing investment in AI-driven drug discovery, as companies including Isomorphic Labs, Iambic and Owkin develop technologies that combine predictive models, generative AI and automated research workflows.
Click here to read the original news story.
