AI startup SandboxAQ, spun out of Google and backed by Nvidia, has released a massive dataset to accelerate drug discovery. Using Nvidia chips, it generated 5.2 million synthetic 3D molecular structures—based on real experimental data—to help predict how drugs bind to proteins, a crucial step in treatment development. This synthetic data, now publicly available, can train AI models to replicate lab results more efficiently. While the dataset is free, SandboxAQ plans to monetize its own AI models built using it. The goal is to simulate binding predictions virtually, saving time and resources in pharmaceutical research with high accuracy.