SandboxAQ has announced a strategic collaboration with NVIDIA aimed at expanding the reach and impact of its Large Quantitative Models (LQMs). Through NVIDIA DGX Cloud on Google Cloud, SandboxAQ is accelerating the development of its AI-native scientific platform—enhancing performance, accuracy, and scale in domains like biopharma, materials science, and cybersecurity.
At the core of this collaboration is the use of DGX Cloud to train and operate LQMs capable of performing high-resolution, equation-based simulations across complex scientific systems. This shift enables SandboxAQ to replace traditional design-make-test cycles with high-throughput simulations that significantly shorten discovery timelines. By enabling faster validation of drug candidates, materials, and chemical processes, the company aims to improve predictability and reduce the risk of costly failures in the R&D lifecycle.
A key outcome of the initiative is the creation of curated, high-fidelity datasets that combine molecular modeling and generative AI techniques. These datasets form the basis for knowledge graphs and predictive models used in therapeutic development, offering more accurate insights into how small molecules interact with biological systems. In addition, SandboxAQ has launched its AI Chemist—an agentic AI system capable of autonomously exploring millions of chemical pathways. This development opens the door to accelerated molecule discovery that extends well beyond the limits of traditional research methods.
The collaboration builds on previous research milestones. In 2024, the companies achieved an 80x speedup in quantum chemistry simulations using GPU acceleration. In 2025, they jointly published research demonstrating orbital optimization at an unprecedented scale—handling systems with over 80 electrons and orbitals, which had previously been computationally infeasible.
