SandboxAQ is an enterprise company delivering AI solutions that address some of the world's greatest challenges at the intersection of machine learning and physics. Spun out from Alphabet in 2022, SandboxAQ develops Large Quantitative Models (LQMs)—AI systems trained on high-fidelity scientific data rather than linguistic patterns—to solve complex challenges in chemistry, materials science, and energy.

For the Oil and Gas sector, SandboxAQ is redefining catalyst discovery with its AQCat25-EV2 family of AI models. While traditional R&D relies on slow, expensive trial-and-error laboratory cycles, AQCat25-EV2 enables high-throughput virtual screening with physics-level accuracy. Trained on the proprietary AQCat25 dataset (which includes over 13 million high-fidelity quantum chemistry calculations) is the first model family to incorporate spin polarization. This allows for the precise modeling of all industrially relevant elements, including the abundant transition metals critical to modern refining and petrochemical processes.
By delivering simulations up to 20,000x faster than traditional methods, SandboxAQ empowers energy leaders to accelerate transition-critical initiatives by identifying winning chemical formulations in days rather than years and significantly de-risking the path to commercial scale-up.

To learn more about our breakthrough catalytic models and datasets, visit sandboxaq.com/aqcat25.