SandboxAQ Brings AI Catalyst Discovery to Claude

Catalysts underpin 80% of global manufacturing, from fertilizer production to sustainable aviation fuel, yet the search for new materials has historically been a bottleneck. Conventional Density Functional Theory (DFT) calculations are precise but prohibitively slow and expensive. AQCat bypasses these constraints by calculating adsorption energy—the primary indicator of a material's viability—with near-DFT accuracy at a fraction of the time.

Unlike standard machine-learning models that often struggle with magnetic properties, AQCat is "spin-aware." It accounts for magnetic behavior in metals like iron, cobalt, and nickel, which are abundant yet notoriously difficult to model. Trained on a dataset of 13.5 million high-fidelity calculations, the model enables research teams to rank thousands of candidates instantly. This allows scientists to prioritize their budgets and physical lab resources for only the most promising materials, effectively shortening innovation cycles that previously spanned years.