Ångström AI

Ångström AI is an AI physics simulation platform for pharmaceutical research, designed to compute molecular properties that would otherwise require laboratory experiments. It calculates solubility, lipophilicity, crystal structure stability, free energy differences, binding conformations and hydration sites, with the stated aim of matching experimental error.

The platform uses machine learning interatomic potentials, including CSP-MACE-Å for crystal structure prediction, and alchemical free energy protocols for solvation free energy calculations. These models run substantially faster than DFT, enabling energy and free energy evaluation across many candidate structures. The company's research also covers generative models for Boltzmann distributions, including flow annealed importance sampling bootstrap and SE(3) equivariant augmented coupling flows.

Ångström AI is aimed at drug discovery teams evaluating solid forms, solvation properties and binding behavior. The company offers demos through direct contact, though specific packaging details such as free plans, trials, editions or pricing models are not stated on the page.