AI Novel Prize
Jury and community award
for the most innovative discoveries
applying AI to chemistry,
physics, and biology
With a jury from:
Meet the 2025 finalists and cast your vote
Chemistry
For discovering a liquid-state high-entropy alloy electrocatalyst that converts nitrate to ammonia.
For discovering an octonary electrocatalyst (Pd-Pt-Cu-Au-Ir-Ce-Nb-Cr) that boosts direct formate fuel-cell performance.
For discovering new electrolyte classes via a unified AI, data, and experiment approach.
For discovering sinter-resistant oxide catalyst supports using interpretable machine learning.
For discovering five new porous transition-metal oxide structures for multivalent-ion batteries using generative AI.
Biology
For the discovery of biologically distinct multiple sclerosis patterns using unsupervised ML on MRI plus sNfL, including an “imaging-first” trajectory where brain structural change can precede rises in the blood biomarker.
For the discovery of a disease-driving “moonlighting” role of PHGDH in Alzheimer’s and identification of NCT-503 as a molecule that blocks that harmful role without simply shutting down PHGDH’s normal enzyme activity.
For the discovery of a deep-learning system MycoBCP that reads TB bacteria microscope images and classifies how compounds damage the cell, pointing to likely mechanisms and targets.
For the discovery of de novo antibiotics designed by generative AI with in vivo efficacy against drug-resistant infections.
For the discovery of an AI-discovered drug-and-target pair validated in a Phase 2a human trial (Rentosertib, TNIK).
Physics
For discovering a physics-informed deep-learning model that predicted fusion ignition at NIF.
For discovering AlphaQubit 2, a fast neural-network decoder for topological quantum error correction (QEC).
For discovering two new superconductors (Be2HfNb2 and Be2HfNb) predicted by an AI pipeline and then made and measured in the lab.
For discovering new altermagnetic materials using an AI search engine (MatAltMag), including “i-wave” altermagnets.
For discovering an AI method that extracts compact, interpretable governing equations from high-dimensional chaotic dynamics.