A four-institution team led by the University of Houston will spend the next three years hunting for the permanent magnets of the future by combining AI-guided prediction, synthesis, and characterization in one closed discovery loop.
Read moreA closed loop for magnet discovery
High-performance permanent magnets sit inside nearly every electric motor, generator, and actuator, and the world needs better ones. GAMBIT unites high-throughput computation, machine learning, advanced synthesis, and rapid magnetic characterization in a single closed loop: models propose new magnetic compounds, laboratories make and measure them, and every result sharpens the next round of predictions.
1. Predict
First-principles simulation and machine-learning models screen vast spaces of candidate boride, carbide, and intermetallic compounds, ranking the materials most likely to outperform today's best magnets.
2. Synthesize
Top candidates move to the lab, where complementary synthesis routes, guided by computed reaction pathways, turn predictions into real materials.
3. Validate
Rapid magnetic characterization measures each new material against the properties that matter for a working magnet, without waiting on slow single-crystal studies.
4. Learn
Every result, success or failure, flows back into the models, so each cycle of the loop predicts better than the last. The discovery engine improves as it runs.
Our multi-institutional team
GAMBIT pairs university research groups with startup agility: computation, synthesis, and characterization under a single roadmap.
-
Prof. Jakoah Brgoch
Machine-learning architectures and the synthesis of boride and carbide magnetic materials.
University of Houston
-
Prof. Joshua Bocarsly
Magnetic characterization and the data infrastructure that keeps every measurement machine-ready.
University of Houston
-
Prof. Geoffroy Hautier
High-throughput first-principles computation and large-scale screening of candidate magnets.
Rice University
-
Prof. Jamie Neilson
Non-canonical synthesis routes and in situ X-ray diffraction of reactions as they happen.
Colorado State University
-
Dr. Matthew McDermott
Reaction-network synthesis planning, candidate ranking, and technoeconomic analysis.
Newfound Materials
From discovery to market
GAMBIT is built to leave the lab: stronger permanent magnets for the electric motors, generators, electronics, and HVAC equipment that electrification depends on, judged on cost and manufacturability from the first prediction onward.
Technology-to-market lead
Newfound Materials leads our path to market, running techno-economic and manufacturability analysis on every candidate.
Application partners
We work with Carrier to ground the program in real HVAC applications.
Building products that depend on high-performance magnets? We welcome commercial partners.