AI can generate thousands of new material candidates in minutes. That part is solved. The bottleneck is everything that comes after.
Problem
Can it be synthesized? Does it meet cost requirements? Does it clear regulatory hurdles? Can it scale to manufacturing? Each of those questions can take years to answer for a material no one has seen before.
Meanwhile, decades of scientific research have produced vast knowledge about materials across medicine, construction, electronics, food science, and industry. That knowledge sits in silos. A compound studied in radiology may have optical properties that are exactly what a building materials engineer needs — and no one has made that connection, because no one reads across all of it.
Approach
We build AI agents that read across all of it.
Our platform combines three things.
We encode thousands of known compounds with their physical properties, synthesis routes, regulatory status, existing applications, and commercial availability — drawn from Materials Project, JARVIS, and scientific literature.
Our agents mine scientific literature and commercial databases across disciplines, identifying materials whose properties in one field match an unsolved problem in another.
Before anything reaches a lab, candidates are ranked using density functional theory and optical simulation — grounding every recommendation in established physical principles, not just pattern matching.
At every step we filter for synthesis feasibility, cost, regulatory compliance, and manufacturing compatibility. By the time a material reaches the experiment stage, we already know its synthesis route, supply chain, regulatory standing, and commercial pathway.
The only remaining question is whether it performs as intended.
Difference
Generative AI materials discovery produces novel compounds, then faces every qualification hurdle from scratch. One well-funded startup recently reported generating over 500 novel material candidates.
One had a plausible synthesis pathway.
We do not generate new materials. We find the right existing ones. That shift eliminates most of the bottleneck — known materials come with known synthesis routes, known costs, known safety profiles, and often existing regulatory approval. What remains is connecting them to new problems, and validating that the physics works.
Team
Nexus Materials LLC is a materials discovery company founded by a computational materials scientist and an experimental chemist.
Principal Investigator
Pratik Dhakal, PhD
Chemical Engineering, with research focused on using machine learning, DFT, and molecular simulation to identify materials for battery applications.
Co-founder
Ajhit Kumar, PhD
Chemistry, with 30 years of research experience, 150 publications, and 17 patents in advanced inorganic functional materials for energy, biomedical, and security applications.