AI materials discovery has never moved faster. Microsoft’s MatterGen generates structures from target properties. Meta FAIR’s OMat24 and Open Catalyst made fast screening widely available. Schrödinger gives experts deep simulation. Citrine Informatics learns from a company’s own lab data. Periodic Labs, Lila Sciences and Orbital Materials pair AI with labs of their own. Each has pushed the field forward.
MS Lattice answers the question a materials team actually asks: which few should we make first?
The numbers the field reports
- Microsoft and PNNL screened 32.6 million candidates on about 1,000 cloud machines, in under 80 hours, to reach 18.
- Materials Nexus analyzed more than 100 million compositions and took three months to design, make and test a magnet.
Remarkable work. And one clear lesson: the number of candidates was never the product. The answer is.
Requirements inside the search, not a filter after it
Proposals that break your limits are turned away before a structure is built.
A shortlist, not a haystack
Hundreds of ranked contenders instead of millions of structures to triage, each with the record of every check.
Intelligence that learns while it runs
Sharper with every evaluation, at about a third of a cent of compute per candidate.
Proof you can rerun
Both sealed targets found blind. 73 of 73 candidates rebuilt from their record.
Ready for your agent, and for the factory
Your AI agent drives MS Lattice over MCP. When a winner emerges, Norck takes it through validation and production to finished parts.









