Lattice Graph × OneIM

Lattice Graph operates a governed materials knowledge graph that links composition, crystal structure, thermodynamic properties, synthesis routes, and IP records across millions of experimentally and computationally characterized materials. When an AI-materials company claims its model has discovered a novel candidate, that claim can be traced through the graph to its underlying evidence: which sources reported it, whether those sources agree or disagree on key properties, and whether the structural prediction survives multi-engine validation from independent physics engines including machine-learning interatomic potentials and density functional theory. The platform does not rely on any single model's output; it computes consensus across engines and flags cases where predictions diverge, giving a diligence reviewer a quantified measure of how much trust a given property claim actually deserves. Beyond property validation, the platform maintains a screening layer across more than 300,000 materials patents, allowing freedom-to-operate and patent whitespace analysis to be run programmatically rather than by hand. For critical-minerals plays, it integrates supply-chain data including geological deposit concentration, Herfindahl-Hirschman Index scores for element-level supply risk, and conversion route economics, so the chemistry and the logistics can be stress-tested together. What makes the platform genuinely distinctive for diligence work is its atlas of labeled negative results: failed experiments, dead-end compositions, and killed candidates that most published datasets and most commercial models have never seen. That corpus of documented failures is the hardest thing for a founder to reproduce, and it is the most useful thing for an investor trying to distinguish a defensible technical moat from a benchmark that happens to look good on the surviving examples.

OneIM occupies a specific and demanding position in deep-tech investing: a long-horizon, multi-strategy firm with patient capital and a mandate broad enough to write checks into AI-for-science, critical minerals, and energy storage, but without the in-house materials science team that most scientists would expect to see underwriting these theses. That combination creates a structural diligence gap. An AI-materials founder can present a benchmark that is technically accurate and still be testing a model on a data distribution that excludes the hardest cases. A critical-minerals company can present a conversion route that is geochemically plausible and still have an element-level supply concentration that makes the thesis fragile. Generalist due diligence teams are rarely staffed to catch either problem, and a multi-year hold amplifies the cost of missing them early. Lattice Graph is built for exactly this gap. The platform exists to produce a third-party, reproducible technical read on materials claims, grounded in a knowledge graph that the target company does not control and cannot retroactively optimize for. For OneIM, this means an investment committee memo can include an exhibit sourced from an independent computational system rather than the founder's own benchmark results, which is a materially different kind of evidence. The two platform products that map most directly to OneIM's diligence questions are the Opportunity and Buyer Intelligence index, which provides a ranked, third-party view of what is genuinely inventable in a target's space and which funded buyers actually pay for it, and the Negatives and Eval-Data Atlas, which tests a target's headline candidates against the documented failures most models never see. The strategic fit is reinforced by the nature of OneIM's deal flow. AI-materials and critical-minerals transactions are arriving into long-horizon capital at a pace that outstrips the ability of generalist diligence to verify the underlying science. The firm's own lineage, which spans structured credit and equity across cycles rather than fund-clock exits, means that the cost of an early technical misjudgment is not diluted by a short hold period. Lattice Graph gives OneIM a way to invest in these theses with the same rigor it would apply to a structured credit underwrite: traceable, reproducible, and grounded in evidence the investor controls rather than evidence the founder presents.

OneIM occupies a specific and demanding position in deep-tech investing: a long-horizon, multi-strategy firm with patient capital and a mandate broad enough to write checks into AI-for-science, critical minerals, and energy storage, but without the in-house materials science team that most scientists would expect to see underwriting these theses. That combination creates a structural diligence gap. An AI-materials founder can present a benchmark that is technically accurate and still be testing a model on a data distribution that excludes the hardest cases. A critical-minerals company can present a conversion route that is geochemically plausible and still have an element-level supply concentration that makes the thesis fragile. Generalist due diligence teams are rarely staffed to catch either problem, and a multi-year hold amplifies the cost of missing them early. Lattice Graph is built for exactly this gap. The platform exists to produce a third-party, reproducible technical read on materials claims, grounded in a knowledge graph that the target company does not control and cannot retroactively optimize for. For OneIM, this means an investment committee memo can include an exhibit sourced from an independent computational system rather than the founder's own benchmark results, which is a materially different kind of evidence. The two platform products that map most directly to OneIM's diligence questions are the Opportunity and Buyer Intelligence index, which provides a ranked, third-party view of what is genuinely inventable in a target's space and which funded buyers actually pay for it, and the Negatives and Eval-Data Atlas, which tests a target's headline candidates against the documented failures most models never see. The strategic fit is reinforced by the nature of OneIM's deal flow. AI-materials and critical-minerals transactions are arriving into long-horizon capital at a pace that outstrips the ability of generalist diligence to verify the underlying science. The firm's own lineage, which spans structured credit and equity across cycles rather than fund-clock exits, means that the cost of an early technical misjudgment is not diluted by a short hold period. Lattice Graph gives OneIM a way to invest in these theses with the same rigor it would apply to a structured credit underwrite: traceable, reproducible, and grounded in evidence the investor controls rather than evidence the founder presents.

The most practical surface for OneIM's analysts is the knowledge-graph explorer, used as a diligence cockpit rather than a research tool. An analyst can pull composition-intelligence reports on a target's flagship materials and walk the graph from composition through crystal structure, thermodynamic properties, synthesis routes, and patent records in a single session, with cross-source trust and disagreement scoring surfaced inline so that shaky predictions are flagged automatically rather than buried in footnotes. The opportunity and buyer intelligence dashboard gives a one-screen read on whitespace and funded-buyer affinity that maps directly onto the investment-committee narrative, condensing what would otherwise require weeks of third-party research into a structured, citable exhibit. For systematic coverage across a pipeline of deals rather than a single name, the platform supports batch screening, allowing a team to run an entire subsector watchlist or a target's full candidate list through the negatives check and opportunity scoring in a single pass. Domain-specific workflows for battery materials, catalysis, and synthesis give deal-specific framing when a transaction sits in energy storage or industrial chemistry. The net effect is that a single analyst, without access to a wet lab or an in-house materials science team, can produce a reproducible technical exhibit for an IC memo in a fraction of the time that conventional expert-network diligence would require, and the output is traceable back to primary evidence in the knowledge graph rather than to an advisor's personal judgment.

The natural entry point for OneIM is a fixed-scope technical diligence engagement on a single live deal: an independent read combining the Opportunity and Buyer Intelligence index and the Negatives and Eval-Data Atlas applied directly to the target's own headline candidates and roadmap claims. A single-deal engagement of this type is estimated in the range of roughly $30,000 to $60,000, with the exact scope set in a brief scoping conversation. The deliverable is a structured technical exhibit suitable for inclusion in an investment committee memo, with findings traceable back to primary evidence in the knowledge graph rather than to a single advisor's opinion. For a firm running multiple AI-materials and critical-minerals transactions per year, the more efficient model is a scoped diligence subscription or retainer that bundles a defined number of named-target reads per period with direct API access to both the Opportunity and Buyer Intelligence and Negatives and Eval-Data products for OneIM's own analysts. A sensible sequence is one paid pilot on a deal already in the firm's active pipeline, calibrated against OneIM's own judgment before any ongoing commitment, then conversion to an annual retainer with optional add-on access to the platform's supply-intelligence, IP screening, and knowledge-graph layers for transactions requiring deeper supply-chain or defensibility work. All figures cited here are estimates for framing and would be confirmed in a direct scoping discussion.