Lattice Graph × Noveon Magnetics

Lattice Graph operates a computational materials-discovery platform built around a knowledge graph that spans millions of compositions, connecting crystal structure, electronic and thermodynamic properties, synthesis history, and patent status in a single governed graph. For a rare-earth magnet manufacturer, the practical payoff is that any Dy/Tb-lean candidate magnet composition, any proposed recovery chemistry for magnet leachate, and any sorbent ligand design can be interrogated across that graph for stability evidence, synthesis precedent, and IP exposure simultaneously rather than through three separate research workflows. Candidate materials are validated by running multiple independent physics engines in parallel. Machine-learning interatomic potentials including MACE and CHGNet, together with DFT calculations, produce phonon-stability and thermodynamic-stability verdicts that only advance when independent methods agree. For Noveon, this consensus approach matters most when screening Dy/Tb-lean magnet compositions for coercivity and thermal stability, and when evaluating new ligand designs for heavy-rare-earth selectivity in a leachate environment where slight coordination-chemistry differences between Dy, Tb, and the dominant Nd fraction determine whether a sorbent is commercially viable. A third layer that separates Lattice Graph from standard computational screens is the atlas of labeled negative results. The platform holds tens of thousands of documented failed experiments, the kind of data that rarely appears in publications or commercial datasets, giving every discovery campaign the ability to avoid dead ends that have already been run internally. Coupled with freedom-to-operate screening across more than 300,000 materials patents, every composition or process route that emerges from a Lattice Graph campaign arrives with both a physics-grounded stability case and a clear IP path, which is exactly what an OEM customer or investor diligence team needs before committing to a qualified recycled-magnet supply chain.

Noveon Magnetics has built its commercial thesis on something structurally difficult to replicate: a magnet-to-magnet recycling process that produces qualified sintered NdFeB from end-of-life and swarf feed rather than virgin mined rare earths. That model shortens the supply chain and reduces dependence on a primary production base that is geographically concentrated and increasingly subject to export controls. The $215M raise signals that Noveon is moving from proof-of-concept into scaled manufacturing, and the pressures that come with that transition are well defined: feed volume and quality, management of dysprosium and terbium content in high-coercivity grades, and a need to tell a credible, data-backed story about how each of those risks is contained. Lattice Graph fits Noveon at the exact point where those pressures are hardest to address with internal R&D alone. The platform's critical-mineral recovery and recycling separations portfolio includes a Dy/Tb-selective ion-imprinted resin scoped specifically to NdFeB magnet leachate, an integrated magnet-recycling flowsheet, and supporting assets for recovering co-product value from process streams. These are not adjacent technologies adapted to the magnet space; they were developed around the specific chemistry of heavy-rare-earth separation from the mixed-lanthanide leachates that magnet recycling produces. For a company scaling a process that generates exactly that stream, licensing or co-developing around those assets is faster and more capital-efficient than building proprietary separation chemistry from scratch. The data layer reinforces this fit. Noveon's feedstock scouting, Dy/Tb exposure quantification, and freedom-to-operate work on Dy/Tb-lean compositions all require the kind of structured, cross-referenced intelligence that a governed materials knowledge graph provides. Live supply-risk data, mineral-deposit concentration indices, and composition-level patent whitespace screening turn Noveon's two largest strategic risks into managed, evidence-backed positions rather than narratives. That evidence is what separates a credible recycled-magnet scale-up story from one that stalls in OEM qualification or investor scrutiny.

Noveon Magnetics has built its commercial thesis on something structurally difficult to replicate: a magnet-to-magnet recycling process that produces qualified sintered NdFeB from end-of-life and swarf feed rather than virgin mined rare earths. That model shortens the supply chain and reduces dependence on a primary production base that is geographically concentrated and increasingly subject to export controls. The $215M raise signals that Noveon is moving from proof-of-concept into scaled manufacturing, and the pressures that come with that transition are well defined: feed volume and quality, management of dysprosium and terbium content in high-coercivity grades, and a need to tell a credible, data-backed story about how each of those risks is contained. Lattice Graph fits Noveon at the exact point where those pressures are hardest to address with internal R&D alone. The platform's critical-mineral recovery and recycling separations portfolio includes a Dy/Tb-selective ion-imprinted resin scoped specifically to NdFeB magnet leachate, an integrated magnet-recycling flowsheet, and supporting assets for recovering co-product value from process streams. These are not adjacent technologies adapted to the magnet space; they were developed around the specific chemistry of heavy-rare-earth separation from the mixed-lanthanide leachates that magnet recycling produces. For a company scaling a process that generates exactly that stream, licensing or co-developing around those assets is faster and more capital-efficient than building proprietary separation chemistry from scratch. The data layer reinforces this fit. Noveon's feedstock scouting, Dy/Tb exposure quantification, and freedom-to-operate work on Dy/Tb-lean compositions all require the kind of structured, cross-referenced intelligence that a governed materials knowledge graph provides. Live supply-risk data, mineral-deposit concentration indices, and composition-level patent whitespace screening turn Noveon's two largest strategic risks into managed, evidence-backed positions rather than narratives. That evidence is what separates a credible recycled-magnet scale-up story from one that stalls in OEM qualification or investor scrutiny.

Noveon's matched portfolio, the critical-mineral recovery and recycling separations portfolio, is the closest alignment in the Lattice Graph holdings to a magnet recycling operation. The highest-priority asset is the ion-imprinted phosphonate-bis-picolinamide resin designed for dysprosium and terbium separation from magnet leachate, which is the precise stream a sintered-NdFeB recycler generates when it dissolves end-of-life or swarf feed. Its predicted Dy/Tb separation factors and its Dy/Nd factors are high enough to enable meaningful heavy-rare-earth fractionation from a mixed lanthanide leachate without the full reagent inventory and capital footprint of conventional solvent-extraction cascades. That makes it a realistic fit for a manufacturer that wants to control Dy/Tb routing at plant scale but is not trying to build a primary rare-earth refinery. The integrated flowsheet platform extends that asset to system level. It combines a magnet-recycling separation train with other recovery cascades into a single unified technology portfolio, providing the form factor Noveon would actually deploy if it brought heavy-rare-earth grade management in-house: a qualified end-to-end process from leachate to separated Dy/Tb fraction rather than a collection of point solutions that must be independently integrated. As Noveon adds lines with its new capital, owning or licensing a validated integrated cascade protects the recovery-to-product chain and strengthens the vertical-integration story that OEM customers and the investors behind the $215M want to see. The universal chelating-resin platform rounds out the portfolio match by addressing what happens to the rest of the leachate. A single crosslinked resin with interchangeable binding groups can recover a broad range of critical oxocations from the same industrial streams. For Noveon, minor metals and impurities in its process streams represent byproduct value that currently exits as cost. A configurable recovery platform lets the company turn those streams into additional revenue as throughput scales, broadening from pure magnet manufacturer toward a recycling-capable, byproduct-monetizing operation without requiring separate development programs for each target element.

The most-used surfaces for a Noveon team would be the supply intelligence dashboard and the knowledge-graph explorer. The supply dashboard, built on the element-level supply risk and mineral-deposit data, gives sourcing and strategy staff a live view of Dy/Tb concentration, HHI trends, and available conversion routes from secondary streams. That dashboard becomes the artifact behind feedstock decisions and the recycled-feed pitch to OEM procurement and investor relations alike. The knowledge-graph explorer lets materials scientists interrogate Dy/Tb-lean magnet candidates and leachate-separation chemistries in a composition-360 view: structure, thermodynamic stability evidence, synthesis precedent, and provenance all visible in one place, with natural-language graph queries that do not require writing code against a database. On the IP and screening side, the freedom-to-operate and patent-whitespace workflow and the invention-opportunity hunt engine let the Noveon IP and R&D team screen candidate compositions and process routes before committing bench resources, and scan for under-protected Dy/Tb-lean and magnet-recycling whitespace worth filing into. Batch screening with the machine-learning formation-energy predictor supports rapid triage of substitution chemistries, so the team can prioritize which Dy/Tb-lean candidates to validate experimentally against the most promising, clear-path options. Dossier generation turns a completed screen into a structured report ready for licensing, investment, or OEM qualification conversations.

The natural entry point for Noveon is a parallel-track engagement: a scoped data and evaluation license covering the three matched API products and a knowledge-graph seat, running alongside a bench validation of the Dy/Tb separation resin on Noveon's own magnet leachate. The data license is operational immediately and gives Noveon's sourcing, strategy, and IP teams the supply intelligence and whitespace tools they need during the scale-up phase. The bench validation answers the specific separation performance question on Noveon's actual feed chemistry, which is the gating result before a license or co-development deal on the separation assets makes economic sense. These two tracks de-risk each other: the data layer informs feed-sourcing decisions while the bench work proceeds, and the bench results inform how broadly Noveon wants to license across the integrated flowsheet and supporting assets. Upon successful benching, the natural conversion is a license or co-development agreement covering the Dy/Tb separation resin and the integrated magnet-recycling flowsheet, with the universal chelating-resin platform available as an expansion option. Commercial terms would be structured around milestones rather than flat upfront fees, reflecting the asset's stage and Noveon's preference for capital efficiency during scale-up. The market-size estimates for these assets, ranging from approximately $500 million to $5 billion depending on the specific technology and the scope of deployment, are provided as context for sizing the opportunity rather than as revenue or pricing commitments; all economics would be scoped in a term sheet following the evaluation phase.