Invention candidates relevant to Absolics (SKC)

Lattice Graph is a computational materials-discovery platform built around a knowledge graph spanning millions of compositions. Priority candidate materials must earn consensus validation across multiple independent machine-learning interatomic potentials — drawn from MACE, CHGNet, MatterSim, and ORB — before advancing, and phonon and thermodynamic stability are required checkpoints, not optional filters. Materials that clear that bar proceed to targeted density-functional-theory simulations and, ultimately, to deposition-ready specifications. Beyond candidate generation, Lattice Graph carries a large, curated atlas of labeled negative results — failed synthesis attempts and disqualified compositions — so the search space collapses before a single wafer is touched. Freedom-to-operate and patent-whitespace screening across our indexed corpus of materials patents is embedded in the same platform, meaning IP risk is assessed at the composition level, not as an afterthought once a formulation is already in the process roadmap. The result is a full-stack discovery engine: from composition hypothesis, through multi-potential validation and DFT refinement, to intellectual-property clarity — all before pilot deposition.

Absolics occupies a singular position in the advanced-packaging supply chain: as the SKC subsidiary that is already manufacturing glass-core substrates at its Covington, Georgia fab and sampling to AMD, it is the one company in the ecosystem that needs chemistry stack solutions now, not at some future inflection point. The approximately $75 million in CHIPS Act funding and approximately $100 million in NAPMP support mean Absolics has the capital and the government mandate to move quickly — and the pressure to arrive at production chemistry with clean intellectual-property standing, because every layer above the glass will be scrutinized by customers, partners, and regulators alike. Lattice Graph maps directly to that commercial reality. The platform's Glass-core advanced-packaging substrates portfolio was built around precisely the liner, barrier, and redistribution-layer dielectric chemistries that Absolics needs to license or co-develop to complete its stack. Rather than directing Absolics toward a portfolio of speculative compositions, we can offer computational candidates that passed multi-potential stability screening, a preliminary patent screen of the relevant landscape (not a freedom-to-operate opinion), and process sequences designed for ALD/CVD deposition — a starting point for Absolics' process integration team, with deposition, measurement and qualification still to be done. The co-development and licensing fit is structural, not coincidental. Absolics ships the glass; Lattice Graph's candidate assets are aimed at what goes on top of it. That combination — first-mover substrate platform plus a computationally screened, preliminary-IP-screened candidate chemistry stack — is, in our view, a proposition hyperscaler and accelerator customers may look for as the industry scales. An engagement now positions both organizations to present a unified, deposition-ready glass-core package to the market before competing organic-substrate incumbents can answer.

Absolics occupies a singular position in the advanced-packaging supply chain: as the SKC subsidiary that is already manufacturing glass-core substrates at its Covington, Georgia fab and sampling to AMD, it is the one company in the ecosystem that needs chemistry stack solutions now, not at some future inflection point. The approximately $75 million in CHIPS Act funding and approximately $100 million in NAPMP support mean Absolics has the capital and the government mandate to move quickly — and the pressure to arrive at production chemistry with clean intellectual-property standing, because every layer above the glass will be scrutinized by customers, partners, and regulators alike. Lattice Graph maps directly to that commercial reality. The platform's Glass-core advanced-packaging substrates portfolio was built around precisely the liner, barrier, and redistribution-layer dielectric chemistries that Absolics needs to license or co-develop to complete its stack. Rather than directing Absolics toward a portfolio of speculative compositions, we can offer computational candidates that passed multi-potential stability screening, a preliminary patent screen of the relevant landscape (not a freedom-to-operate opinion), and process sequences designed for ALD/CVD deposition — a starting point for Absolics' process integration team, with deposition, measurement and qualification still to be done. The co-development and licensing fit is structural, not coincidental. Absolics ships the glass; Lattice Graph's candidate assets are aimed at what goes on top of it. That combination — first-mover substrate platform plus a computationally screened, preliminary-IP-screened candidate chemistry stack — is, in our view, a proposition hyperscaler and accelerator customers may look for as the industry scales. An engagement now positions both organizations to present a unified, deposition-ready glass-core package to the market before competing organic-substrate incumbents can answer.

The Glass-core advanced-packaging substrates portfolio is the primary fit. It contains the through-glass-via liner, copper diffusion barrier, redistribution-layer dielectric, and integrated stack claims that map one-for-one to Absolics' process integration roadmap. The aluminum nitride thermal liner converts a via wall from a thermal resistance into an active heat path — a direct answer to the thermal management challenge that arises as Absolics scales to high-power accelerator packages for customers like AMD. The tungsten boride copper diffusion barrier is proposed, from a computed Cu migration barrier, to block copper at sub-TaN thickness, which would preserve via geometric budget in the high-aspect-ratio through-glass vias that Absolics' glass platform enables. The ordered fabrication method claims cover the full deposition sequence and are structured to be licensed as a manufacturing process, not just as materials claims. The Integrated packaging, storage and PFAS-treatment systems portfolio adds two additional layers of value. Its bandgap-graded borate and oxynitride multilayer dielectric stack is designed to reduce carrier injection, as an input toward sub-2-micron redistribution-layer pitch — the fine-pitch RDL that can differentiate glass-core from conventional organic laminate at the electrical level. The barium hafnate Ruddlesden-Popper high-permittivity dielectric offers a path to integrating high-density MIM capacitors directly into the package, a capability hyperscaler customers are increasingly demanding for on-package power delivery. Both assets carry a preliminary FTO screen across the screened patent landscape. The High-power thermal-interface materials portfolio rounds out the picture at the assembled-package level. As Absolics' glass substrate moves from sampling to volume supply, the thermal interface between the package and the heat spreader becomes a co-optimization problem — and materials in this portfolio have been validated against the same multi-potential thermodynamic framework as the via and dielectric chemistries, ensuring compatibility across the full stack.

The Lattice Graph application gives Absolics' materials team a single workspace for every stage of stack development, from initial composition screening through qualification reporting. Multi-potential validation results — covering MACE, CHGNet, MatterSim, and ORB in parallel — are displayed side by side so engineers can immediately see where consensus holds and where models diverge, which is the signal that matters for de-risking a deposition decision. Phonon stability and thermodynamic stability assessments are integrated into the same view, not presented as separate analysis artifacts, so the team moves from hypothesis to a qualified candidate in one workflow rather than across disconnected tools. The negative-results atlas is surfaced directly in the application, meaning that when an engineer is evaluating a barrier or liner candidate, failed prior compositions in the same chemistry space are visible as context — with the specific failure modes labeled — rather than buried in literature or internal notebooks. A preliminary patent-screen flag (not a freedom-to-operate opinion) is rendered at the composition level alongside stability data, so the process integration team at Covington can see an early IP read on each candidate under consideration, for counsel to review. For Absolics' reporting obligations under its CHIPS and NAPMP agreements, the application's provenance tracking provides an auditable record of how each qualified composition was validated, which simplifies the technical documentation that federal program milestones require.

A Lattice Graph engagement with Absolics begins with a targeted chemistry-stack audit: we map Absolics' current via liner, barrier, and RDL dielectric process targets against the topAssets in the Glass-core advanced-packaging substrates portfolio and the Integrated packaging, storage and PFAS-treatment systems portfolio, identify the closest validated compositions, and deliver freedom-to-operate assessments for each. This first phase is designed to run in parallel with Absolics' ongoing process integration work at Covington, so there is no delay to the sampling program already underway with AMD. Deliverables include ranked candidate specifications, deposition-parameter guidance derived from multi-potential validation, and a composition-level IP landscape report covering the most relevant patent families. From there, an engagement typically moves to a co-development structure or a direct license, depending on Absolics' preference for internal ownership versus speed to qualification. In a co-development arrangement, Lattice Graph provides ongoing platform access — including Knowledge-Graph API and freedom-to-operate API calls — as the Absolics team iterates on deposition conditions, with the platform's negative-results atlas actively informing which failure modes to anticipate at each integration step. Licensing is structured around the specific asset or stack claims that enter Absolics' process roadmap, with pricing calibrated to the addressable market size of the relevant portfolio rather than a flat platform fee. Both paths are designed to deliver a chemistry stack that arrives at AMD qualification with a clean intellectual-property chain — the outcome Absolics' commercial position and federal funding obligations both require.