What We're Looking For

What We're Looking For

What We're Looking For

We're seeking computationally de‑risked inventions that can be patented and licensed across NMK Capital's innovation challenge and the 10 focus areas we discussed:

1. Telecom / RAN & Core

2. Satellite / Ground Segment

3. IoT / Edge / Embedded

4. Semiconductors / EDA / Packaging

5. Energy Storage / Batteries

6. Electric Power / Grid Operations

7. Robotics / Autonomy / Facilities (Open‑RMF/ROS)

8. CFD / Thermofluids / Aero

9. Materials & Chemistry (non‑biomedical)

10. Quantum / SDR / Physics

We're looking for systems, methods, devices, compositions, processes, and diagnostics/monitoring inventions that can be proven computationally, replicated from your repo/zip, and claimed cleanly—analogous to the way you'd build and validate an AI‑scientist project, but outside life sciences. Examples include:


  • Control and orchestration methods (e.g., RIC xApps/rApps, grid controllers, multi‑robot facility orchestration, flight autonomy supervisors).


  • Design flows & tool‑calling agents (EDA flows, packaging/assembly planning, CFD/thermal topology optimization, grid planning toolchains).


  • Compositions & materials/structures (battery electrodes/electrolytes/interlayers, surface treatments/coatings, composites, metamaterials), where the performance uplift can be demonstrated in simulation/digital twins.


  • Form‑factor, deployment‑route & device innovations (edge vs. cloud split, packaging/thermal designs, radio front‑end/antenna structures, device‑level integration).


  • Combination systems with measured "synergy" (coordinated controllers/waveforms/schedules/policies that beat baselines beyond additive effects).


  • Sensor‑/telemetry‑guided methods of use (thresholds/decision rules for safe performance or efficiency).


  • Diagnostics/monitoring/imaging for selection, operations, or predictive maintenance (e.g., grid state estimation, link‑quality prediction, thermal early‑warning, structural health monitoring).

1) What "good" looks like (selection lens)

1) What "good" looks like (selection lens)

1) What "good" looks like (selection lens)

We prioritize inventions that:


  • Show a clear performance edge (efficiency, safety, reliability, latency/throughput, cost, convenience) and a path to real‑world adoption.


  • Demonstrate novelty with an explicit prior‑art map and a draft claim (see "IP & Claiming" in the form).


  • Are replicable from your repo/zip (data inventory, prompts chain, environment files, commands).


  • Include a human‑contribution log (who conceived what; how AI was used; why a human is an inventor).

2) Patentable & high‑value claim types (with why they matter)

2) Patentable & high‑value claim types (with why they matter)

2) Patentable & high‑value claim types (with why they matter)

Systems & Methods (control/orchestration/scheduling; network, grid, robotics, autonomy, spectrum). Most licensable in complex systems—claims tie directly to measurable KPIs (SLA, reliability, energy, latency).


Devices/Architectures (hardware structures, packages, antennas, flight stacks, facility interfaces). Durable protection where the architecture itself yields unexpected performance.


Materials/Compositions & Structures (electrodes/electrolytes/interlayers, composites, coatings, metamaterials, phases/polymorphs). Strong exclusivity—solid‑state variants, dopants, and engineered microstructures can materially change performance/CMC and claim scope.


Processes & Manufacturing Methods (semiconductor steps, deposition/anneal windows, assembly, test, calibration). Practical exclusivity that OEMs license because it embeds in production.

Interoperability/Protocol Methods (interfaces, standard‑aligned behaviors—e.g., RIC Service Models, ROS message/services schemas). Can rise to SEP‑like licensing if adopted.

Diagnostics/Monitoring (sensing, estimation, anomaly/predictive maintenance) when they unlock claimable method linkages (selection, control, or maintenance).

Systems & Methods (control/orchestration/scheduling; network, grid, robotics, autonomy, spectrum). Most licensable in complex systems—claims tie directly to measurable KPIs (SLA, reliability, energy, latency).


Devices/Architectures (hardware structures, packages, antennas, flight stacks, facility interfaces). Durable protection where the architecture itself yields unexpected performance.


Materials/Compositions & Structures (electrodes/electrolytes/interlayers, composites, coatings, metamaterials, phases/polymorphs). Strong exclusivity—solid‑state variants, dopants, and engineered microstructures can materially change performance/CMC and claim scope.


Processes & Manufacturing Methods (semiconductor steps, deposition/anneal windows, assembly, test, calibration). Practical exclusivity that OEMs license because it embeds in production.

Interoperability/Protocol Methods (interfaces, standard‑aligned behaviors—e.g., RIC Service Models, ROS message/services schemas). Can rise to SEP‑like licensing if adopted.

Diagnostics/Monitoring (sensing, estimation, anomaly/predictive maintenance) when they unlock claimable method linkages (selection, control, or maintenance).

Systems & Methods (control/orchestration/scheduling; network, grid, robotics, autonomy, spectrum). Most licensable in complex systems—claims tie directly to measurable KPIs (SLA, reliability, energy, latency).


Devices/Architectures (hardware structures, packages, antennas, flight stacks, facility interfaces). Durable protection where the architecture itself yields unexpected performance.


Materials/Compositions & Structures (electrodes/electrolytes/interlayers, composites, coatings, metamaterials, phases/polymorphs). Strong exclusivity—solid‑state variants, dopants, and engineered microstructures can materially change performance/CMC and claim scope.


Processes & Manufacturing Methods (semiconductor steps, deposition/anneal windows, assembly, test, calibration). Practical exclusivity that OEMs license because it embeds in production.

Interoperability/Protocol Methods (interfaces, standard‑aligned behaviors—e.g., RIC Service Models, ROS message/services schemas). Can rise to SEP‑like licensing if adopted.

Diagnostics/Monitoring (sensing, estimation, anomaly/predictive maintenance) when they unlock claimable method linkages (selection, control, or maintenance).

Note: Stay at an in‑silico level here—no lab/factory build instructions. We evaluate computation + rationale + claimability.

3) Concrete angles we love (examples, not limits)

3) Concrete angles we love (examples, not limits)

3) Concrete angles we love (examples, not limits)

  • Deployment & integration upgrades: cloud→edge splits; packaging/thermal routes that deliver SLA‑stable latency or energy savings under realistic loads; standardized device/stack integration methods.


  • Service‑life / reliability engineering: controller/design variants that measurably extend asset life or reduce variance (e.g., grid components, batteries, RF front‑ends, actuators) with bounded risk.


  • "Heavy‑element/structure" strategies: dopants, interlayers, coatings, metamaterials, and surface/phase engineering that plausibly shift stability, efficiency, or safety—computationally demonstrated.


  • Activation/triggered systems solving bottlenecks (e.g., temperature/field‑responsive materials, protocol activation gates) with specific rationale and safe operating windows.


  • Synergistic combinations with quantified uplift over baselines (e.g., coordinated schedulers/controllers/waveforms/policies with robust statistical tests) and dose‑window analogs for safe operation.


  • Telemetry‑defined methods of use (predictive, enrichment, or safety thresholds) with a decision rule (sensor features, limits, inclusions/exclusions).


  • Repurposing (new use contexts) where physics/controls evidence supports feasibility (standards alignment, deployment constraints).

4) What we don't accept

4) What we don't accept

4) What we don't accept

✔ Content that includes PII, secrets, or controlled/dual‑use details enabling misuse.

✔ Public postings of enabling details pre‑filing (kills foreign rights).

✔ Unverifiable "black‑box" claims without data lineage and reproducibility.