Industrial AI
Industrial AI
Industrial AI is AI put to work inside industrial operations: plants, grids, mines, pipelines, rail, ports and factories. It reads the operation's own systems of record and sensors, joins them in one object model, reasons over that model with open-weight models running on the operator's hardware, and puts each decision in front of the person who owns it. CodeNinja Atoms has published 44 industrial AI designs, one per real operation, across 12 sectors in the United States (19), Saudi Arabia (13), Pakistan (11), Germany (1), with 643 typed objects between them.
Industrial AI use cases by sector
One row per sector. Each design is one industrial AI use case at a real operation, written up as a complete reference architecture.
The architecture of industrial AI
Every published design follows the same seven layers, bottom up. Across the 44 designs the object model layer holds 643 objects and 560 typed links, the inference layer names 127 open-weight model choices, and 44 of 44 designs name the person who approves each action. The sovereign boundary is what makes it sovereign: every layer runs on hardware the operator owns or controls, in its own country, with no third-party AI service in the serving path.
| Layer | What it holds |
|---|---|
| 1. Sources | the systems of record the operation already runs, read only; they stay the system of record |
| 2. Sensing | cameras, detectors and instruments, existing or added, sized from the streams they produce |
| 3. Adapters | one family per kind of source, owning its contract, retries and ordering; the only way data enters |
| 4. Object model | the operation as typed objects, properties and links (an ontology), held on infrastructure the operator controls |
| 5. Inference | open-weight models only, placed at the edge, the site or one in-country node by latency and link loss |
| 6. Services | what the design computes over the object model: forecasts, detections, joins, evidence |
| 7. Surfaces | where people see the result and a named person decides; actions are written one way into the owning system |
- Industrial AI reference architecture: every design by sector, with patterns, layers, compute, cost and DOI
- Industrial data ontology: the data model under each industrial AI design
- Physical AI system design: how each one was designed, decision by decision
Industrial AI starts with the data the operation already has
No design here begins with a new data platform. Each reads the systems of record the operation already runs, read only, and joins them in an object model; the models come after. Browse every object model at industrial ontology examples.
Built from the published designs only: each paper's At a glance table, its object model and the dataset. Designed on Praxis; object models load into Hyper Ontology. Also: Physical AI · Ontologies · Self-improving systems · sovereign AI. Updated 2026-11-14.
Want one of these designs for your own operation? Talk to us.