# Industrial AI MOC
Industrial AI is the application of machine learning and optimization to physical processes: factories, plants, airports, utilities. The core challenge is not the AI. It's bridging the gap between messy physical reality and clean algorithmic assumptions.
Every industrial AI company faces the same existential question: are you a consultancy with software, or software that occasionally consults? The answer lives in [[Deployment Velocity]].
## Foundations
1. [[Digital Twins]]
2. [[Surrogate Models]]
3. [[Simulation-Based Optimization]]
4. [[Knowledge Graphs for Industrial Data]]
5. [[Reinforcement Learning for Process Control]]
6. [[Industrial MLOps]]
## The Physics-to-Software Pipeline
The full stack for industrial AI runs: sensor data > data contextualization > physics model > surrogate compression > optimization agent > human-in-the-loop decision.
Each step is where companies live or die:
7. [[Sensor Data and Historian Systems]]
8. [[Auto-Generated Physics Models]]
9. [[Neural Network Compression of Simulations]]
10. [[Human-in-the-Loop Systems]]
## Scaling Dynamics
The economics of industrial AI are brutal. Every deployment touches bespoke engineering: different sensor vendors, different process chemistry, different operator workflows. The companies that win are the ones that automate the bespoke parts.
11. [[Deployment Velocity]]
12. [[Consultancy-to-Platform Transition]]
13. [[Bespoke Engineering in Industrial AI]]
14. [[Model Maintenance at Scale]]
## Competitive Landscape
15. [[Incumbent Bundling Risk]]
16. [[Technical Moat Assessment Framework]]
17. [[Sovereign AI Positioning]]
## Investment Lens
18. [[Industrial AI Unit Economics]]
19. [[Land-and-Expand in Enterprise AI]]
20. [[Key-Person Risk in Deep Tech]]
21. [[IP Strategy for Deep Tech Startups]]
22. [[Technical DD Framework]]
## Cross-Links
- [[First Principles and Mental Models MoC]] : [[Wright’s Law]] applies to deployment learning curves
- [[knowledge graphs]] : foundational concept for industrial data contextualization
- [[Data Centre First Principles]] : adjacent physical infrastructure domain
- [[Investing/Investing System MoC]] : investment frameworks apply directly to evaluating this space
---
Tags: #deeptech #systems #kp