# 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