Where should AI run?
Placement of AI workloads across cloud, edge, vehicle and plant, under latency, cost, power and safety constraints. The evidence here helps decide what runs where, and why.
Decision pathway
Research by decision
The decisions this research informs. Each pathway leads with curated reading, then every item tagged to that decision, newest first.
Decisions
-
-
Can AI act?
Whether an AI system or agent may take actions on its own, and what authority, governance and evidence that requires before it does.
-
Is a supplier concentration acceptable?
Exposure to a single critical supplier, input or region, and whether that concentration is acceptable, priced in or avoidable.
-
How to evaluate an SDV platform
Criteria for choosing or assessing a software-defined vehicle platform: kernel and hypervisor, safety certification, update architecture and who owns the platform.
-
How to move an industrial AI pilot to deployment
Moving an industrial AI or robotics pilot into production: the safety case, the data and skills the plant needs, and the economics that decide whether it scales.