The Layer Model

The layer model organizes the AI ecosystem into horizontal stacks where each layer depends on the layer below it. The standard AI ecosystem stack has six layers: infrastructure (chips, data centers, networking), compute platform (cloud providers, GPU-as-a-service), foundation models (frontier labs, open-source models), middleware (orchestration, RAG frameworks, fine-tuning platforms), application (vertical AI solutions, copilots, agents), and end user (enterprise buyers, consumer users). Mapping actors to layers reveals market structure, competitive dynamics within layers, and the value chain dependencies between layers.

1. Populate Each Layer List 5-15 actors per layer, prioritized by market share and strategic relevance. For each actor, note: market position (leader, challenger, niche), trajectory (growing, stable, declining), and key relationships to actors in adjacent layers. The goal is a complete but manageable view — not every startup, but every actor whose moves could affect your strategy.
2. Map Cross-Layer Dependencies Draw the dependencies between layers. Which foundation model providers depend on which compute platforms? Which application vendors depend on which middleware? Cross-layer dependencies reveal strategic vulnerabilities: if a compute platform raises prices, which foundation model providers are affected, and which application vendors downstream are affected in turn?
3. Identify Layer Dynamics Each layer has a dominant dynamic: consolidation (few large players emerging), fragmentation (many small players, no dominant share), or integration (layer boundaries blurring as actors expand vertically). The dynamic determines competitive behavior within the layer and strategic opportunities at the layer boundaries.