Industry Map
AI Capital Map
From models and applications to compute, manufacturing, packaging, interconnect, data centers, and energy — map where AI capital actually lands.
2 min read
AI is not only a story about models and applications.
When a technology cycle expands, capital reallocates along the full value chain. BGA uses the AI Capital Map to break research into layered structure and focus on the infrastructure growth truly depends on.
Nine layers
- Models & applications: How demand forms, and which use cases drive scaled capital spending.
- Compute: Training and inference capacity set the pace of upstream expansion.
- Advanced manufacturing: Chip manufacturing ability constrains supply elasticity.
- High-bandwidth memory: Memory and bandwidth become critical cost and performance variables.
- Advanced packaging: Packaging and interconnect determine system-level performance release.
- Servers & high-speed interconnect: Network efficiency shapes deployment density.
- Data centers: Construction, cooling, siting, and delivery cycles.
- Energy infrastructure: Power, grids, and storage become longer-cycle constraints.
- Data & storage: Data assets and storage systems support durable application expansion.
Follow The Bottleneck
Hotspots only show what markets are watching. Higher-value research asks:
- Which critical links does growth depend on?
- If demand keeps expanding, where might the next bottleneck appear?
- Which links remain under-recognized or mispriced?
Follow The Bottleneck — follow constraints, not only narratives.