Junping Yi
English中文

The AI Infrastructure & Semiconductor Investment Atlas

Follow demand from applications into accelerator systems, data centers, manufacturing, and power—then test which securities retain the economics.

Read the atlas

Package → rack → facility → grid

What this book connects

The atlas follows economic value through the full physical stack. It connects a model workload to chips, memory, packaging, networks, data-center power, manufacturing capacity, national policy, and the public or private entities exposed to each constraint.

Chip-to-grid anatomy from accelerator package through data center and electrical grid

Read the atlas

Orientation

Orientation

  1. →Who turns AI importance into shareholder returns?

Part I

The Central Contradiction

  1. →Who ultimately pays for the AI build?

Part II

The Stack

  1. 2.1Models & the intelligence frontier
  2. 2.2Applications & monetization
  3. 2.3Power & data centers
  4. 2.4Cloud & the software moat
  5. 2.5AI chips & accelerators
  6. 2.6Interconnect & networking
  7. 2.7Memory & advanced packaging
  8. 2.8Foundry & fabrication
  9. 2.9Equipment & EDA
  10. 2.10Materials & inputs
  11. 2.11Capital & market structure
  12. 2.12Geopolitics, policy & risk
  13. 2.13From GPU to AI factory

Part III

Mutual Dependencies

  1. →Where would each system stop first?

Part IV

Where the Bottleneck Moves

  1. →Where the constraint migrates next

Part V

Investment Decisions

  1. →Why an important company can still be a poor investment

Part VI

Audit the Atlas

  1. →How to challenge and recompute the conclusions

Choose a reading path

InvestorFrom strategic map to investable exposureOperatorFrom package and rack to the 100 MW facilityPolicy readerGeopolitics, controls, and two diverging stacks