AI Photonics — Interlinked Chokepoint Map

Focused visual map of the AI photonics supply chain built from the repo's scorecard and reports. It emphasizes the narrowest points: photonics-SOI wafers, InP substrates, MOCVD and lithography tooling, SiPh foundry capacity, laser and analog components, AEC and optical DSP, transceiver production, and the system-layer demand anchors that pull the whole chain through. Chokepoint values use the repo's fixed 25-point, 5-question framework.

GitHub folder Scorecard Chokepoint method Research notes CHIPS status notes
BUY rated
HOLD rated
AVOID rated
Chokepoint
SiPh path
InP / laser path
Assembly / system path
Cross-path dependency
CHIPS final
CHIPS proposed
CHIPS R&D / hub
Silicon Photonics Path
InP / Laser Path
Assembly, Packaging & System Pull

Materials & Substrates

Deep upstream choke points that determine whether photonic wafers and laser die can even be made.

Fab Tools & Foundries

Where the chain narrows from raw materials into qualified manufacturing capacity.

Components & Connectivity

Laser, analog, DSP, retimer, and AEC positions with the strongest pricing power.

Modules & Assembly

Transceiver, assembly, and packaging nodes that turn components into shippable systems.

Systems & Demand

Switches, networking systems, and hyperscaler demand that monetize the entire optical stack.

This is a curated map of the key interlinks, not every company in the photonics universe. It currently visualizes the highest-signal nodes from a broader 78-report scorecard. A node is tagged as a chokepoint when its systematic `chokepoint_score` is `16/25` or higher. CHIPS labels track direct U.S. CHIPS incentives awards, CHIPS R&D awards, and CHIPS-linked hub funding as of 2026-03-24; the full scorecard sweep lives in the scorecard and CHIPS notes linked above.