The ~90-company CPO supply chain has no clean map — that's a sourcing and diligence problem

~90 companies sit in the co-packaged optics (CPO) supply chain, and I have yet to find one clean, sourceable map of them. For anyone doing supply-chain strategy or diligence on optical modules, that gap is the actual bottleneck — not the technology.

Here's why it matters. Pluggable optical modules are running into power and density limits as AI training and inference clusters scale. CPO moves the optics next to the switch ASIC, cutting electrical reach and power per bit. The roadmap is reasonably well covered in trade press. The supply chain underneath it is not.

When I try to assemble that ~90-company picture, it fragments fast:

• Photonic integrated circuit (PIC) design and fabs
• Laser sources and external laser modules
• Optical engines and packaging / advanced assembly
• Connectors, fiber shuffles, and detachable optical interfaces
• Switch ASIC and system vendors pulling it together
• Test, burn-in, and reliability houses that rarely show up in public decks

Each layer has a few credible names and a long tail of single-source or pre-revenue players. For a strategist that's the whole game: where is the concentration risk, who is actually shipping versus sampling, and which "CPO companies" are really one customer away from zero.

A few notes for anyone building their own map:

1. Separate demonstrated production from roadmap slides. Sampling is not supply.
2. Treat laser sourcing as its own risk axis — it concentrates faster than packaging.
3. Watch packaging and test capacity; it is the quietest constraint and the hardest to add.
4. Be careful with financials. Funding and valuation figures float around this space; treat any number as unverified unless it is reported by Bloomberg, Reuters, or the WSJ, and say so explicitly.

I am not naming targets or giving investment advice — the point is methodology. A defensible map is built layer by layer, with each company tagged by role, maturity, and customer concentration, then stress-tested against who disappears if one switch vendor changes course.

If you are sourcing or doing diligence in optical modules: are you working from a real bottom-up map, or from the same five names everyone repeats? I'd genuinely like to compare frameworks.

Source-traced from a verified signal on the CPO sourcing gap; not investment advice.