Yesterday, a friend in the comments sent me that SemiAnalysis article, saying that CPO has been postponed, and asking if that means the entire optical communication sector is doomed.

After reading it, I’m actually not that pessimistic. I think the biggest problem with the market is that it tends to misinterpret "time delay" as "demand logic is gone."

If Kyber really gets delayed, it affects the timeline for large-scale CPO adoption, but it doesn’t mean AI data centers suddenly no longer need optical communication.

GPUs are still increasing, racks are still getting bigger, and data is still growing — none of that has changed.

The more FUD there is, the more it’s worth doing serious research and contrarian thinking.

What the market is punishing now is the expectation that CPO will soon be fully commercialized. But the long-term demand for companies making lasers, optical modules, DSPs, and optical components — I don’t think it disappears because of one piece of news.

Technology upgrades are rarely completed overnight. GPUs improve generation by generation. Networks also improve generation by generation.

Will CPO eventually come? I think it's highly likely.

Technology routes like Pluggable, AEC, LPO, and optical DSP will continue to ramp up in the coming years, and they won’t lose value just because CPO is delayed.

The stocks that have been hit hardest in the past few days aren’t necessarily the ones with the biggest fundamental changes.

AAOI dropped from $235 to a little over $110.

LITE, for example, fell from a high of nearly $1,100 to around $730.

Let me give an example.

In LITE’s recent earnings calls, they mentioned that demand for AI data center-related lasers is still very strong, and capacity for some products is basically booked by customers through 2027, and the company has been expanding production.

If demand for AI optical interconnects were really weakening, why are those orders still there? Why continue to expand capacity?

Fundamentals haven’t changed, but you’re given a discount opportunity. I think we should seize these weekly-level adjustment opportunities.

There are a lot of optical earnings reports in August. I prefer to see stocks drop or consolidate before earnings; if earnings beat expectations even slightly, it can trigger massive short covering and fundamental catch-up rallies.

A lot of the time, the market sells off an entire sector just because of one delay. But the key is to look at which companies are only affected by sentiment and which ones are actually seeing changes in orders, customers, or competitiveness.

The best opportunities come when fundamentals haven’t changed, but people start doubting, short-sellers start talking it down, and all kinds of FUD is flying around.

So far, I haven’t seen any weakening in AI’s demand for high-speed interconnects and optical communication. CPO may come later, but optics won’t die just because CPO is delayed.