Recently, LightCounting, a research institution in the optical communication industry, wrote in its latest market report that the Ethernet optical module market will double in growth by 2024. If InP chip suppliers had sufficient production capacity, the market size could have doubled again by 2025.

LightCounting predicts that the shortage of EML and CW laser chips will constrain market growth until the end of 2026. At that time, demand may gradually slow down to help the supply chain restore balance. Although LightCounting's prediction points to a soft landing, supply chain balance is often achieved through market corrections - the earlier the correction comes, the less painful the pain.

The following figure shows the impact of artificial intelligence on the Ethernet optical module market, with predictions covering the use of re determined optical modules, LPO, and CPO schemes in Scale Out and Scale Up networks. From 2026 to 2030, the demand for AI applications will continue to dominate the market, with growth rates exceeding expectations. It can be said that Al has fully controlled the market dominance.

Another interpretation of the above question is: Has AI gained self-awareness and begun to manipulate human decisions? This may explain why the investment frenzy is so intense. Even if this situation has not yet occurred, can we detect it when it actually arrives?

The key question is whether the current investment boom stems from sound financial planning or from a fear of missing out? We need to face this situation calmly.

This investment frenzy was sparked by the early success of ChatGPT, a collaboration between OpenA1 and Microsoft. To defend its dominant position in search engines, Google has invested heavily in the Gemini project, whose performance is now comparable to ChatGPT. Precision advertising based on Al enhanced search engine is a trillion dollar profit battlefield for Google and Microsoft to compete in. Currently, the AI investment scale of the two companies has a reasonable foundation

But are there any other validated profitable AI business scenarios? The answer is negative. Most application scenarios are still in the early stages of development, and success is not guaranteed. Perhaps it is only a matter of time before investors lose patience and withdraw their funds and leave. At present, investors' investment in new technologies is no less than that of the Internet foam and the telecom foam 25 years ago - unless there is really self-conscious AI manipulating them.