Nice article. One thing I found particularly interesting in this analysis is that it implicitly shifts the discussion away from photonic performance and toward qualification and manufacturing scalability.

Most investors still seem focused on whether III-V devices can outperform existing solutions. Increasingly, I find myself asking a different question: if AI optical infrastructure continues scaling as projected, where does the next manufacturing bottleneck emerge?

If the industry ultimately becomes constrained by foundry portability, wafer-scale integration, qualification throughput, packaging complexity, or optical manufacturing capacity rather than device physics itself, the value may accrue very differently than most current models assume.

Curious which of the eight opportunity areas you believe is most likely to achieve qualification first. Defense SWIR appears to have the shortest path, while datacom and CPO arguably offer the largest long-term upside. Which market do you believe is most likely to validate the platform initially?