Kewzinvasty
127 posts



AI's $1T photonics bottleneck isn't lasers. It's testing them. When I attended $NVDA GTC, a comment from the CTO of $CRWV at a niche side event stuck with me. “If there’s a day where I don’t have to hear about failed optics, I will be so happy." This brought me down a rabbit hole. And what I found completely changed how I think about the AI trade. Photons are unreliable and optics failures are costing data centers billions of dollars. Before any GPU cluster goes live every optical component has to pass brutal, failure proof testing. One bad laser in a 100,000 GPU cluster isn’t a bug. It’s a multi-million dollar outage. Now optical testing isn’t a side market. It’s going from 19% → 30% of the entire photonics stack this decade. It's growing faster than the optics themselves! This means the real leverage in the photonics trade might not be in the laser makers themselves but in the gatekeepers signing off on every photon. I allocated 20% of my $3M fund tracking photonics stocks to two stocks positioned to outperform as the optics guardians. To explain the thesis, @KawzInvests and I just dropped our photonics testing deep dive on Substack. Link in bio or below. $VIAV $AEHR $COHR $LITE $AAOI $CIEN $TSEM

Credo $CRDO is acquiring DustPhotonics for $750 million cash plus equity. This is the biggest move they've made. My concern with CRDO was always copper. Their business was built on Active Electrical Cables, and the shift toward silicon photonics was the one thing that could break the thesis. This deal answers that. DustPhotonics is a fabless Israeli company with 70 engineers and a SiPho PIC portfolio spanning 400G, 800G, and 1.6T with a 3.2T roadmap. Their L3C laser coupling technology lowers power, simplifies manufacturing, and removes the EML supply chain problem that has been holding the industry back. CEO William Brennan said it plainly: "a vertically integrated connectivity platform that spans from copper to optical and from chip to cluster." That is the whole thesis now. $CRDO $COHR $LITE $AAOI











