The single development that animated @aleabitoreddit today was a TrendForce report indicating that AMD — the No. 2 AI accelerator and CPU maker — is placing multiple large procurement orders to lock up continuous-wave (CW) laser supply. CW lasers are the light source for silicon photonics and co-packaged optics (CPO), the technique of placing the optical engine right next to a switch or GPU die. He reads AMD's move as the first domino after Nvidia, framing it as game theory: once one buyer signs long-term agreements, rivals scramble to avoid being choked out.

This is the crux of his entire thesis. He argues Western independent CW capacity is nearly exhausted — incumbent Lumentum (LITE), a high-quality large-cap he owns and uses as his sector benchmark, is itself CW-constrained and, by his read, buying from maxed-out Japanese suppliers after Nvidia and Coherent locked up multi-year deals. That leaves few merchant options: Sivers (SIVE), the tiny Swedish indium-phosphide laser maker that is his largest and highest-conviction position, and Applied Optoelectronics (AAOI), the vertically integrated Texas transceiver maker he also owns. Both, he contends, sit inside an increasingly invaluable chokepoint.

He defended AAOI at length against bears, noting the stock has run from roughly $30 to $170. His case: scarce 800G/1.6T transceiver capacity that hyperscalers want, management guidance pointing toward roughly $471M in monthly revenue in 2027 (about $5.6B annualized against a ~$13.5B market cap), and a ramp he likens to Nebius. He flagged the offsetting risks he's named before — dilutive at-the-market share sales (~$600M) and volatility from analyst notes, including bearish takes on Lumentum tied to what he called false CPO-delay rumors.

Adjacent packaging names featured too. AEHR Test Systems jumped about 11% to $116 after a follow-on production order from a silicon-photonics customer for wafer-level burn-in systems; he noted the stock sat in the $30s two months ago. AEHR is his test/burn-in chokepoint play. He reiterated conviction in LPKF (LPK), the German maker of glass-core substrate patterning equipment he calls the "ASML" of next-gen packaging, pointing to its claim that 80% of major players have selected its tools and arguing the re-rating comes simply as volume timelines approach. He also lamented that X-Fab (XFAB), the small European foundry he owns for its silicon-photonics and power-semi exposure, gets no credit for leading Europe's sovereign-AI photonics chain.

In Neoclouds — AI cloud-capacity providers — he revisited his long-running pairing: Nebius (NBIS), his core long, hit all-time highs, while bearish target Iris Energy (IREN) languished, which he tied to IREN's ~$6B equity offering diluting holders. On Intel (INTC), his policy-driven turnaround long, he mocked a Bernstein call, noting the firm's $36 target in January preceded a move to $118. He also dismissed a bearish Kioxia call, and wryly noted Jim Cramer endorsing the SpaceX (SPCX) IPO, which he views as the defining mega-listing of the moment.

This is derived commentary, not advice, and reflects one investor's paraphrased views.