Serenity, the small-cap-focused stock-picker who back-maps the AI buildout to its most obscure upstream suppliers, had a busy day spanning valuation arbitrage, a fresh catalyst in one of his European holdings, and a recap of his two dominant themes.

The NAV-discount hunt. He opened by siding with activist Palliser Capital on Taiwan-listed WUS Printed Circuit (2316), a PCB maker he owns. His gripe: WUS trades near a ~$1B market cap while holding an 11.26% stake in its Kunshan subsidiary worth several billion — a glaring NAV discount, where a company trades below the value of what it owns. He argued management could sell a fraction of that locked-up stake, fund PCB growth into the AI demand inflection, and still leave the market to re-rate the rest. Later he asked followers for other extreme NAV-discount names that also have independent AI growth, citing ACMR — a US-listed semiconductor cleaning-equipment maker with a China-based operating arm — and again WUS, versus the better-known SK Square.

Riber catalyst. His French holding Riber (ALRIB) released general-meeting notes. Serenity flagged that a second "ROSIE" system is due shortly for a leading US quantum-computing customer, plus intensifying business development in photonics materials. Riber makes molecular-beam-epitaxy (MBE) tools in an effective duopoly with Veeco, and his core thesis rests on OSINT-sourced evidence that a hyperscaler's quantum program quietly depends on its machines — making it the kind of sub-$1B chokepoint he favors. He called the update a positive development.

Photonics philosophy. Responding to Chinese-language followers, Serenity reaffirmed that laser companies like Applied Optoelectronics (AAOI) and Sivers (SIVE) are his favorites because their revenue can expand from selling lasers into full optical modules and external light sources. AAOI is his Made-in-America vertically integrated transceiver bet; SIVE, his largest position, makes the indium-phosphide light sources he sees as the chokepoint for co-packaged optics (placing the optical engine next to the switch die). He contrasted them with Coherent (COHR), a vertically integrated photonics compounder that extends upstream into substrates. His key point: photonics is still extremely early, with the real revenue ramp arriving in 2027.

Memory victory lap. Serenity reminded readers that sentiment was broadly bearish on memory three months ago, then ticked off the moves — Micron (MU), Sandisk (SNDK) and the Korea ETF EWY (his Samsung/SK Hynix proxy) all sharply higher. MU and SNDK anchor his "memory supercycle" thesis of structural AI-driven DRAM, HBM and NAND shortages. He used the rally to argue investors should think independently of consensus narratives, suggesting optical names face the same skepticism today.

Macro and policy. He noted SpaceX (SPCX), the headline mega-IPO he's tracked, dipped after a Jim Cramer endorsement — a wry aside, not a position. And he flagged President Trump's announced Intel–Apple partnership sending Intel (INTC) up 8%, reportedly surprising Intel executives. INTC is his policy-driven bet that Washington won't let its domestic foundry champion fail; he quipped the President is "hard carrying" the stock, with Apple cast as a pressured customer rather than a holding.

This is derived commentary, not investment advice.