A busy session for @aleabitoreddit, who toggled between his two favorite obsessions — value trapped behind asset discounts, and the still-early optical buildout — while taking a victory lap on memory.

The NAV-discount hunt. He opened and closed the day on Taiwan-listed WUS Printed Circuit (2316 TW), a PCB maker whose roughly $1B market value he calls absurd given it holds an 11.26% stake in WUS Kunshan, a Chinese affiliate worth several billion. Echoing activist Palliser Capital, he argued management could sell even a quarter of that stake and pocket cash exceeding the parent's entire market cap, funding PCB growth into AI demand while still leaving billions in assets. It's a textbook "NAV discount" play — a holding trading below the value of what it owns — and he later widened the net, asking followers for other deep NAV discounts that also carry independent AI growth. He flagged ACMR (ACM Research), a US-listed maker of semiconductor wafer-cleaning and packaging tools with a China-heavy business, as another lopsided example he already owns, contrasting both against Korea's SK Square.

Riber's quantum thread tightens. France's Riber (ALRIB) — a molecular-beam-epitaxy equipment maker he frames as a duopoly with Veeco and a chokepoint for compound-semiconductor and photonics materials — released general-meeting notes. He highlighted a second "ROSIE" system due shortly for a leading US quantum-computing customer, plus stronger commercial interest in photonics applications. That dovetails with his long-running, OSINT-sourced thesis that a hyperscaler quantum program quietly depends on Riber's machines.

Optics: still early. Responding to Chinese-language coverage of his book, he restated his core conviction that laser names like AAOI (Applied Optoelectronics) and Sweden's Sivers (SIVE) have the most room to expand — moving beyond selling lasers into full optical modules and external light sources for co-packaged optics. He cited Coherent (COHR), a vertically integrated photonics player that also makes substrates, as the model for climbing upstream to lift margins. His key point: the real revenue ramp doesn't arrive until 2027, so optics remains in the early innings of its supercycle.

Memory victory lap. He reminded readers how bearish sentiment was three months ago, citing big runs in Micron (MU), Sandisk (SNDK) — his template "supercycle" NAND name — and the iShares South Korea ETF (EWY), his leveraged Samsung/SK Hynix proxy, alongside the Korean memory giants themselves. The lesson he drew: ignore the macro fear narratives (oil, LNG, Iran) and think independently — and expect the same skepticism to give way on optical names into 2027–28.

Intel gets a presidential lift. He noted Intel (INTC) jumped about 8% after a Trump-announced partnership with Apple (AAPL) — reportedly a surprise even to Intel executives. It feeds his thesis that Washington won't let its domestic foundry champion fail and will steer megacaps toward it; he wryly suggested the President is single-handedly carrying the stock.

Lighter notes. He joked that even SpaceX (SPCX), the blockbuster space IPO he's watched as a liquidity event, couldn't escape the "Jim Cramer curse," and shared informal survey results on the net worth of finance-focused X users.

Derived commentary, not advice.