Serenity, the small-cap-focused stock-picker whose reconstructed trades this site tracks, spent Monday reaffirming one of his oldest convictions — the memory supercycle — while sketching out where he thinks the next, less obvious opportunity lies.

He opened by noting a strong session for the two giants at the heart of the trade: SK Hynix and Samsung Electronics, the world's dominant memory makers. His preferred vehicle for that exposure remains EWY, the iShares MSCI South Korea ETF, which he treats as a concentrated proxy on both names because their pass-through ownership structures push real exposure above the headline index weights. It is one of his highest-conviction core holdings. Alongside it he flagged a DRAM-focused vehicle, underscoring that he is still positioned long the broad memory complex.

The reasoning was familiar. He argued that if the multi-year capacity agreements now being signed across the industry extend another three to five years — and if analyst models calling for roughly 2.8-3.3x growth into 2027 prove out — the risk/reward on staying long over time remains attractive. This is the structural version of his thesis: AI-driven demand for DRAM, high-bandwidth memory (DRAM stacked beside GPUs) and NAND flash creating a durable pricing upcycle rather than a normal boom-bust.

What stood out today was where he pointed next. Serenity singled out legacy memory — older, non-cutting-edge DRAM and NAND — as a particularly interesting trade right now, driven by the potential for average selling price (ASP) hikes as supply tightens across every node. His logic is a classic supply-chain back-map: the big players don't make everything themselves. As leaders like Micron (MU), a core U.S. holding he has likened to "the next Nvidia" for its Made-in-America memory moat, discontinue certain legacy product lines, that demand has to be filled somewhere — either through alliances or via direct supply into Samsung. Smaller suppliers picking up that slack, he suggested, may carry higher earnings sensitivity to the price moves than the giants themselves.

To find those beneficiaries he offered a research method rather than names: track the wafer allocations at Powerchip Semiconductor (PSMC), the Taiwanese foundry that fabricates a large share of the industry's legacy memory and specialty chips. Following who is buying that capacity, he implied, is a way to identify the obscure companies stepping into the gap Micron and others are leaving behind. Characteristically, he declined to publicly name the specific tickers he has in mind, framing the post as a set of ideas for others to run down.

The note contained no disclosed trades or position changes — it read as a reaffirmation of the memory thesis plus a nudge toward the legacy corner of it. For a reader new to his book, the throughline is consistent: buy the memory upcycle broadly through EWY and MU, then hunt one layer upstream for the smaller, higher-torque names that the giants quietly depend on.

*Derived commentary, not investment advice.*