The evolution of the AI trade plays into SK Hynix's strengths: Rayliant Global Advisors

Rayliant Global Advisors’ chief research officer outlined portfolio construction options for SK Hynix—using Korea-listed shares versus U.S. ADRs—while assessing whether growth can sustain the current enthusiasm for memory chip stocks. He also discussed where the AI trade could potentially broaden beyond memory names, framing the opportunity as an ongoing positioning decision rather than a clear new catalyst.

Analysis

The main market mechanism here is not “AI enthusiasm” in the abstract; it is whether memory turns from a scarcity trade into a capacity trade. SK Hynix has been one of the cleanest expression of HBM tightness, but once sell-side models start capitalizing multiple quarters of peak margins, upside shifts from multiple expansion to actual shipment beats and ASP durability. That makes the next 1-2 quarters critical: if hyperscaler capex keeps rising, the stock can re-rate further; if not, the name can de-rate quickly because expectations are already crowded.

Second-order winners are the picks-and-shovels behind the memory bottleneck: advanced packaging, test, substrates, and lithography/equipment suppliers. If AI broadens beyond GPUs and HBM, the incremental alpha likely migrates toward less obvious beneficiaries with more diversified revenue bases, while pure memory names become more exposed to cyclical mean reversion. Micron is the closest public peer to watch for confirmation, but it also highlights the risk that the market is pricing an industry upcycle rather than a company-specific moat.

The contrarian risk is that investors are underestimating how fast supply can catch up once Korean and U.S. fabs fully allocate capex to HBM. The trade is most vulnerable over 6-18 months, not days, because the immediate flow can stay momentum-driven even as the fundamental slope flattens. Falsifiers: any pause in hyperscaler capex, weaker HBM pricing, or guidance that implies inventory normalization by mid-year rather than tightness into next year.

Key Decisions for Investors

• If initiating exposure, use SKHYV as a tactical long only on pullbacks after post-earnings digestion; treat it as a 1-2 quarter momentum trade, not a core hold, with a thesis break on any HBM ASP compression or guide-down.
• Pair trade: long SKHYV / short MU for a 3-6 month relative-value expression if you want the cleaner HBM scarcity winner versus the more cyclical U.S. memory exposure; thesis fails if Micron’s pricing commentary turns equally constructive and narrows the mix gap.
• Use SMH or SOXX as a hedge against single-name memory beta if the goal is AI participation without relying solely on memory pricing; this reduces risk if the AI trade broadens into semicap, networking, or software rather than memory alone.
• Watch for a rotation into equipment/substrate suppliers as the better 6-18 month second-order trade; if capex remains elevated, that basket can outperform memory names once the market starts discounting supply normalization.