By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel is divided on EWY's rally, with concerns about concentration risk, potential HBM price compression, and capex moderation post-2026. While some argue it's a durable AI-driven trend, others see it as a liquidity-driven bubble or a timing call rather than a long-term trend.
Risk: Potential HBM price compression and capex moderation post-2026 could derail EWY's rally.
Opportunity: Sustained AI-driven demand and long-term contracts could support EWY's rally.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
- iShares MSCI South Korea ETF (EWY) is up 87% year-to-date through May 6, 2026, outpacing the iShares Semiconductor ETF (SOXX) which is up 68% YTD and the SPDR S&P Semiconductor ETF (XSD) which is up 65% YTD, driven by concentration in memory-chip stocks tied to AI infrastructure demand. Applied Materials (AMAT) is up 67% YTD as Korean fabs order equipment from US toolmakers participating in the same cycle.
- Korean memory-chip manufacturers are capturing more of the global AI semiconductor trade than US-focused benchmarks reflect, as hyperscalers worldwide drive demand for Korean HBM chips and the economic cycle compounds across Korean fabs and their US equipment suppliers.
- The analyst who called NVIDIA in 2010 just named his top 10 stocks and iShares MSCI South Korea fund wasn't one of them. Get them here FREE.
The AI semiconductor trade has a passport, and lately it has been stamped more often in Seoul than in San Jose. On a recent Reuters Morning Bid podcast segment titled Chip stock boom, the host pushed back on the assumption that the chip rally is a Wall Street phenomenon, pointing to Korea's market surge as evidence of something bigger.
The framing from the host: "this is not solely a demand and an AI buildout and a chip demand from Wall Street or from the US. This is all around the world." The host added that "the scramble for those hot stocks is more around in Asia certainly than it is on Wall Street itself."
The numbers back the thesis. The iShares MSCI South Korea ETF (NYSEARCA:EWY) is up 87% year-to-date (YTD) through May 6, 2026, after delivering a 95% total return in 2025 that crowned it the world's top-performing major equity market. For comparison, the iShares Semiconductor ETF (NASDAQ:SOXX) is up 68% YTD and the SPDR S&P Semiconductor ETF (NYSEARCA:XSD) is up 65% YTD.
The analyst who called NVIDIA in 2010 just named his top 10 stocks and iShares MSCI South Korea fund wasn't one of them. Get them here FREE.
That gap matters. US chip benchmarks are running hot, and Korea is still pulling ahead. The host noted Korea's 75% surge was "led by two very big stocks," without naming them on air. We will respect that gap. What is publicly disclosed by BlackRock is that the top two holdings in EWY together account for 45% of the fund, and that concentration is the engine behind the move, fueled by memory-chip pricing tied to AI infrastructure demand.
Korea's boom is part of a broader cycle. Wafer-fab equipment makers tied to Korean memory capex are participating in the same cycle. Applied Materials (NASDAQ:AMAT) shares are up 67% YTD, roughly in line with the broader chip ETFs. That is the "global story" in practice: Korean fabs order from US toolmakers, US hyperscalers buy Korean HBM, and the cycle compounds.
Four leading AI models discuss this article
"The rally in Korean equities is driven by a structural shift in memory demand that is currently priced for perfection, leaving no margin for error in the upcoming HBM supply-demand balance."
The 87% YTD surge in EWY is a classic 'second-derivative' play on the AI infrastructure cycle. While the market focuses on NVIDIA’s logic chips, the bottleneck is High Bandwidth Memory (HBM). Korea’s dominance in HBM production makes it the essential plumbing of the AI buildout. However, this rally is dangerously concentrated. With 45% of EWY tied to two names—presumably Samsung and SK Hynix—the ETF is less a diversified market play and more a leveraged bet on memory pricing cycles. If hyperscalers like Microsoft or Google hit a 'capex fatigue' wall or if supply-side capacity additions lead to an HBM glut by late 2026, the drawdown in EWY will be violent and liquidity-constrained.
The thesis ignores the 'Korea Discount' and geopolitical tail risks; if tensions on the peninsula escalate or China accelerates domestic memory production, the valuation premium currently priced into Korean fabs will evaporate overnight.
"EWY's outperformance highlights Korea's memory leaders seizing AI HBM market share, extending the global semi upcycle beyond US benchmarks."
EWY's 87% YTD gain through May 6, 2026—after a 95% 2025 return—outstrips SOXX (68%) and XSD (65%), propelled by 45% fund weight in two memory-chip giants (likely Samsung Electronics and SK Hynix), riding AI-driven HBM demand from global hyperscalers like Nvidia. AMAT's 67% rise underscores Korean fab capex spilling to US equipment makers, amplifying the cycle. This isn't just Wall Street; it's a Seoul-centric story with legs if HBM stack production ramps continue. But post these parabolic moves, memory pricing volatility looms large.
Against the bullish case: Memory cycles are notoriously boom-bust—DRAM prices crashed over 50% in 2022 amid inventory gluts—and EWY's 45% concentration in two names risks a sharper plunge if AI capex slows or US foundries localize HBM supply.
"EWY's 87% gain is driven by valuation re-rating in a concentrated two-stock portfolio, not earnings growth, making it vulnerable to HBM price normalization or demand disappointment."
Korea's 87% YTD gain on EWY masks a dangerous concentration risk: the top two holdings represent 45% of the fund, almost certainly SK Hynix and Samsung Electronics. This isn't broad-based outperformance—it's a two-stock bet on HBM pricing that has already moved 75%+. Applied Materials' 67% gain suggests US equipment suppliers are pricing in sustained Korean fab capex, but memory-chip cycles are notoriously volatile. The article conflates 'global demand' with 'Korean memory dominance,' ignoring that NVIDIA, TSMC, and US-based chip design still capture the narrative and margin premium. Most critically: HBM spot prices have already compressed 30-40% from 2024 peaks, yet EWY keeps climbing—suggesting valuation expansion, not just volume growth.
If HBM demand from hyperscalers remains structurally undersupplied through 2027 (not just cyclical), Korean memory makers could sustain pricing power and justify current multiples; meanwhile, the article's point stands—Wall Street indices underweight the actual beneficiaries of the AI capex cycle.
"EWY’s ascent is driven by a concentration in a few heavyweight memory names amid a global AI-capex cycle, making it a high-conviction trade but with outsized idiosyncratic risk to those names."
The piece argues Korea is leading the AI semiconductor cycle, with EWY up 87% YTD (as of May 6, 2026) driven by concentration in two memory-name stocks; US toolmakers like AMAT ride the same cycle. That suggests a durable, global AI-capex upcycle rather than a pure Wall Street story. Yet the rally looks fragile: EWY’s roughly 45% top-holdings concentration means a miss on Samsung Electronics or SK Hynix could derail the ETF; memory pricing cycles can snap back as supply/demand rebalances. A cooling AI compute impulse, Sino-US frictions, or tighter financing could sap capex. Korea’s export-led growth also ties risk to macro shocks. Valuations may be catching up to stretched earnings growth.
The rally could be unsustainable: a de-rating or reversal in Samsung/SK Hynix memory demand would hit EWY hard, and a slower AI capex cycle or tighter funding could sharply compress earnings—making a one-way bet risky.
"The rally in EWY is driven by speculative multiple expansion despite deteriorating HBM spot prices, signaling an unsustainable valuation bubble."
Claude, your observation on HBM spot price compression is the most critical disconnect here. If EWY is rallying while HBM margins are under pressure, we are witnessing a speculative 'multiple expansion' decoupled from underlying commodity pricing. This suggests the market is pricing in a permanent structural shift to 'AI-grade' memory premiums that may never materialize if capacity catches up. We aren't looking at a growth cycle; we are looking at a liquidity-driven valuation bubble.
"HBM spot weakness hides contract-driven earnings strength justifying EWY's valuation."
Gemini, dismissing EWY as a 'liquidity-driven bubble' overlooks HBM's bifurcated pricing: spot compression (30-40%) masks robust long-term contracts with Nvidia/Microsoft at elevated levels through 2026, per company filings. SK Hynix's HBM revenue share hit 40% last quarter; margins hold above 30%. Rally tracks earnings acceleration (consensus +45% FY26 EPS growth), not pure multiple expansion. Risk is real, but premature de-rating call.
"Long-term HBM contracts mask a 2027 cliff risk when renewal rates face pressure from normalized capex cycles."
Grok's long-term contract defense is credible but incomplete. If 40% of SK Hynix HBM revenue locks in at 2024-25 prices through 2026, what happens in 2027 when those contracts reset? The market is pricing in perpetual AI capex acceleration, but hyperscalers have already signaled capex moderation post-2026. EWY's 87% YTD gain assumes contract renewal at current or higher rates—a bet that hasn't been stress-tested against a normalized capex environment.
"2027 HBM contract resets and supply-demand normalization will erode margins, turning EWY's rally into a timing call rather than a durable AI-capex story."
Grok, your defense that HBM margins stay above 30% thanks to long-term contracts risks underestimating the 2027 reset. Even if 40% of HBM revenue is locked in at current levels, the remaining 60% will reprice as new fabs come online and memory supply catches up. A cycle normalization, plus potential price erosion from competition or softness in hyperscale capex post-2026, could compress EWY's multiple just as earnings plateau, making the rally a timing call rather than a durable trend.
The panel is divided on EWY's rally, with concerns about concentration risk, potential HBM price compression, and capex moderation post-2026. While some argue it's a durable AI-driven trend, others see it as a liquidity-driven bubble or a timing call rather than a long-term trend.
Sustained AI-driven demand and long-term contracts could support EWY's rally.
Potential HBM price compression and capex moderation post-2026 could derail EWY's rally.