The investment seeks long-term capital appreciation... Show more
KSMH is an actively managed, non-diversified equity ETF that provides targeted exposure to South Korea's semiconductor ecosystem, with an emphasis on companies enabling artificial intelligence hardware and infrastructure. Listed on Nasdaq and launched in August 2026, the fund uses swaps and direct holdings to build a concentrated portfolio of roughly 20 to 25 positions spanning memory, chip equipment, packaging, substrates, and testing. Its net expense ratio is 0.65%.
The portfolio is dominated by the two global leaders in high-bandwidth memory (HBM), the stacked DRAM used in AI accelerators: Samsung Electronics and SK hynix, which together represent roughly 40% of assets. Smaller but meaningful positions include Hanmi Semiconductor, Samsung Electro-Mechanics, DB HiTek, Leeno Industrial, and Isupetasys. Because the fund is both single-country and single-sector, its performance is tightly linked to the memory cycle and investor sentiment toward AI infrastructure spending.
Over the last 30 days, KSMH has fallen approximately 21%, from a level near $25.45 to around $20.16. The decline was not gradual: the fund peaked near $30.88 in early October and gave back roughly 35% from that high, with the sharpest leg occurring in a single November session that took shares down more than 17%.
Because the fund is newly launched, its "quarter" is effectively its life-to-date record. From its August inception near $24.63, KSMH initially climbed about 25% as AI-memory enthusiasm peaked, then reversed into a volatile downtrend that now leaves it about 18% below its launch level. The pattern has been characterized by rapid moves in both directions rather than a steady, trendless range.
The last 30 days compressed several distinct shocks into a short window. First, the broader AI infrastructure trade began to de-rate as investors questioned the sustainability of debt-financed data center buildouts and whether enormous capital spending would translate into durable returns. Hotter-than-expected US inflation revived concerns that borrowing costs could stay elevated, pressuring high-multiple technology assets.
Second, China's ChangXin Memory Technologies (CXMT) reported progress on a denser DRAM architecture, fueling fears that memory pricing power and HBM leadership could eventually face new competition. This hit Samsung Electronics and SK hynix directly, given their dominance of the memory market.
The decisive move came in early November, when a Democratic sweep in US midterm elections raised expectations of tighter oversight of data centers and AI development. The tech-heavy Nasdaq suffered its largest one-day drop in more than a year, and Korean markets followed: the Kospi fell more than 8%, triggering a temporary trading halt, while SK hynix and Samsung slumped sharply. KSMH's concentrated, single-country structure meant the fund absorbed these moves nearly one-for-one, producing its steepest daily decline.
The three-month pattern reflects a complete reversal of sentiment toward Korea's AI-memory trade. In August, the fund launched into a market still celebrating record memory profits, HBM sold-out through long-term agreements, and a large share-buyback program at SK hynix. Foreign capital had poured into the same handful of Korean semiconductor names, and KSMH rode that momentum to its early-October peak.
The unwind that followed was structural as well as headline-driven. Samsung and SK hynix approached completion of combined buyback programs, removing a steady marginal buyer just as foreign investors turned net sellers of Korean equities. Earnings remained strong, yet shares fell anyway, signaling that investors had shifted from rewarding growth to questioning how long pricing power and the AI spending cycle could persist. This de-rating, rather than any deterioration in current chip demand, has been the dominant force over the quarter.
Tickeron's AI Screener is an AI-powered stock and ETF discovery platform that helps investors scan thousands of securities using technical indicators, fundamentals, volatility metrics, price patterns, industry filters, AI-generated signals, and performance characteristics. Rather than reviewing tickers one at a time, users can apply customizable criteria to surface securities matching specific conditions, from breakout candidates to sector trends. The tool is designed to make systematic screening faster and more consistent than manual research, supporting both retail and professional workflows. Investors can explore the platform's capabilities through the AI Screener to identify opportunities aligned with their own criteria.
The path ahead for KSMH depends less on current chip demand than on whether the market regains confidence in the durability of AI infrastructure spending. Key factors include whether major cloud providers maintain or trim capital-spending guidance, the trajectory of US inflation and interest rates, and any concrete regulatory action on data center development in Texas and Georgia following the midterms.
On the supply side, investors should monitor memory pricing and the HBM4 ramp, along with whether CXMT's technology progress translates into commercial scale. The completion of Samsung and SK hynix buyback programs also removes a source of support that had helped steady these shares. Because KSMH concentrates roughly 40% of its assets in two names within one cyclical sector, single-stock and single-country risk will remain a defining characteristic, and the fund's modest assets under management (AUM) add liquidity considerations that may influence trading costs.
The information on this webpage is provided for general informational and educational purposes only and is not intended as investment advice, a recommendation to purchase or sell any security, or an offer or solicitation related to investments. It does not consider your personal financial situation, goals, or risk profile, and all investing carries inherent risks, including the possibility of losing your entire investment. For more details, please review our full disclaimer.