Generated July 2026 from current fund data.
Overview
These four ETFs zero in on memory semiconductors—DRAM, NAND, and high-bandwidth memory (HBM)—and the supply chains that feed AI infrastructure. DRAM and KMEM are buy-and-hold growth vehicles with no distributions, while DRMP pursues weekly income via put spreads on memory stocks and HBMX seeks annual appreciation through concentrated holdings. The critical divide: DRMP generates 38.88% yield but uses options overlay to do it; the others rely on stock picking and price appreciation alone.
How they differ
DRMP stands apart by deploying a systematic put credit spread strategy to generate weekly distributions, targeting a 38.88% distribution rate—a synthetic income approach fundamentally different from the other three. DRAM and KMEM both charge 0.65% expense ratios and distribute nothing, favoring pure capital gains; HBMX, actively managed for concentrated long-term gains, charges 0.95% and distributes annually. Scale matters too: DRAM holds $23.0B in assets, while HBMX has $40.6M, KMEM has $25, and DRMP has $6.41M—a massive gap in liquidity and fund maturity that carries operational risk for smaller vehicles.
Who each is best for
DRAM: Fits investors seeking broad thematic exposure to memory semiconductors with minimal costs, zero distribution drag, and the liquidity and track record of a $23.0B fund.
DRMP: Fits investors with high current income needs who are comfortable with weekly distributions and understand that put spreads on concentrated sectors carry leverage risk and may not fully hedge downside moves.
HBMX: Fits investors who believe concentrated, active stock picking in memory and HBM producers will outperform, accept annual distributions, and tolerate the higher idiosyncratic risk of a $40.6M fund with a tighter holdings list.
KMEM: Fits investors seeking memory semiconductor exposure through a basket approach at a 0.65% cost, though the $25 AUM signals an extremely early-stage fund with minimal operating history and no proven ability to execute at scale.
Key risks to know
- DRMP's options overlay risk: Put credit spreads generate income by selling downside protection. A sharp drop in memory stocks forces the fund to absorb losses on those positions while still obligated to pay distributions, risking rapid NAV erosion and capital loss for shareholders.
- NAV erosion and synthetic yield risk in DRMP: A 38.88% distribution rate on equities almost certainly relies on return-of-capital treatment and declining NAV per share over time; the put spread income cannot sustainably replace 39% annual equity returns without material principal bleed.
- Concentration and single-sector risk across all four: Memory semiconductors are cyclical and driven by commodity pricing, capex cycles, and geopolitical supply constraints. A downturn in AI spending or memory chip oversupply hits all four funds severely; no diversification backstop.
- Liquidity and track record risk for KMEM and HBMX: Both are newly launched (June 2026) micro-cap funds with minimal assets under management. Execution risk, index rebalancing slippage, and the ability to scale operational infrastructure are unproven.
- Active management and concentration in HBMX: Narrow focus on HBM, advanced packaging, and equipment vendors makes performance highly dependent on manager stock picks in a small investable universe; underperformance can persist if the fund's bets disagree with sector momentum.
Bottom line
If you want broad, liquid, low-cost exposure to memory semiconductors without distributions, DRAM's $23.0B scale and 0.65% expense ratio offer simplicity. If you crave weekly income and can tolerate options risk and NAV decay, DRMP's 38.88% yield appeals—but verify you understand put spread mechanics and can stomach capital drawdowns. HBMX and KMEM bet on active or basket-based stock picking in a nascent, micro-cap fund structure; both lack the track record to validate their approach. The core tradeoff is income and active selection versus capital preservation and passive exposure. Past performance does not predict future results.
AI-generated analysis for educational purposes only. Verify important details independently; past performance does not guarantee future results.
