Generated July 2026 from current fund data.
Overview
DRAM and HBMX are both thematic ETFs targeting semiconductor memory and AI infrastructure, but they take fundamentally different approaches. DRAM is a passive, $23.4B fund focused narrowly on memory chip producers, while HBMX is an actively managed, concentrated $30.2M fund that extends beyond memory makers to include the packaging, testing, and equipment suppliers powering the memory ecosystem. The choice hinges on whether you want broad, liquid exposure to memory stocks or a concentrated, manager-driven bet on the entire memory-infrastructure supply chain.
How they differ
DRAM tracks memory semiconductor producers with a passive approach and broad AUM; HBMX uses active management to concentrate holdings across memory, packaging, testing, and equipment companies. That structural difference means DRAM offers the simplicity and liquidity of index-like exposure, while HBMX depends on manager selection and tactical positioning.
Second, HBMX distributes annually and carries a 0.95% expense ratio, versus DRAM's 0.65% fee and no distributions. The cost difference is modest in absolute terms, but HBMX's smaller AUM of $30.2M versus DRAM's $23.4B creates wider bid-ask spreads and less trading depth—a material friction for frequent traders.
Third, DRAM's passive design means it moves with the broader memory-chip sector; HBMX's active mandate and concentration strategy could produce meaningfully different returns depending on how the manager weights memory suppliers against downstream infrastructure plays. Neither fund incurs capital-gains leakage from distributions, and both carry the risk of semiconductor cyclicality and memory-price volatility.
Who each is best for
DRAM: Fits investors seeking straightforward, liquid exposure to memory semiconductors without manager risk or the cost of active oversight. Suits allocators who want the thematic AI-memory bet without concentration in a single-digit holding count.
HBMX: Designed for investors comfortable with concentrated, actively managed positions and willing to accept lower liquidity in exchange for manager-directed exposure across the memory supply chain—from chipmakers to equipment providers. Fits those who believe the supporting ecosystem (packaging, testing, equipment) offers better risk-reward than memory makers alone.
Key risks to know
- Semiconductor cyclicality and memory pricing volatility. Both funds depend on memory-chip demand and pricing power. When memory oversupply emerges (a recurring industry pattern), prices and margins compress sharply, dragging returns regardless of passive or active structure.
- Concentration in memory chip suppliers. DRAM's narrow focus on memory makers creates significant single-sector exposure. HBMX's concentration is even tighter—it holds a smaller number of stocks—amplifying drawdowns if memory demand disappoints or a major holding faces regulatory or competitive pressure.
- Liquidity and execution friction in HBMX. With $30.2M in AUM, HBMX trades far less volume than DRAM's $23.4B fund. Larger trades or exits may face wide spreads and move the market, raising the true cost of entry and exit beyond the stated 0.95% expense ratio.
- Active-management execution risk in HBMX. Concentrated, active strategies can underperform their benchmark or sector peer through manager timing, positioning, and stock-picking missteps. Past returns do not predict future results.
- AI infrastructure hype and valuation resets. Both funds are thematic plays on AI-driven memory demand. If market enthusiasm for AI infrastructure cools or expectations shift to mature-phase growth, valuations could face sharp compression.
Bottom line
If you want broad, low-friction access to memory semiconductors, DRAM's passive structure and $23.4B scale offer simplicity and liquidity at a 0.65% cost. If you're convinced the real edge lies in the suppliers and service providers surrounding memory makers—and you're comfortable with a smaller fund, wider spreads, and manager-dependent returns—HBMX's concentrated active approach may justify the 0.95% fee. Past performance does not guarantee future results.
AI-generated analysis for educational purposes only. Verify important details independently; past performance does not guarantee future results.
