Generated August 8, 2026.
Figures quoted in this analysis are from its generation date and may lag the live snapshot table above, which always shows the latest data.
Overview
DRAM and LUMA are both thematic technology ETFs targeting AI infrastructure, but they focus on different hardware layers. DRAM invests in memory-chip companies—the semiconductors that store and access data—while LUMA targets the photonic and optical infrastructure that moves data between systems. Neither pays distributions; both are growth-oriented, not income-focused.
How they differ
The core difference is their underlying exposure: DRAM holds equity in memory-chip manufacturers (DRAM, NAND flash, and related memory companies), whereas LUMA invests in photonic and optical hardware makers across a broader, emerging supply chain. DRAM is far larger, with $23.9B in assets under management compared to LUMA's $2.15M, which reflects the maturity of the memory-chip market versus the nascency of photonic infrastructure for AI. DRAM also charges a lower expense ratio at 0.65% versus LUMA's 1.00%. LUMA includes exposure to both public and private companies, introducing valuation and liquidity complexity that DRAM avoids by holding only listed equities.
Who each is best for
DRAM: Fits investors seeking capital appreciation from established semiconductor memory demand, who are comfortable with a consolidated, mature subsector of AI infrastructure and prefer lower-cost passive exposure to a large, liquid market.
LUMA: Designed for investors with higher risk tolerance and a longer time horizon, interested in earlier-stage optical infrastructure buildout for AI, and willing to accept higher fees and significantly smaller asset base in exchange for exposure to an emerging hardware category.
Key risks to know
- Semiconductor cyclicality and oversupply risk: Memory-chip markets are historically prone to boom-bust cycles driven by capex swings and inventory imbalances. DRAM's concentrated exposure to one hardware layer amplifies this cyclical risk relative to broader chip or AI infrastructure plays.
- Photonic market immaturity and adoption uncertainty: LUMA's underlying market—optical interconnects and photonic infrastructure for AI—is early-stage. Revenue concentration among few customers and unproven demand scaling represent material execution risk that does not yet apply to DRAM's established memory-chip customers.
- Private-company valuation and liquidity in LUMA: LUMA's inclusion of private companies means some portfolio holdings lack market pricing and trading liquidity. A sharp repricing of private valuations or a market downturn affecting LUMA's size could create wider bid-ask spreads and tracking error versus the fund's stated holdings.
- AI cycle dependency: Both funds are highly exposed to AI capex timing and intensity. A slowdown in data-center buildout or AI spending would pressure both memory and optical infrastructure demand simultaneously.
Bottom line
If you're seeking exposure to a mature, well-capitalized hardware layer of AI infrastructure at low cost, DRAM's scale and 0.65% expense ratio stand out. If you believe optical and photonic infrastructure is the next critical bottleneck and you're willing to accept higher fees, illiquidity, and private-company risk in a much smaller fund, LUMA offers that concentrated bet. Past performance does not predict future results, and both funds' value depends entirely on sustained AI infrastructure investment.
AI-generated analysis for educational purposes only. Verify important details independently; past performance does not guarantee future results.
