DRAM vs LUMA: Which Is the Better Pick in 2026?

Overview

DRAM and LUMA are both thematic technology ETFs betting on AI-adjacent hardware infrastructure, but they target different bottlenecks. DRAM invests in memory semiconductor manufacturers—the chips that store and access data in AI systems—while LUMA focuses on photonic and optical hardware that moves data across networks and data centers. Both are nascent funds launched in 2026 with no dividend component, making them pure capital-appreciation plays rather than income vehicles.

How they differ

The core difference is supply-chain position: DRAM targets memory chips (storage and retrieval), while LUMA targets the optical infrastructure (data transmission) that feeds those chips. DRAM commands vastly larger assets at $25.9B versus $9.71M, reflecting either earlier institutional adoption or different market reception since inception. DRAM's expense ratio of 0.65% undercuts LUMA's 1.00%, a notable gap for thematic funds competing on the same theme-coverage principle. Both ETFs launched recently—DRAM in 04/02/2026 and LUMA in 07/14/2026—so neither has a track record long enough to assess how well their holdings have positioned them within the AI infrastructure rally.

Who each is best for

• DRAM: Fits investors who believe memory-chip makers will be primary beneficiaries of AI scaling, with enough conviction and time horizon to tolerate concentrated exposure to semiconductor cycles and DRAM pricing volatility.

• LUMA: Designed for investors targeting optical infrastructure as the less-crowded segment of AI hardware, or those who want to hedge memory-focused portfolios with exposure to the data-transmission layer of the stack.

Key risks to know

• Sector concentration and cyclicality. Both funds concentrate on narrow segments of semiconductor and technology hardware. Memory chips and optical components are subject to severe industry cycles—overcapacity, price compression, and demand destruction can arrive suddenly, especially if AI capex growth slows faster than markets expect. DRAM's larger base is less acute but still young enough that structural fund flows (inflows or outflows) could amplify volatility.

• Single-theme concentration risk. Neither fund diversifies across multiple AI-adjacent hardware categories. Investors holding both DRAM and LUMA may face hidden overlap in the same supply-chain companies (contract manufacturers, materials suppliers) that serve both segments—a risk that requires holdings review.

• Limited track record in a new market narrative. Both funds launched during the AI boom and have no performance history through a technology downturn or margin squeeze. It's unclear how their constituent companies will behave if AI adoption growth disappoints or capex cycles reverse.

Bottom line

If you're betting on memory semiconductors as the primary AI infrastructure winner, DRAM offers lower fees and more established investor backing; if you see optical networking as the overlooked layer, LUMA presents a smaller, less-trodden thematic bet. Neither offers income or defensive characteristics, and both carry significant execution and liquidity risk given their youth. Past performance does not predict future results.