This New Memory ETF Has Already Doubled. Is It Still a Buy?

The DRAM – Roundhill Memory ETF (DRAM), launched April 2, has already more than doubled as DRAM prices surge on supply constraints and accelerating HBM demand for AI/GPU chips. The fund’s top holdings—Micron, SK Hynix, and Samsung—account for over 73% of assets, with the big three seeing revenue and gross margins expand while beginning to sign longer 3–5 year HBM supply deals to improve visibility and reduce cyclicality. The article also flags added volatility/risk from the ETF’s use of leverage/total return swap derivatives (not daily leverage), positioning DRAM as a tactical way to play a potential ongoing memory supercycle.

Analysis

The market is still pricing memory like a commodity, but the shift to multi-year HBM commitments changes the mechanism: customers are paying for capacity assurance, not just bits. That should support a higher sustainable multiple for MU because the next downcycle is less likely to be a clean air-pocket in pricing and more likely to be a slower normalization in utilization and mix.

Second-order winners extend beyond the obvious memory names. NVDA benefits if HBM allocation is secured, but the cleaner way to play the theme may be the equipment and advanced-packaging complex (AMAT, LRCX, KLAC) if fabs keep spending to improve yields and back-end capacity. The underappreciated losers are memory-intensive OEMs and cloud/server builders, where input-cost inflation hits gross margin before demand visibly rolls over.

The contrarian risk is that investors are extrapolating scarcity too far. If Korean and Japanese capacity ramps faster than expected, the pricing tailwind can fade within 1-2 quarters even if contracts remain longer-dated. The key falsifiers are any sign of HBM lead-time compression, capex acceleration, or spot DRAM softness; over 6-18 months, those would matter more than today’s headline strength.

Key Decisions for Investors

• Long MU on 5-10% pullbacks; 6-12 month view for a multiple re-rate if HBM mix keeps rising. Falsify if next guide shows gross margin flattening or HBM allocation slipping.
• Pair trade: long MU / short SOXX for 1-3 months to isolate memory outperformance from broad semiconductor beta. Cover if MU underperforms the index after the next earnings cycle or if spot DRAM weakens.
• Avoid DRAM and RAM as core holdings after the move; use them only as tactical vehicles because concentration plus swap/leverage mechanics can amplify reversals if the cycle turns.
• Watch AMAT, LRCX, KLAC, and NVDA into the next capex/earnings updates; add only if Korean supply discipline stays intact and AI accelerator demand remains allocation-constrained.