Roundhill's DRAM ETF notches $6B on memory demand, here’s their next big idea
Speakers
Dave Mazza
— CEO, Roundhill Investments
Drew Pettit
— Research Director of U.S. Equity Strategy and ETF Strategy, Citi
Summary
The episode covers Roundhill's DRAM ETF as a successful launch driven by AI memory demand. Guests discuss memory chips as the key bottleneck, next opportunities in optical and industrial enablers, and geopolitical hedges involving oil and US large cap growth. Emerging markets like Korea and Taiwan are favored for AI exposure.
- Memory chips are the biggest AI bottleneck with a multi-year supply-demand imbalance.
- Roundhill's DRAM ETF has seen significant inflows and performance.
- Optical networking companies (Credo, Coherent, Lumen) are the next bottleneck.
- Industrial names benefit from data center buildout but not specified as trades.
- Korea and Taiwan are preferred emerging markets for AI memory exposure.
- US large cap growth is recommended as a geopolitical hedge.
- Oil as a commodity is a tail risk hedge against Iran conflict escalation.
- Earnings momentum supports semiconductor stocks despite large price moves.
Trade Ideas
DRAM memory is a persistent AI bottleneck
Memory chips are the biggest bottleneck in the AI buildout, with a significant supply-demand imbalance that will persist for years because it takes three to five years to build new fabrication plants. The demand from data centers and hyperscaler capex is shifting memory from cyclical to structural growth, extending the imbalance into 2026-2028. This makes DRAM stocks attractively valued despite recent price momentum.
Korea and Taiwan are AI-driven EM buys
Emerging markets, specifically Korea and Taiwan, are favored because they have strong exposure to the AI memory story, which is driving earnings revisions. The narrowing focus outside the US on quality AI plays makes these countries attractive investments.
Optical networking is next AI bottleneck
Optical networking components are the next bottleneck in AI infrastructure because they enable semiconductor chips to communicate effectively. As AI demand grows and new chips like Nvidia's Vera Rubin require even more memory, optical companies will see sustained demand. Specific names include Credo, Coherent, and Lumen.
Oil commodity is tail risk hedge
Oil as a commodity is a good tail risk hedge because the market is not pricing in the extended Iran conflict scenario. Equities are not recommended for this hedge; instead, long oil directly provides protection against geopolitical escalation.
US large cap growth is geopolitical hedge
US large cap growth is the preferred geopolitical hedge because it is relatively insulated from global conflicts and energy cost shocks. With hyperscaler cash returns on investment improving, mega-cap growth acts as a defensive trade in a higher-for-longer oil environment.
