About this episode

The episode discusses the emerging importance of memory in the AI trade and the debut of the Roundhill Memory ETF.

For a while, the artificial intelligence trade was almost comically simple: just buy Nvidia. But as the buildout of massive data centers accelerates, and the AI models running on them get more complex, some long-overlooked parts of the semiconductor world are suddenly back in view. One of the most important: memory. Dominated by a small group of global giants—SK Hynix, Samsung and Micron among them—it’s a corner of the market that’s been hard for investors to access in a targeted way. Until now. On this episode of  Trillions , Eric Balchunas and Joel Weber speak with Bloomberg Intelligence semiconductor analyst Jake Silverman to unpack the explosive debut of the Roundhill Memory ETF (DRAM), which pulled in $1 billion in just 10 days. They break down why this fund could become one of the most successful thematic exchange-traded funds ever, what actually makes memory different from other chips, and why—despite its critical role in AI—it still trades a lot like a commodity. See omnystudio.com/listener for privacy information.

Topics covered

• artificial intelligence
• semiconductors
• memory chips

More episodes of Trillions

• Will Prediction Markets Come to ETFs? · April 9, 2026 · 32 min
• That Mutual Fund Has an ETF Now · March 26, 2026 · 32 min
• The Complexity of the Oil Trade · March 12, 2026 · 30 min
• Freedom Has Paid Off For this Indie ETF · February 26, 2026 · 31 min
• Metal Mania · February 12, 2026 · 41 min
• How to Invest in Future Unicorn IPOs · January 29, 2026 · 50 min

Explore listener stats, chart rankings, contacts and more on the Trillions podcast page.