Episode summary

AI Memory Trade's Biggest Questions Following SKHY & STX Earnings

Schwab Network · · Guests: Brendan Burke, Nitin Sacheti

AI Memory Trade's Biggest Questions Following SKHY & STX Earnings – Schwab Network

Summary

The discussion centered on why the AI memory trade remains strong after earnings from Seagate and SK Hynix, and whether the current demand surge looks more durable than past memory cycles. One guest argued that AI buildouts have turned memory into a more multi-year, contracted business, with backlog giving customers reason to keep upgrading as accelerator roadmaps evolve. On that view, pricing and demand stay supported even as more capacity comes online.

The other guest framed the key uncertainty as not whether demand is strong today, but how long the cycle can last. He said SK Hynix looks attractive because of its mid-single-digit earnings multiple and aggressive buyback, which he sees as a way to translate strong free cash flow into long-term value. He also compared SK Hynix’s setup to Seagate in 2011, when buybacks helped sustain earnings power.

Both guests pushed back on the idea that current shortages will quickly become oversupply. They argued that continued innovation in HBM, DRAM, NAND flash, and hard drives should keep demand above supply through 2030, while inference spending may prove more durable than training if AI capital spending tightens.

Key takeaways

  • The guests said AI memory demand has become more contractual, with backlog pushing customers toward next-generation upgrades rather than spot buying.
  • One guest expects DRAM pricing to stay strong in 2027 and said demand could remain above supply every year through 2030.
  • The guests argued oversupply is unlikely soon because HBM, DRAM, NAND flash, and HDD demand are all still expanding.
  • Nitin Sacheti said SK Hynix looks compelling because it trades at a mid-single-digit earnings multiple and is aggressively buying back stock.
  • The panel said inference spending may outlast training spend if financing gets tighter, supporting Western Digital and Seagate as long-tail beneficiaries.

Notable quotes

“Memory is now in a backlog. That means that customers are going to be looking for the next generation of memory improvements”

— A speaker on Schwab Network at 0:38

“I think we can never really know whether that lasts, at least from my perspective, that lasts throughs obviously been a lot in ter jusd of incremental data center CapExet funded from private credit”

— A speaker on Schwab Network at 1:34

“when you're trading at a mid-single digit earnings multiple and you're essentially, you know, using all of that free cash flow to buy back stock or pay divid”

— A speaker on Schwab Network at 2:26

“I don't see that coming, Jenny, because there's a continued innovation in HBM system, Dram, Nand flash”

— A speaker on Schwab Network at 3:42

“if we just basically froze at fable 5.1. ws of AI, you know I'm sure you know, we're all using muse here”

— A speaker on Schwab Network at 5:26

Topics

  • ai memory demand cycle
  • memory chip earnings
  • chip supply and shortages
  • customer upgrade backlog
  • buybacks and valuation

Companies and tickers mentioned

  • SKHY
  • STX
  • SK Hynix
  • Seagate

Watch the episode

Watch “AI Memory Trade's Biggest Questions Following SKHY & STX Earnings” on YouTube

Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.

More from Schwab Network

All Schwab Network episodes →

Schwab Network, this episode and its artwork belong to their owners. Octagon is not affiliated with the show; this page is an independent summary.