They said semiconductors were peaking out... The real variable the market missed, the HBM cycle isn't over yet | Kim Jang-yeol, Unistory Asset Management Research Center Head [Today's Stock Focus]

Watch on YouTube ↗  |  June 05, 2026 at 11:00  |  39:16  |  3PRO TV (삼프로TV)
Speakers
Kim Jang-yeol — Reporter, The Bell

Summary

Kim Jang-yeol, Head of Unistory Asset Management Research Center, discusses the recent SoCAM-related selloff in memory stocks, arguing it is a misinterpretation and presents a buying opportunity for Samsung Electronics and SK hynix. He explains why the semiconductor cycle is not peaking early, citing AI data center demand, long-term contracts, and HBM die penalty. He also recommends buying Alphabet and Broadcom as core AI infrastructure plays, while cautioning about AMD's high valuation.

  • SoCAM news about LPDDR reduction was misinterpreted; actual driver is CPU shortage and strong demand.
  • AI data center demand accounts for ~60% of DRAM, making the cycle less price-sensitive than past consumer-driven cycles.
  • HBM production creates die penalty, constraining general DRAM supply and further supporting prices.
  • Alphabet's capital raise signals aggressive AI investment; Kim recommends buying Google.
  • Broadcom is a second AI infrastructure pick worth buying below $400.
  • AMD's high P/E warrants caution despite strong growth.
  • Market is not at bubble extremes yet, but turbulence and selectivity are advised.
  • FOMC meeting on June 18 and Micron earnings on June 25 are key near-term catalysts.
Trade Ideas
Kim Jang-yeol Reporter, The Bell 9:35
Buy memory on SoCAM misinterpretation
The SoCAM news that LPDDR per CPU is halved was misinterpreted as bearish for memory. In reality, it is driven by extreme CPU demand and DRAM shortage; Nvidia is splitting the memory to produce more CPUs and fulfill strong AI server demand. This does not reduce total memory TAM and the selloff creates a buying opportunity for Samsung Electronics and SK hynix, supported by long-term AI data center contracts and HBM die penalty that constrains general DRAM supply.
Kim Jang-yeol Reporter, The Bell 26:21
Buy Google as AI infrastructure core
Alphabet (Google) is a core AI infrastructure asset. Its recent debt and equity capital raise signals strong confidence in AI investment. With a full stack including TPU, data centers, models, and services, Alphabet is positioned to be a low-cost AI provider, similar to Apple's ecosystem. He explicitly states he would buy Google.
Kim Jang-yeol Reporter, The Bell 26:24
Buy Broadcom if below $400
Broadcom is his second AI infrastructure pick after Google. After a pullback, the P/E has dropped to 38x, and he would consider buying if the stock falls below $400, viewing it as an attractive entry point into a key semiconductor and networking player for AI.
Kim Jang-yeol Reporter, The Bell 35:57
Watch AMD for high valuation
AMD trades at a high P/E of 40x+ for 2026 earnings, and while growth is strong, the valuation is rich. He cautions that investors should be careful when assigning 30-40x multiples, implying the stock is worth watching rather than aggressively buying at current levels.
Up Next

This 3PRO TV (삼프로TV) video, published June 05, 2026, features Kim Jang-yeol discussing 005930.KS, 000660.KS, GOOGL, AVGO, AMD. 4 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Kim Jang-yeol  · Tickers: 005930.KS, 000660.KS, GOOGL, AVGO, AMD