HBMX delivers concentrated exposure to companies developing high-bandwidth memory (HBM) technology — the specialized chips that make AI computing possible. This ETF bets that memory makers, not just AI chip designers, will capture outsized profits as data centers race to build more powerful AI systems.
How It Works
The fund holds a concentrated portfolio of companies across the HBM supply chain, from memory manufacturers like SK Hynix and Micron to equipment makers enabling next-generation memory production. Unlike broad semiconductor ETFs that dilute exposure across dozens of chip types, HBMX focuses specifically on the memory bottleneck constraining AI advancement. The concentrated approach means individual positions can exceed 10% of the portfolio.
Key Features
- Pure-play exposure to HBM memory makers who control AI's critical bottleneck
- Concentrated portfolio of 15-25 names versus 50+ in typical semiconductor ETFs
- Captures the full HBM value chain from chip makers to specialized equipment suppliers
Risks
- Memory pricing cycles can crush profits — HBM makers lost 50%+ in previous downturns
- Extreme concentration means a single company's stumble could tank 10-20% of the fund
- Technology risk if alternative AI architectures reduce dependence on HBM memory
Who Should Own This
Built for investors who believe memory, not processors, is the real AI infrastructure play and want concentrated bets rather than diluted semiconductor exposure. Best suited as a 2-5% satellite position for those already overweight tech who want to sharpen their AI thesis beyond buying NVDA.