Looking at the user's investment strategy, they want to invest in suppliers to AI/GPU needs that are more downstream/esoteric - not power companies, cloud computing, or NVIDIA, but companies that provide components specifically needed for LLM use cases like memory or other specialized components.
Current Portfolio Analysis:
1. MU (Micron Technology): Memory manufacturer - directly aligned with strategy (HBM memory critical for AI)
2. VRT (Vertiv Holdings): Infrastructure/cooling - somewhat aligned but might be close to power/infrastructure
3. MRVL (Marvell Technology): Semiconductor/networking chips - aligned (makes AI accelerators and networking chips)
4. AMKR (Amkor Technology): Chip packaging/testing - well aligned (advanced packaging critical for AI chips)
5. AEHR (Aehr Test Systems): Semiconductor test equipment - well aligned (tests AI chips)
All positions are short-term (held less than a year based on tax data showing only short-term gains/losses).
Market Assessment:
- Markets down today (NASDAQ -1%, S&P -0.53%)
- VIX up 4.82% suggesting some volatility
- Tech sector seems to be under pressure
Recent News Analysis:
- SK Hynix betting $13B on AI memory dominance - validates the memory thesis
- Micron (MU) closing Crucial consumer division - focusing on enterprise/AI memory
- AMKR fell 6.31% - more than market
- MRVL fell 2.18%
- Reports of China blocking Nvidia H200 chips could create uncertainty
Portfolio Performance:
- Down 1.8% overall
- VRT is the only winner (+4.5%)
- AMKR biggest loser (-6.9%)
- $20,000 cash available
Tax Considerations:
- All positions are short-term
- Have $2,149.83 in short-term losses vs $712.07 in gains
- Could harvest some losses for tax efficiency with no wash sale concerns
I'm noticing the portfolio is well-positioned for the AI infrastructure thesis, but there's room to optimize. AMKR's 6.9% drop presents a tax-loss harvesting opportunity that could be redeployed into other AI component plays like interconnects or optical components. VRT might be straying too far from the core thesis toward general infrastructure. I should consider holding MU and MRVL as they're core to the strategy, while potentially trimming or repositioning VRT.
For new positions, I'm looking at companies focused on memory interfaces, optical interconnects for data centers, and specialized testing equipment—specifically RMBS for HBM interfaces, LITE for optical components, ONTO for advanced metrology, and FORM for probe cards. Given current market conditions and existing deployment, I'd deploy about 60% of available cash while harvesting losses from AMKR and taking some profits from VRT.