Why Two Top Investors Are Avoiding Memory Stocks Despite the AI Boom
TREE NEWS reports: In the midst of a massive AI infrastructure spending spree, memory chip stocks have been among the biggest beneficiaries, with high-bandwidth memory (HBM) prices soaring threefold, fourfold, or even tenfold. Yet two prominent investors—Cathie Wood of ARK Invest and tech strategist Ben Thompson—are publicly cautioning against the sector. Their reasoning converges on a single, contrarian thesis: the very high prices and profits in memory are sowing the seeds of their own destruction.
What Happened
Cathie Wood recently explained on a podcast why she does not hold memory stocks. She argued that memory is the most commoditized and cyclical segment of the semiconductor industry, and that the current price surge is not a sign of health but a warning signal. Meanwhile, Ben Thompson, in a sharp analogy, compared memory makers to Iran blockading the Strait of Hormuz—effective in the short term, but ultimately forcing the world to find alternative routes.
Market Impact Analysis
Stocks: The immediate impact is on semiconductor and AI-related equities. Memory names like SK Hynix, Samsung, and Micron have rallied on AI demand, but these warnings could trigger profit-taking or a re-rating if investors begin to question the sustainability of memory pricing. Conversely, AI chip designers that avoid HBM, such as Cerebras and Groq, could gain attention as alternatives.
Bonds: The broader bond market is unlikely to react directly, but if memory prices fall, it could ease inflationary pressures in the tech supply chain, potentially supporting duration-sensitive bonds.
Crypto: The impact on crypto is indirect. However, if AI infrastructure spending slows due to a memory crunch, it could affect the broader tech sentiment that often correlates with risk assets like Bitcoin.
Commodities: Memory chips are not commodities in the traditional sense, but the shift away from HBM could reduce demand for certain raw materials like cobalt, which is used in some memory manufacturing. Wood’s comparison to Tesla removing cobalt from batteries highlights this substitution effect.
Currencies: For Asian exporters like South Korea and Japan, memory is a major export. A decline in memory prices or demand could weigh on their currencies, particularly the Korean won and Japanese yen.
Why This Matters for Investors
The debate underscores a critical risk in the AI trade: the assumption that memory demand will continue to grow linearly with compute. Wood and Thompson argue that engineering innovation and supply chain politics will actively work to reduce reliance on memory, especially HBM. For investors, this means:
- Diversification: Don’t concentrate AI bets solely on memory suppliers; consider chip designers with alternative architectures.
- Monitor Price Signals: If memory prices start to fall, it could signal the beginning of a cyclical downturn.
- Watch for Substitution: Keep an eye on companies like Apple pushing for new suppliers, and algorithm optimization trends that reduce memory usage.
In the short term, memory stocks may continue to rally on strong earnings, but the long-term structural risks highlighted by these two voices are worth heeding.
