Why not buy storage stocks? Here are the answers from two industry leaders

Amid the AI infrastructure investment boom, storage chip stocks appear to benefit from the surge in computing power demand, but two renowned investors and technology analysts remain cautious. Their reasoning coincidentally points to a central judgment: high prices and high profits for storage are breeding forces destined to eliminate themselves.

Cathie Wood, founder of ARK Invest, recently explained on a podcast why she does not hold storage stocks. She believes storage is the most commoditized and cyclical link in the semiconductor supply chain, and the current sharp rise in prices is not a normal state for the technology industry but is essentially a negative signal.

Meanwhile, technology strategy analyst Ben Thompson issued an even sharper warning by analogy, stating that storage manufacturers are repeating the mistake of Iran’s blockade of the Strait of Hormuz—effective in the short term, but eventually pushing the entire industry to circumvent them in the long run.

Their assessments pose a direct challenge to current market sentiment enamored with storage stocks. Against the backdrop of rapidly expanding AI inference demand, high-bandwidth memory (HBM) suppliers' stock prices are continuously sought after, but the logic presented by these two investors reminds investors to be wary of structural risks from technical route switches and demand-side proactive avoidance.

Cathie Wood: High Prices Are a Warning, Not a Positive Signal

In a recent video, Cathie Wood directly addressed the public's doubt as to why she doesn’t hold storage stocks. She admitted that years of investment experience may have made her more alert to cyclical industries.

In her view, storage is the most commoditized part of the entire semiconductor chain and has historically seen drastic cyclical fluctuations. The current trend, where high-bandwidth memory prices have tripled, quadrupled, or even increased tenfold, is not normal for the technology sector. "Most people see this as a huge positive, but in reality, it’s a negative signal," she said.

She further substantiated this view from a cash flow perspective. She mentioned a recent chart that showed the free cash flow of chip stocks moving in the exact opposite direction from hyperscalers' free cash flow—benefiting the former and pressuring the latter. However, she emphasized that this state is temporary.

More importantly, Cathie Wood pointed out that technological innovation is actively eliminating the dependence on high-bandwidth memory from the demand side. She cited Cerebras and Groq as examples—with ARK’s venture capital fund holding shares in Groq—both of these inference chip companies design architectures that do not require high-bandwidth memory.

She likened this trend to Tesla removing cobalt from its batteries: once the price or risk of a certain supply chain element rises, engineers will find ways around it. "In inference, we are witnessing technological advancements replacing the need for high-bandwidth memory," she said.

Ben Thompson: Storage Vendors Are Making Themselves a Target

Technology strategy analyst Ben Thompson’s concerns about the storage industry focus more on competitive dynamics. He used a geopolitical analogy to describe the strategic dilemma currently facing storage manufacturers.

"I compare storage vendors to Iran," Thompson said. His logic is: the deterrent power of the Strait of Hormuz lies in always being a card that can be played. Once actually used, however, it triggers opponents’ determination to permanently bypass it. "Now they’ve played this card, and it did work. But the UAE and Saudi Arabia will build pipelines and new ports, and won’t let this happen again."

He believes the current high-price strategy of storage manufacturers is producing the same effect. On one hand, Apple is lobbying to introduce Chinese memory suppliers to break the current supply structure; on the other hand, the top optimization direction for algorithms has already become "how to reduce memory usage."

"I'm worried that storage vendors may have done the same thing," Thompson said. "No one will let themselves fall into such a passive situation regarding memory issues again." His conclusion: in the long run, by creating such a significant "target," storage vendors may eventually suffer the consequences themselves.

Two Logics, One Conclusion

Although Cathie Wood and Ben Thompson frame their analysis differently, their conclusions are highly consistent: the current strength of storage stocks is accelerating their own process of being replaced.

Cathie Wood starts from the path of technological evolution, emphasizing that innovation in inference chip architectures will systematically reduce reliance on high-bandwidth memory; Ben Thompson approaches the issue from the perspective of competitive dynamics and supply chain politics, noting that high prices and concentration will drive buyers to look for alternatives, whether through algorithmic optimization, introducing new suppliers, or redesigning system architectures.

Their judgments jointly point to a risk worth paying attention to for investors: within the AI infrastructure investment narrative, the benefit logic of storage stocks may be more fragile than the market expects, and the current boom may, in fact, be the strongest driver prompting the industry to accelerate its search for alternatives.