Sandisk Corp (NASDAQ:SNDK) has become one of the market’s biggest AI winners, but one ETF expert believes the stock’s breathtaking run may have gone too far. Asked which AI stock he’d avoid despite the hype, GraniteShares CEO Will Rhind didn’t point to a struggling company.
Instead, he singled out one of the market’s biggest success stories, arguing that SanDisk’s remarkable rally has left investors with little margin for disappointment. “This isn’t a knock on the business. NAND demand is very real,” Rhind told Benzinga in an exclusive email interview. “It’s about the setup.”
The comments offer a reminder that in AI investing, even great businesses can become difficult investments when expectations outpace reality.
Why SanDisk’s Rally Raises The Bar
Rhind noted that SanDisk has surged roughly 30-fold over the past year, making it one of the market’s strongest-performing AI-related stocks.
“When a name goes that far that fast, the expectations baked into the price leave almost no room for a stumble,” he said.
While artificial intelligence has revived demand across the memory industry, Rhind argues investors shouldn’t overlook one key difference between NAND flash and high-bandwidth memory (HBM), the premium memory powering Nvidia’s AI accelerators.
Unlike HBM, which remains in short supply and commands significant pricing power, NAND has historically been a far more cyclical business, with prices swinging sharply as supply and demand shift.
For Rhind, that history makes valuation just as important as the long-term AI narrative.
AI Doesn’t Eliminate Cycles
The broader AI boom has created enormous opportunities across semiconductors, sending investors searching for the next Nvidia. But Rhind believes not every AI-related stock offers the same risk-reward profile after such dramatic gains.
Rather than chasing companies where perfection is already reflected in the share price, he prefers businesses where AI demand can continue driving earnings without equally stretched expectations.
“The stock is up something like thirty times over the past year,” Rhind said. “Memory has always been cyclical. I’d rather own a story where the price hasn’t already assumed perfection.”
That view stands in contrast to his bullish stance on AI infrastructure names such as Nebius, which he believes still have a long runway as demand for AI compute continues to outstrip supply.
For investors, the message is clear: the next phase of the AI trade may not simply be about buying every company tied to artificial intelligence. As valuations climb, distinguishing between companies benefiting from durable structural demand and those priced for perfection could become just as important as identifying the next technological breakthrough.
Image courtesy GraniteShares
