
USD 580, Sandisk's target price raised by 93%! Bernstein: AI is creating an unprecedented storage supercycle

Wall Street's focus on AI hardware is shifting from computing power to storage. Bernstein has significantly raised its target price for Sandisk, believing that the demand for long context in AI will drive a surge in NAND demand, initiating a storage supercycle. However, JP Morgan holds a conservative view, believing that the current high prices are a cyclical phenomenon and may return to normal after capacity expansion
The narrative on AI hardware on Wall Street is shifting from pure "computing power" to "storage capacity."
According to the Chasing Wind Trading Desk, on January 14th, Bernstein analyst Mark Newman and his team pointed out in their latest report that the AI-driven data explosion is creating an "unprecedented storage supercycle," raising SanDisk's (SNDK) target price violently from $300 to $580.
For investors, this is an important macro puzzle regarding the shift in AI infrastructure bottlenecks, but at the same time, JP Morgan's previously released bearish report also reminds us that the specter of cyclicality still lingers.
"Context" Becomes the New Bottleneck
Bernstein's core logic is based on the technological path revealed in NVIDIA's CES speech. As large models evolve towards long context and multi-turn dialogue, the traditional HBM (High Bandwidth Memory) capacity has become a bottleneck. The report sharply points out that NVIDIA's Vera Rubin architecture has officially elevated rack-level SSDs to the critical path of AI inference, with SSDs no longer just being cold storage but transforming into an "active context layer" that helps GPUs improve utilization.
"'Context is the new bottleneck, and storage must be restructured'—we believe Jensen Huang's statement at CES is a game changer for various rules in the NAND industry. Our analysis shows that this could increase the NAND demand per GPU by five times."
Price Surge Amid Supply-Demand Mismatch
While demand is exploding, supply is exceptionally restrained. Bernstein emphasizes that, apart from YMTC, there is almost no new NAND capacity being added across the industry. This extreme supply-demand mismatch is driving up prices, with the average selling price (ASP) of NAND and DRAM experiencing sharp increases. The report believes that as long as supply remains disciplined, this high-price environment will continue to support manufacturers' profits.
"AI model training and inference workloads, richer content creation, and longer data retention requirements are collectively driving this data explosion, creating an insatiable (and price-insensitive) demand for storage and memory, leading to unprecedented price increases."
Is It Prosperity or a Bubble?
However, just as Bernstein is proclaiming a "supercycle," JP Morgan earlier poured cold water on the situation. In its December report, JP Morgan only gave SanDisk a "neutral" rating, setting a target price of $235—less than half of Bernstein's target price.
JP Morgan analyst Harlan Sur believes that SanDisk's current high profits are more of a "cyclical boom" rather than a structural improvement. With SanDisk holding only a 2%-3% market share in the critical enterprise SSD market for AI, it is in a follower position, and its long-term profitability is not solid.
"Short-term excess profits are unlikely to support long-term valuation increases, and it may revert to the historical 'boom-bust' cycle pattern. As major manufacturers initiate a new round of capacity expansion around 2027, the supply-demand structure will tend to loosen." This is a classic game about "whether this time is different." Bernstein bets that AI has completely restructured the demand function for storage, believing that the market underestimates the value of "contextual memory"; while JP Morgan adheres to cyclical laws, warning that mean reversion is inevitable when capacity is released. For investors, the difference between SanDisk's target prices of $580 and $235 represents a significant gap in the current market's understanding of the sustainability of AI hardware
