Wall Street is recalibrating its expectations for the AI infrastructure trade, shifting focus from the processors that power the intelligence to the storage that remembers it. SanDisk (SNDK) has emerged as a primary beneficiary of this shift, with analysts now positioning the company as the AI boom’s pure-play NAND winner due to its concentrated exposure to flash memory pricing.
The momentum is backed by a surge in demand from hyperscalers who are aggressively ramping enterprise solid-state drives (SSDs) to support massive AI training clusters. This demand is driving NAND average selling prices higher and faster than previous forecasts suggested, pushing SanDisk shares to a fresh all-time high.
The bullish sentiment reached a peak this week as two prominent firms issued aggressive price targets. Mark Newman of Bernstein raised his target for SanDisk to $1,250 from $1,000, maintaining an Outperform rating. Simultaneously, C.J. Muse of Cantor Fitzgerald increased his target to $1,000 from $800, keeping an Overweight rating. Both analysts pointed to the acceleration of NAND pricing as the primary catalyst for the move.
For investors, the appeal lies in the “pure-play” nature of the business. Unlike diversified memory giants, SanDisk’s earnings are tied almost exclusively to the storage market, meaning every percentage point increase in NAND pricing flows more directly to the bottom line without being diluted by other product lines.
The Pure-Play Advantage: SanDisk vs. Micron
To understand why SanDisk is being singled out, it is necessary to look at the structural difference between it and peers like Micron Technology. Micron operates a broad portfolio that includes DRAM, NAND, and high-bandwidth memory (HBM). While this diversification provides a safety net, it also creates a “dilution” effect during specific price spikes.
Micron recently posted record fiscal Q1 DRAM revenue of $10.8 billion, a 69% increase year over year. Yet, its NAND revenue grew by a more modest 22% to $2.7 billion. In contrast, SanDisk’s focused strategy allows it to capture the full force of the storage surge. In its fiscal Q2 2026 report, SanDisk saw total revenue jump 61% year over year to $3.03 billion.
The impact is most visible in the data center segment. SanDisk’s data-center revenue hit $440 million, representing a 76% year-over-year increase. This growth is fueling a significant expansion in profitability, with gross margins climbing toward 50% as the company leverages the pricing tailwind.
| Metric | SanDisk (Pure NAND) | Micron (Diversified) |
|---|---|---|
| Q2/Q1 Revenue Growth (YoY) | +61% | +69% (DRAM focus) |
| Data-Center Revenue Growth | +76% | Not specified (Mixed) |
| Trailing 12-Month Revenue | $8.93 Billion | >$20 Billion (Run-rate) |
Deconstructing the $1,250 Price Target
The Street-high call from Bernstein is not based on speculation, but on a specific earnings model. Mark Newman’s base case applies an 11-times multiple to average non-GAAP earnings of $114 per share across fiscal 2026 through 2029. For fiscal 2027 alone, Newman projects earnings of $144 per share.
The potential for further upside exists in what Newman describes as a “bull case.” If NAND pricing remains sustained at these elevated levels, he models earnings of $224 per share, which could push the stock’s valuation north of $3,000 in a blue-sky scenario. This projection suggests that the stock’s recent rally—which includes a 2,228% gain over the past year—may still have significant room to grow if the AI capex cycle holds.
Market reaction was immediate. SanDisk shares closed at $851.57, rising 9% on volume of 19.8 million shares, more than double the 30-day average. This volume indicates strong institutional conviction behind the new price targets.
Risk Factors and the Memory Cycle
Despite the optimism, the memory market is historically volatile. The “cycle” is the defining characteristic of the industry; prices can turn on a dime when supply finally catches up to demand. SanDisk’s trailing 12-month EPS currently shows a $7.49 per share loss, a stark reminder of the weakness the company faced before the current AI-driven rebound.
There are two primary headwinds that could pressure the stock by late 2027:
- Capex Slowdown: If hyperscalers reduce their spending on AI training clusters, the artificial demand for high-end enterprise SSDs could evaporate.
- New Capacity: The introduction of new fabrication plants (fabs) could increase supply, potentially capping the price growth analysts are currently modeling.
However, current data suggests these risks are not immediate. Most analysts believe that NAND shortages will persist into 2028, providing a multi-year runway for the current pricing trend to continue.
What This Means for AI Infrastructure
The focus on SanDisk signals a broader realization in the market: AI is not just a compute problem, but a data problem. As models grow in size and complexity, the require to store and retrieve massive datasets with minimal latency becomes critical. This makes NAND flash—the technology that allows for high-speed, non-volatile storage—as essential as the GPUs themselves.
For investors, the “pure-play” leverage means that SanDisk acts as a high-beta bet on the storage layer of AI. While a diversified company like Micron offers stability, SanDisk offers a more direct correlation to the pricing of NAND flash. With a current market cap of $125.7 billion across 147.6 million shares, the company is now a heavyweight in the semiconductor space, though it remains susceptible to the inherent volatility of the memory cycle.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Investing in semiconductors involves significant risk due to market volatility and cyclicality.
The next major catalyst for SanDisk will be the release of its Q3 revenue figures. With analysts expecting revenue to rise more than 50% sequentially at the midpoint of guidance, the upcoming filing will confirm whether the NAND pricing acceleration is sustainable or reaching a plateau.
We desire to hear from you. Do you believe the “pure-play” strategy is a better bet than diversification in the current AI climate? Share your thoughts in the comments below.
Related reading
