SanDisk: The 1,000% Surge Miracle After the Spin-Off

In February 2025, Western Digital spun off its NAND flash business into the independent company SanDisk (NASDAQ: SNDK). The spin-off price was around $36, and by June 2026 it had soared to approximately $2,180—a gain of more than 1,000% in just over a year.

On June 18 alone, SNDK surged more than 6%, closing near a 52-week high, fueled by the AI storage super-cycle narrative and a broad tech rally.

"SanDisk is experiencing its own 'NVIDIA moment.'" — This single sentence captures the market's狂热 sentiment toward SNDK.

But the question is: With a P/E of 74.6x and the stock soaring from $36 to $2,180, is this rally sustainable? Let me break it down using the methods of a Wall Street research analyst.

*The storage demand of AI servers—every AI training cluster requires massive SSDs, and this is SanDisk's growth engine.*

I. Core Business Segment Breakdown

1. Enterprise AI Data Storage

Core Products:

• AI Server SSDs: High-capacity (30TB+) NVMe SSDs for AI training data loading
• HBF (High Bandwidth Flash): Proprietary high-bandwidth flash, an NAND equivalent of HBM, targeting AI inference
• Large Model Training Cluster Storage: Distributed storage solutions supporting thousand-card GPU clusters

Long-Term Agreements (LTAs):

SanDisk has locked in a minimum of $42 billion in contracted revenue. Major customers include cloud giants such as Microsoft and Google, with contract terms of 3-5 years. This gives revenue a high degree of predictability.

Capacity Expansion Plans:

• Joint-venture wafer fabs with Kioxia (Yokkaichi and Iwate, Japan)
• Investment in advanced packaging technology
• U.S. domestic capacity build-out

2. Consumer Storage

Demand Drivers:

• AI PC refresh cycle → demand for larger-capacity SSDs
• Smartphone storage capacity upgrades (128GB → 256GB+)
• Drones and action cameras → high-speed memory cards
• Game consoles (PS5/Xbox) → NVMe SSD expansion

Competitive Landscape

SanDisk's unique positioning: 1) A pure NAND play (unlike Samsung/Hynix/Micron, which have DRAM for diversification); 2) Proprietary HBF as a technology moat; 3) $42B in locked-in LTAs = highest revenue visibility

II. Deep Dive on Financial Fundamentals

Four Consecutive Quarters of Trends

Key Valuation Metrics

*A P/E of 74x—the market has already priced in the best-case scenario for the next 2-3 years. The question is: is it enough?*

Is 74.6x P/E high? In absolute terms, outrageously so. But factoring in: 1) revenue +251%, 2) gross margin still expanding, 3) NAND sold out through 2026, and 4) $42B in LTAs—this P/E reflects the "super-cycle peak" expectation.

CapEx and R&D

Shareholder Returns

• As a newly spun-off company, it currently does not pay a dividend, focusing on growth investment
• No large-scale buyback program at this time
• All cash flow is allocated to capacity expansion + HBF R&D

III. Core AI Growth Drivers

1. AI Server Storage: A Hard Requirement

AI GPUs need HBM for short-term memory, but training data + model weights + checkpoints require massive SSD capacity for long-term storage. For every H100 added, several TB of NVMe SSD is needed on the back end.

2. HBF (High Bandwidth Flash) Technology Moat

HBF is SanDisk's secret weapon:

• Combines NAND Flash's large capacity with HBM-like high bandwidth
• Targets AI inference: model loading speed >> traditional SSDs
• Proprietary packaging technology that competitors cannot easily replicate in the short term
• Strong product premium → gross margin higher than ordinary NAND

3. Long-Term LTAs

SanDisk's LTA model resembles Micron's HBM lock-ins:

• Cloud giants sign 3-5 year supply contracts in advance
• Prices are semi-fixed (with periodic adjustments)
• Minimum guaranteed revenue of $42 billion
• This dramatically reduces the impact of memory-cycle volatility on profits

4. AI Storage Incremental Forecast

IV. Industry Competition and Core Risks

🟢 Bull Case Logic

1. NAND 2026 Capacity Sold Out

SanDisk management says NAND supply-demand tightness will persist into 2026 and beyond. Kioxia has confirmed that NAND + SSD for the full year of 2026 is sold out. Customers are already seeking 2027 supply guarantees.

2. NAND Spot Prices Have Doubled

Over the past six months, NAND Flash spot prices have doubled. TrendForce forecasts 2026 Q2 NAND contract prices up 70-75% QoQ.

3. AI Storage Super-Cycle

AI servers' storage demand is 8-15x that of traditional servers. A global AI CapEx explosion → SSD demand explosion → SanDisk benefits directly.

4. $42B in LTAs = Revenue Floor Supported

Even if flash prices pull back, the minimum revenue guaranteed by long-term contracts limits SanDisk's downside risk.

5. HBF Technology Monopoly

Proprietary high-bandwidth flash has no direct competitor, giving the product strong pricing power.

🔴 Bear Case Risks

1. 74.6x P/E = A Cyclical Top Signal

Historically, peak P/E in the memory industry tends to appear at the cyclical top—because EPS is highest at the top. Once NAND prices fall, EPS will collapse and P/E will balloon mechanically → the stock will crater.

2. Samsung / Kioxia New Capacity Shock

• Samsung is aggressively expanding NAND capacity
• Kioxia's Iwate new fab is slated to come online in 2027
• Once the three giants complete their expansions simultaneously, supply-demand tightness can reverse quickly

3. Cloud Vendor AI CapEx Contraction

If Microsoft / Google / Meta trim AI infrastructure spend (even just a slowdown in the growth rate), NAND demand could cliff-dive.

4. Weak Consumer Electronics Demand

Consumer NAND accounts for ~35% of SanDisk. If the global economy slows, the smartphone + PC refresh cycle may underperform expectations.

5. Valuation Bubble Correction Risk

SNDK is up 1,000% in a year—historically, such rallies rarely sustain. Once market sentiment turns, the pullback could be severe.

6. Supply Chain Cost Inflation

Wafer foundry costs, advanced packaging materials, and power supply—each one is rising.

V. Scenario-Based Valuation Forecast

Valuation Assumptions

Price Target Calculation

⚠️ Critical Warning: Based on the above calculation, even under the bull scenario, SNDK's target price ($1,760) is below the current stock price ($2,180). This means the market has already priced in expectations even more optimistic than "bull."

VI. Final Trading Conclusions

Short-Term Speculation (1-3 Months): ⚠️ Extremely High Risk

Short-Term Actions:

• If holding → Strongly recommend setting a trailing stop (15-20%)
• If not holding → Chasing at $2,180 is not recommended
• Options strategy: Consider selling far-month deep ITM calls to lock in profits

Long-Term Allocation (1-3 Years): ⚠️ Cautious—Wait for a Major Pullback

Long-Term Actions:

• Establishing a position at the current price is not recommended—all scenarios imply negative expected returns
• Wait for a pullback to the $800-1,000 range before reassessing
• If it pulls back to below $500, consider scaling in (the neutral scenario still implies +60% upside)

Who It's Suitable For (and Not)

Key Items to Track Going Forward

Conclusion: The Double-Edged Sword of the Super-Cycle

SanDisk's story is exhilarating—NAND sold out, AI storage exploding, $42B in LTAs, an HBF technology moat. All of this is real.

But a P/E of 74.6x means the market has already priced in all the good news—and more. History tells us:

**Memory stocks' peak P/E always appears at the cyclical top. Because EPS is highest at the top, P/E *looks* "reasonable." But once prices fall and EPS collapses, you'll realize "cheap" was an illusion.**

SNDK is a good company, but "a good company" and "a good stock" may not be the same thing at this moment.

The most conservative advice: Admire the rally, but don't chase at the top. Wait for the cycle to pull back and valuations to return to a reasonable range before becoming a long-term investor.

"Buy when no one wants it, stand aside when everyone wants it."—That is the essence of investing through the memory cycle.