₩1M Became ₩33M, SanDisk's Wild One-Year Run
📅 Published: 2026-05-09 · 🏢 Company: SanDisk (Ticker: SNDK, NASDAQ) · 💰 Price: $1,562.34 (~₩2.267M) · 🚀 +46.8% in 9 sessions post-earnings, fresh 52-week high · 1-year +3,685% surge
🎯 One-Line Summary
"The USB-stick maker landed on an AI-era money mountain" 🚀
SanDisk reported its quarterly results on April 30. Revenue was 3.5x last year (+251%), beating estimates by 26%. The company also said next quarter's revenue will grow another 33%. The stock has run +46.8% in 9 sessions since.
🏢 What Does SanDisk Actually Do?
One line: it's the semiconductor company that makes the "memory chips (NAND)" that live inside computers, smartphones, and servers 🧠
The same SanDisk you've seen on USB drives, SD cards, and laptop SSDs. It used to be inside Western Digital (WD), but spun off as a separate listing in February 2025. Think of it as the older brother sticking with the family business (hard drives) while the younger one branched off to chase the new opportunity (NAND memory).
Three things they do:
- 🔵 Data center SSDs (25% of revenue): big-capacity storage for cloud servers like Google and AWS. Exploding in the AI era! (+233% in a single quarter)
- 🔵 Built-in memory for PCs and smartphones (53% of revenue): laptop SSDs, Galaxy and iPhone storage
- 🔵 SD cards and USB drives (14% of revenue): camera cards, USB sticks
Customers: Google, Microsoft, Meta, AWS + laptop and smartphone manufacturers.
🌍 How Big Is This Company?
| Item | Number | In Plain English |
|---|---|---|
| 🌐 NAND market share | 12.4% (5th) | 1 in 8 NAND chips comes from SanDisk |
| 🥇 AI data center SSD growth | +233% (in one quarter) | Like a tteokbokki shop tripling sales in a month |
| 💰 Market cap | ~₩358T | 1.8x SK Hynix (~₩200T) |
| 📈 Price since spin-off | +3,300% (1 year, 1 month) | ₩10K → ₩330K |
| 🔥 YTD return | +558% | ₩1M on Jan 1 → ₩6.58M today |
🆚 Rival Comparison
| Company | Share | One Line |
|---|---|---|
| 🥇 Samsung | 32.3% | "Big brother", diversified memory powerhouse |
| 🥈 SK Hynix | 19.3% | "Second brother", HBM-first |
| 🥉 Kioxia (Japan) | 15.3% | "Friendly rival", JV partner |
| 4th Micron | ~12% | "American sibling" |
| 5th SanDisk | 12.4% | "100% NAND-pure youngest", fastest growing |
Analogy: the older brothers run department stores, SanDisk runs a NAND-only specialty shop. The AI boom only blew up the NAND aisle, so the youngest brother is making money faster than the rest.
🚀 Why Did It Surge? (5 Reasons)
1️⃣ Revenue jumped 3.5x YoY (+251%)
| Item | Last Year | This Quarter |
|---|---|---|
| Revenue | ~$1.7B | $5.95B |
| Estimate | $4.73B | $5.95B (+25.8% beat) |
Analogy: a tteokbokki shop that sold 10K orders last year now sells 35K the same month, and at higher prices.
2️⃣ 78.4% of revenue is gross profit! (Highest margin ever)
Gross margin, how much of every $100 in sales becomes profit.
- 1 year ago: loss
- 6 months ago: 27.8%
- 3 months ago: 51.1%
- This quarter: 78.4% ($78.40 of every $100!) 😱
Analogy: not a regular memory business, closer to luxury handbags or a SaaS company in margin terms.
3️⃣ Company itself says next-quarter revenue will grow another 33%
"Q4 revenue ~$8.0B (~₩11.6T), EPS ~$31.5 (~₩46K) expected"
The market expected next-quarter revenue to decline. The company says it will grow +33%, almost double what the Street modeled.
4️⃣ AI giants pre-booking 5 years of memory ($42B locked in!)
The most shocking part. NBM (multi-year supply contracts), Google, Microsoft, and Meta are pre-booking memory for five years.
- Contracts signed: 5
- Minimum revenue committed: $42B (~₩61T), twice Korea's annual government R&D budget
- Financial guarantees: $11B (~₩16T), money they get even if customers go bust
- One-third of FY27 (next year) memory output is already sold
Analogy: a typical memory company is hand-to-mouth, make today, sell today, prices everywhere. SanDisk is now a kimchi factory with five years of orders pre-booked.
5️⃣ Inventor of "HBF," the new AI-era memory
SanDisk is building HBF (High-Bandwidth Flash), temporary memory for AI inference. Today, AI uses expensive HBM (where SK Hynix is #1). HBF is cheaper while storing more.
- HBF NAND chips: shipping late 2026
- HBF systems: early-to-mid 2027
If it works, a brand-new market opens up alongside SK Hynix's HBM.
📊 Post-Earnings Price Action
| Date | Close | Change | Cumulative |
|---|---|---|---|
| 4/29 (D-1) | $1,064 | +3.0% | 0 |
| 4/30 (Earnings) | $1,096 | +3.0% | +3.0% |
| 5/1 | $1,187 | +8.3% | +11.6% |
| 5/4 | $1,256 | +5.8% | +18.0% |
| 5/5 | $1,406 | +12.0% | +32.1% |
| 5/7 | $1,340 | -5.0% | +25.9% |
| 5/8 | $1,562 | +16.6% | +46.8% 🚀 |
→ +47% in 9 sessions. Not all in one shot, gradual, as the market slowly digests "wait, this is real?".
🧠 Big-Picture Analysis: +3,300% in 1 Year, Real or Bubble?
🤔 33x in 1 Year, How?
Spin-off market cap ~$5B → 5/8 ~$170B+, a 33x move.
4 Drivers Firing at Once
| Driver | Change |
|---|---|
| 1️⃣ NAND memory cycle turn | 2024 losses → 2026 explosion |
| 2️⃣ AI inference era arrives | Memory is as critical as GPUs |
| 3️⃣ NBM multi-year contracts ★ | Tames price volatility |
| 4️⃣ HBF new tech invention | Next-gen memory for AI inference |
🤯 But Real Risks Too
1️⃣ Memory is a cyclical industry
- Historically NAND swings boom/bust every 2–3 years
- We may be near the cycle peak
Analogy: like apartment prices that surged in a real-estate boom. Buying now means buying expensive.
2️⃣ Beta 5.04 = 5x market volatility
- If S&P 500 falls 1%, SNDK falls 5%
- Macro sell-offs can mean -30 to -50%
3️⃣ +47% in 9 sessions
- Strong short-term mean-reversion pressure
4️⃣ NBM 5-year contracts = double-edged sword
- Stabilizes pricing but caps upside on future price hikes
- If memory prices keep rising, competitors benefit more
🎯 Verdict: "Real + Cycle-Peak Risk = Staggered Entry Only"
| View | Weight | Comment |
|---|---|---|
| Justified re-rating | 50% | NBM + HBF + AI inference = real |
| Cycle-peak bubble | 30% | Memory is cyclical; downcycle risk |
| More upside ahead | 20% | FY27 guide raise could mean +30% |
In a cyclical industry, "where you are in the cycle" matters more than "the company itself". NBMs partly tamed cycle volatility, but if memory prices roll over, SNDK rolls over with them. Strong fundamentals ≠ a buy signal.
🚦 Traffic Light Summary
| Item | Signal | Reason |
|---|---|---|
| Revenue growth | 🟢🟢 | +251% YoY |
| Margin | 🟢 | 78.4% all-time high |
| Next-quarter guide | 🟢🟢 | $8B (2x consensus) |
| AI demand | 🟢🟢 | NAND is core AI infra |
| NBM contracts | 🟢🟢 | $42B locked + $11B guarantee |
| Current price | 🟡 | Up too fast, stretched |
| Volatility (beta) | 🔴 | 5.04, 5x the market |
| Margin peak concern | 🟡 | 78.4% all-time high; normalization risk |
💸 ₩1M Invested → 1 Year Later?
| Scenario | Probability | After 1 Year |
|---|---|---|
| 🟢 Bull (to $1,800) | 35% | ~₩1.21M (+21%) |
| 🟡 Neutral (to $1,038) | 45% | ~₩660K (-34%) |
| 🔴 Bear (to $513) | 20% | ~₩330K (-67%) |
Average expected: ~+12–15% return.
⚠️ Today's $1,562 is 34% above the average scenario PT. Buying in the $1,300–$1,400 zone is safer.
🧠 Buy or Don't Buy?
Short answer: good company long-term, dangerous to chase short-term. The stock just ran +47% in 9 sessions.
✅ Suitable for
- People who can hold 1+ years → bet on the AI/NAND supercycle
- People who can buy in tranches → split into $1,300–1,400
- People who can stomach 25%+ volatility
❌ Not Suitable for
- Short-term traders, already up too much
- Income/dividend seekers, almost no dividend, high volatility
- Those who can't tolerate ₩1M → ₩660K, bear case is -34%
🎯 Suggested Strategy
- Don't buy now, wait for the $1,300–1,400 zone
- Stagger purchases, not all at once, split into 3–4 entries
- Cap at 5–10% of portfolio (beta 5.04)
- Hold through the August 13, 2026 earnings; check NBM additions and Q4 margins
- Stop-loss at $1,200, below that, the thesis breaks
💱 Notes for Korean Investors
| Item | Detail |
|---|---|
| 💵 FX | $1 = ₩1,450.8 (stable) |
| 📊 Per-share price | $1,562.34 = ~₩2.267M (one share is expensive!) |
| ✂️ Toss Securities | Fractional shares let you buy in ₩10K / ₩50K increments |
| 💰 Capital gains tax | 22% above ₩2.5M annual exemption; volatile names benefit from active realization |
👀 What to Watch Next
Next earnings: August 13, 2026 (Q4 FY26).
| Priority | What? | Bull | Bear |
|---|---|---|---|
| 🥇 | Q4 revenue (guide mid $8.0B) | ≥$8.5B | <$7.7B |
| 🥈 | NAND ASP | Further upside | Slowing/declining |
| 🥉 | New NBM signings | +3 or more | 0 |
| 4 | Stargate revenue recognition | $400M+ | Delayed |
| 5 | First FY27 guide | Revenue +30% | Conservative |
→ 3+ 🟢 means PT raised to $2,000+; 3+ 🔴 means cut to the $1,000s.
🤔 Frequently Asked Questions
Q1. Is +3,300% in a year really real? A. Yes. ~$48 right after spin-off → $1,562. NAND cycle turn + AI era + NBM contracts all triggered together. Rare, and could be near a cycle peak. Chasing short-term is risky.
Q2. Aren't we near the cycle peak already? A. That is the core risk. NAND has 2–3 year cycles, and we're potentially near the recovery peak after the 2024 trough. NBMs partly tamed it, but not full immunity. Downcycle could mean -50%.
Q3. SK Hynix / Samsung vs SanDisk? A.
- SK Hynix / Samsung: HBM (AI training) leaders, overwhelming revenue
- SanDisk: invented HBF (AI inference), next-gen bet
- Different domains. SNDK has more next-gen exposure.
Q4. ₩1M investment strategy? A.
- ₩0: now (no chasing ❌)
- ₩500K: stagger into $1,300–1,400 zone
- ₩300K: after Q4 earnings (August)
- ₩200K: when first FY27 guidance drops
- Stagger sells too (capital gains tax)
🚨 Risk Warnings
⚠️ Extreme volatility, beta 5.04, ±47% in a month is normal. ⚠️ Cyclical industry, downcycle could mean -50%. ⚠️ No chasing, +47% in 9 sessions, mean-reversion risk. ⚠️ Strongly not recommended for conservative investors.
📚 Glossary
| Term | In Plain English |
|---|---|
| NAND | Memory that holds data even with the power off. Core of USB, SSD, SD cards |
| SSD | Storage built from many NAND chips |
| HBM vs HBF | HBM (SK Hynix #1) = for AI training, HBF (SanDisk) = for AI inference |
| NBM | Multi-year supply contract pre-booking 5 years |
| Spin-off | Splitting a business unit out of a larger company as its own listing |
| YoY | Compared to the same quarter last year |
| Margin | How much profit per $100 of revenue. 78.4% = $78.40 |
| EPS | Earnings per share. $23.41 means each share earned $23.41 |
| Beta | How much the stock moves when the market moves 1%. 5.04 = 5x |
📌 One-Sentence Conclusion
"Riding the AI-era NAND golden age, the youngest sibling rocketed +3,300% just one year and one month after spin-off, and is now using NBM multi-year contracts to tame even cyclical volatility. The +47% in 9 sessions is short-term-stretched, so staggered buying in the $1,300–1,400 zone is the answer." 🚀
If you want the data, logic, and 12M PT derivation in detail, see the analyst report version. For comparison with peers in the same earnings season, DigitalOcean +40% / Datadog +31% / Cloudflare -24% / ServiceNow -17%, you'll see how the AI capex cycle separates winners from losers. New to stocks entirely? Start with the stock investing beginner guide.
⚠️ This post is educational reference only; investment decisions are your own responsibility. Talk to a parent/guardian and never invest with borrowed money.