SanDisk 725% Rally Hit by Korean Market Crash|Record Fundamentals Cant Stop the Sell Call

Chapter 1: The Record That Broke Overnight

SanDisk closed at an all-time high Monday and opened down 11% on Tuesday. The trigger came from Seoul, not from anything in SanDisk's own business. South Korea's Kospi fell 10% on June 23, its steepest single-day drop in over three months, as AI valuation concerns collided with forced liquidations in the country's chip-heavy index. SK Hynix and Samsung each slid more than 12%. Because those two names sit alongside Micron as top holdings of the Roundhill Memory ETF, the selloff crossed the Pacific before U.S. markets opened. That contagion logic is the first thing holders need to understand: SanDisk does not make DRAM. Its product is NAND flash — a different memory type with a different supply chain. Yet the stock fell in lockstep with Korean DRAM names, because it had been grouped into the same "AI memory supercycle" basket trade. What gets bought as a basket gets sold as a basket. The provisional answer for why a record-fundamentals company opened down 11% is not fundamental deterioration — it is basket-level positioning unwinding after one of the most concentrated rallies in semiconductor history. SanDisk had gained 725% year to date and 4,342% over the prior twelve months before the session that broke the streak. Names that move like that tend to reverse hard on the first serious wobble in the macro frame that carried them. The wobble was Korea. The mechanism, though, runs deeper than a single day's contagion. What the Korean crash exposed is that the stock's price had been moving on a different clock than its fundamentals, and that disconnect is what now puts both holders and watchers in an impossible position.

Chapter 2: The Same Data, Two Opposite Conclusions

The numbers from SanDisk's most recent quarter are unambiguous. Q3 FY26 revenue hit $5.95 billion, up 251% year over year, beating consensus by a wide margin. EPS of $23.41 crushed the $14.66 estimate by 59.67%. Datacenter revenue exploded 645% year over year to $1.467 billion. Gross margin expanded from 22.5% to 78.4%. The company retired $650 million in debt and now carries zero long-term obligations. CEO David Goeckeler called it "a fundamental inflection point" driven by a deliberate mix shift toward the highest-value datacenter customers. Five signed multi-year agreements lock in hyperscaler demand. Those facts are not in dispute. What is in dispute is what they are worth at a stock price of roughly $1,958. 24/7 Wall St. looked at those exact numbers and issued a sell rating with a twelve-month price target of $1,663 — implying 15% downside at 90% confidence. Their argument is not that the business is broken but that the stock has already priced in two to three years of the bull thesis. A trailing price-to-earnings ratio of 72, combined with the historical pattern of NAND pricing reversals, leads them to conclude that even perfect execution cannot generate meaningful returns from current levels. The Motley Fool looked at the same numbers and projected a price of $3,661 by 2027, based on a 20-times forward earnings multiple applied to fiscal 2027 EPS estimates of $183. Their argument is that the structural NAND shortage extends through 2028, that BiCS8 technology gives SanDisk a cost advantage, and that High Bandwidth Flash positions it for AI inference workloads that haven't fully emerged yet. The conflict between these two readings is not a pro-con exercise on a single fact. Both sides have named their assumptions and both sides are drawing on the same Q3 print. That is the most dangerous kind of disagreement in markets — not the kind where one side is clearly wrong, but the kind where the outcome depends entirely on which assumption controls: cyclical reversion or structural demand.

Chapter 3: The Hidden Assumption Separating $1,663 from $3,661

Both the bear and the bull case rest on a shared foundation they never state explicitly: that Q4 guidance will hold. SanDisk guided Q4 FY26 revenue of $7.75 billion to $8.25 billion with non-GAAP EPS of $30 to $33. Every analyst projection that reaches $3,000 or beyond for 2027 threads through that Q4 print confirming the acceleration. Every bear argument that lands at $1,663 implicitly assumes that Q4 delivers in-line results at best, and that forward multiples compress as the cycle matures. The buried assumption the bull side treats as given is that hyperscaler demand does not moderate before NAND supply catches up. TrendForce data showed memory contract prices up more than 100% in the first half of 2026. Omdia forecasts global DRAM revenue at $372 billion in 2026, up 147% year over year. Apple CEO Tim Cook called memory price increases "unavoidable" and said the supply backdrop had become "unsustainable" — language that confirmed NAND suppliers hold pricing leverage over even the world's largest consumer electronics buyer. Mizuho's Jordan Klein estimated memory bill-of-materials costs had risen from the mid-teens to 25–30% of smartphone and PC cost, with at least $100 to $200 in expected iPhone 18 price increases as a result. These signals support the bull case. The bear counter-assumption is equally grounded: the entire semiconductor industry — Micron, SK Hynix, Samsung — is collectively committing more than $75 billion annually to new capacity. New fabs take two to three years to come online, pointing to 2027–2028 as the period when supply catches up. SanDisk's Consumer segment already declined 10% sequentially in Q3, the first visible crack in demand outside the datacenter. The sell case does not require a crash — only that supply growth eventually catches demand, compressing the margins that are driving both the EPS estimates and the multiple. What neither side will know until August is whether Q4 revenue lands near $8 billion with margin expansion intact, or whether it prints at the guidance floor with Consumer continuing to soften. That single number will determine which assumption was carrying more weight.

Chapter 4: What the Holder and the Watcher Each Need to See

The bear evidence in the pool is real: 24/7 Wall St.'s $1,663 price target is grounded in cyclical margin math, not speculation. The stock's forward multiple of 33 leaves little room if Q4 guidance disappoints. The Consumer segment's sequential decline in Q3 is a concrete data point that the AI-fueled mix shift is not reaching all of SanDisk's business. That counter-evidence does not break the bull thesis, but it means a holder cannot dismiss it by citing datacenter strength alone. The main risk is an August earnings print that hits the guidance midpoint without acceleration in Q5 forward estimates — a result that would be operationally fine but strategically uninformative, leaving the contested assumption unresolved and the stock rangebound. A holder's confirmation criterion is Q4 FY26 earnings in August: revenue above the $8 billion midpoint, gross margin at or above Q3's 78.4%, and Q5 guidance that implies another sequential step up. All three need to hold together. A top-line beat with margin compression, or in-line revenue with flat guidance, does not resolve the debate — it hands a point to the bear. A watcher's entry criterion is simpler but requires patience: the thesis is not broken, but the price needs to compress toward the $1,400-1,600 zone where the model becomes constructive on a forward basis before the Q4 print, not after. Buying into a 725%-YTD stock two months before the earnings that decide the structural thesis is not a risk-adjusted entry. The variable that decides which direction SanDisk takes from here is not the Korean Kospi or macro sentiment. It is August's Q4 report, specifically the margin line and the forward revenue outlook. If that number confirms, the $3,661 scenario comes back into play. If it misses, the 2022–2023 memory cycle — where Micron fell 50% and inventories hit 31 weeks — becomes the reference frame that replaces the supercycle narrative.