SNDK at $820, are you bottom-fishing or catching a falling knife?
From 2300 down to 820, halved twice. Revenue surged 378%, EPS beat expectations, $5 billion buyback—yet the stock price crashed like a dog. On the eve of the stock split, is 820 the golden pit for AI storage, or a meat grinder halfway up the mountain?
First, look at the surface: explosive earnings report, but the stock price collapsed.
On October 29, the earnings report showed revenue of 11 billion, up 378% year-over-year; data center revenue 4.8 billion, up a staggering 1600%; gross margin 85%; $5 billion buyback; and an 8-for-1 stock split announced. For any company, this would be a script for a limit-up. Instead, SNDK crashed from 2300 to 820, down over 60%. Earnings reports are like a truth mirror, stock price is an emotion gauge—the better the performance, the harder the sell-off.
Why? Because expectations were already maxed out. Everyone knows about the AI storage super cycle. You buy the “future,” sell the “realization.”
First point: this is not a fundamental collapse, it’s a leverage collapse.
AI training, inference, KV cache, data centers—all consuming storage. Enterprise SSDs already account for nearly half of NAND shipments, supply is tight through 2027. The company is debt-free, cash-rich, with long-term contracts locking capacity, and customers prepaying deposits.
Is the fundamental strong? Yes. But perpetual contract funding rates have been positive—longs are paying to hold positions. When the earnings beat came out, leveraged longs collectively panicked; this is not value destruction, it’s leverage cleansing.
Second point: the stock split is not a positive, it’s a truth mirror.
On November 19, the 8-for-1 split takes effect, with record date November 9. Many get excited hearing “stock split”: “It’s cheaper now, time to buy!”
Wake up. A stock split doesn’t change any fundamentals, only the per-share price. It’s candy for retail investors, not meat for institutions. Liquidity expands and volatility intensifies before and after the split, so price spikes will be even sharper. You think 820 is the bottom, after the split it becomes 102, but you can still get swept out by a spike.
What really matters is whether the 800-820 demand zone holds. If it holds, post-split is a new narrative; if not, expect 750-780.
Third point: the AI storage cycle isn’t over, your patience is.
TrendForce raised the 2026-2027 memory market size forecast, NAND supply-demand imbalance continues, Agentic AI further boosts storage demand. SanDisk’s backlog is huge, gross margin 85%, this is not a cyclical peak bubble, it’s structural shortage.
But you must distinguish: correct mid-to-long-term logic doesn’t mean short-term prices won’t crash. From 2300 to 820, what fell was valuation excess and sector sentiment, not AI storage demand. Analysts’ average target price is still above 2000, but that’s a one-year outlook.
Bull vs. bear, judge for yourself
On one side:
Revenue +378%, data center +1600%, gross margin 85%
$5 billion buyback, no debt, cash rich
NAND shortage continues to 2027, pricing power in hand
Strong demand zone 800-820, oversold rebound momentum
Post-split retail inflow, liquidity expansion
On the other side:
60% drop from 2300, lower highs and lows, bearish structure
Positive perpetual funding rates, long leverage still present, liquidation risk not cleared
Semiconductor/AI sector overall correction, linked to BTC/Nasdaq
Volatility intensifies around split, high spike risk
Memory cyclicality, AI CapEx slowdown means Davis double kill
Key levels
Resistance above: 900-950 → 1000-1100 → 1200+
Support below: 800-820 (strong demand) → 750-780 (breakdown target)
Trading strategy
Conservative bulls:
Scale in lightly at 800-820, total position 5-10%, leverage ≤5-10x. Strict stop loss below 780. First target 900-950, second target 1000+. Add positions on volume-supported hold above 850-900 with higher lows. Avoid all-in; building positions gradually on pullbacks is better.
Short-term traders:
Wait and watch, enter small longs on 4H volume surge or breakout of short-term downtrend, quick profit-taking. If rebound meets resistance at 900-950 with weak volume, consider light short positions targeting 820/800 retest. Below 800, trend continuation short but beware oversold bounce.
Long-term believers:
If you believe in the AI storage cycle, 800-820 is a region worth scaling in gradually. Use spot tokens or low leverage; avoid high-leverage perpetual bets on the split. Liquidity traps abound around the split; staying alive is more important than bottom-fishing.
SNDK now looks like Nvidia in 2022—
Explosive earnings, stock halved, everyone said “cycle topped.” What happened? AI came, from 100+ to over 1000.
$BTC$SNDK$NVDA
Disclaimer: OKX Orbit content is provided for informational purposes only. Learn more
