Hook & Thesis
Sandisk plunged alongside Micron after headlines about Google's new TurboQuant compression algorithm roiled the memory complex. That reaction is understandable - the idea of 6x less memory needed for large language models is scary - but it's a classic example of headline-driven extrapolation. The market is treating a potential efficiency gain as an immediate demand destroyer for NAND, when in reality adoption, performance trade-offs, enterprise buying cycles, and supply dynamics will mute and delay any structural impact.
We see a tactical long opportunity on a mid-term basis. The company still prints free cash flow, carries minimal net leverage, and is trading below near-term technical resistances after a sharp correction. Entry at $570, stop at $525, target $730, horizon mid term (45 trading days) - this is a disciplined swing trade that bets on the overreaction fading and AI infrastructure demand reasserting itself.
What Sandisk Does and Why the Market Should Care
Sandisk develops and sells NAND flash-based storage: SSDs, memory cards, and USB flash drives. These are core components of data centers and AI infrastructure where high-performance flash is used for model training, caching, and archival tasks. When data center operators scale AI deployments, they buy storage capacity and high-performance SSDs; that direct demand is the single most important fundamental driver for Sandisk.
Snapshot of the Fundamentals
| Metric | Value |
|---|---|
| Current Price | $574.00 |
| Market Cap | $84.5B |
| Free Cash Flow (TTM) | $1.449B |
| Price / Sales | 9.46x |
| EV / EBITDA | 63.9x |
| EPS (TTM) | -$7.05 |
| Cash / Current Ratio | Cash: 0.93; Current: 3.11 |
Two items stand out from the numbers: first, Sandisk generates meaningful free cash flow ($1.449B), which supports capital allocation flexibility and buybacks if management chooses; second, valuation on sales and EV/EBITDA looks elevated (P/S ~9.5x and EV/EBITDA ~64x), reflecting the market's willingness to price in durable high-margin growth from AI demand. That combination - positive cash generation but a high multiple - is why headline shocks like TurboQuant trigger outsized moves: expectations are priced to perfection.
What's Behind the Recent Drop
On 03/30/2026 markets sold NAND and memory names after reports that Google's TurboQuant could compress model memory needs by roughly 6x. The immediate read was simple: if models need less RAM/VRAM/NAND, hardware demand could soften. Add a general risk-off backdrop from high oil prices and geopolitical risk in the Middle East, and the result was heavy selling across semiconductors.
Technically the stock is below its 10- and 20-day averages (SMA10 ~ $683, SMA20 ~ $646), with RSI ~42 and MACD showing bearish momentum. That technical context amplifies headline-driven exits and creates a tactical buying window for patient traders.
Trade Plan
- Direction: Long
- Entry: $570.00
- Stop Loss: $525.00
- Target: $730.00
- Horizon: mid term (45 trading days) - expect the market to remobilize around fundamentals and/or digest further updates on TurboQuant's real-world adoption over this period.
Rationale: entry at $570 buys a discount to recent intraday levels and leaves room for short-term noise while the stop at $525 limits downside if the market shifts to structurally repricing NAND demand. The $730 target is a recovery toward recent swing highs around the high-$700s and reflects a reversion toward the trend that prevailed through March. Our mid-term horizon aligns with the time needed for enterprise testing, vendor commentary, and inventory-cycle shifts to become clearer.
Catalysts to Watch (2-5)
- Vendor commentary from major cloud/data center customers about TurboQuant adoption and any resulting demand changes.
- Sandisk earnings or operational updates that show revenue/mix stability and continued FCF generation.
- NAND pricing stabilization or another round of pricing increases driven by capacity discipline.
- Competitor supply signals (capex plans, inventory levels) that support a tighter supply/demand balance.
Valuation Framing
At roughly $84.5B market cap and an enterprise value just under $84B, Sandisk is priced for sustained, premium growth. EV/EBITDA near 64x and P/S around 9.5x are rich by historical semiconductor memory standards, but they reflect the market's view that AI-driven demand will be durable and margins will remain high. If that thesis holds, the stock can justify current multiples. If it does not, multiple compression is the primary downside path.
Importantly, free cash flow of $1.449B gives management options: de-risk via buybacks, capex discipline, or M&A. Low reported leverage (debt to equity ~0.06) reduces balance-sheet risk, a constructive feature in uncertain macro conditions.
Risks and Counterarguments
- TurboQuant is real and widely adopted. If Google's compression becomes a widely used standard for large models quickly, incremental memory demand per model could fall materially, shrinking long-term unit demand for NAND. That would be a structural downside to the thesis.
- Macroeconomic slowdown or capex pullback. Geopolitical shocks (e.g., rising oil prices or broader recession fears) could depress data-center spending, exacerbating downside beyond headline compression fears.
- Valuation multiple risk. The stock already embeds high expectations. If revenue growth or margin expansion disappoints, multiple contraction could accelerate declines even if fundamentals remain decent.
- Short-term technical momentum may persist. With MACD bearish and price below shorter-term SMAs, momentum traders could keep selling into weakness, delaying a recovery beyond our 45-trading-day horizon.
- Counterargument: TurboQuant's compression gains do not instantly translate into equivalent real-world demand declines. Enterprise adoption cycles, data locality, latency and performance trade-offs, and redundancy needs mean customers will still buy substantial flash capacity. Combining that with supply discipline in NAND manufacturing, demand could remain strong enough to support current fundamentals.
What Would Change Our Mind
We would abandon the long if any of the following occur: a) credible, repeated vendor confirmations that TurboQuant is being deployed at scale and materially reducing new storage purchases; b) Sandisk reports a sudden drop in enterprise SSD orders or guidance that implies durable demand erosion; or c) NAND pricing collapses due to aggressive capacity additions or a sudden inventory glut.
Conclusion - Clear Stance
Short-term headlines are creating a tactical buying opportunity. Our view is constructive: buy a measured position at $570 with a defined stop at $525 and a target at $730 on a mid-term (45 trading days) horizon. The trade is not a statement that TurboQuant is irrelevant - it could be consequential - but rather a disciplined risk/reward play that assumes the market has overshot on headline risk while underappreciating Sandisk's FCF generation, limited leverage, and the sticky nature of enterprise storage demand.
Execution note: keep position size consistent with the stop and your portfolio risk limits. If you prefer a staggered approach, consider starting half the size at $585 and adding to $570 to lower average cost on additional weakness.