Prediction: Sandisk Will Reclaim Its All-Time High by the End of the Year

SanDisk (SNDK) reported fiscal Q4 revenue of $8.97B vs guided “up to” $8.25B, a 51% sequential revenue jump, and guided fiscal Q1 2027 revenue up to $10.8B (implying ~20% sequential growth). GAAP net income reached $6.9B, and management emphasized “growing and durable free cash flow,” with the forward P/E falling below 20 after the stock’s pullback. The article argues hyperscaler AI capex is likely to keep memory demand strong, reducing the likelihood of an inventory-driven price crash.

Analysis

SNDK is the clearest direct beneficiary of a tighter memory cycle, but the second-order winner is the broader AI storage stack: if hyperscalers are truly locking in multi-quarter capex, enterprise SSD demand and controller pricing can stay firmer than consensus expects. That said, the read-through to NVDA and AVGO is mostly sentiment support, not incremental earnings today; they gain from capex validation, while SNDK gets the actual mix shift in the near term. META and AMZN are the more subtle losers: not because AI spend stops, but because every extra dollar allocated to storage and data-center buildout pressures free cash flow discipline and raises the bar for later return on invested capital.

The key risk is timing. Memory equities can stay strong for months, but they can also reverse in a single quarter if inventory normalization or procurement discipline shows up first at the hyperscalers; this is a faster cycle than compute. The market is likely underweighting peak-earnings risk: a sub-20 forward multiple looks cheap only if gross margins and pricing hold, and that assumption can fail quickly if spot NAND softens or channel inventory rebuilds complete. Watch for any downward revision to next-quarter margin guide, weaker hyperscaler capex commentary, or MU signaling easing memory pricing.

Consensus is missing that "AI beneficiary" is not a monolith: memory tends to be more volatile and more easily mispriced than GPU demand, so the upside can be real but the drawdown risk is also larger. The best expression is relative value, not a naked long at any price, because the stock is already being valued on a durable-scarcity narrative that needs continued proof. If that proof continues, the rerating can extend for 1-3 months; if it doesn't, the multiple can compress faster than EPS falls.

Key Decisions for Investors

• Buy SNDK on pullbacks rather than chasing strength; use 3-6 month call spreads to express the view, targeting upside if hyperscaler demand remains tight. Falsify if next-quarter margin or revenue guidance decelerates materially.
• Pair trade: long SNDK / short MU into the next memory-pricing and inventory data points. The trade works if NAND tightness keeps driving SNDK faster than the market rewards MU's broader memory exposure; exit if MU commentary confirms sustained pricing strength across the memory stack.
• Use any post-earnings weakness in META or AMZN as a source of funds to add SNDK exposure, not as an outright bearish short. The read-through is higher capex intensity, not a collapse in AI demand.
• Avoid adding NVDA or AVGO solely on this note; their thesis is already crowded and this is more confirmation than incremental catalyst. Reassess only if cloud capex guides re-accelerate again in the next 1-2 quarters.