Morgan Stanley defends Micron, Sandisk stocks amid memory concerns By Investing.com

Sandisk shares fell 12% over the past week to $677.86 but remain up >1,100% year-to-date from a 52-week low of $27.89. Morgan Stanley reiterated overweight on memory names (including Micron and SanDisk) citing durable strength and that memory supply is becoming a primary bottleneck for AI, while analysts forecast $41.82 EPS for SanDisk in fiscal 2026. Corporate activity: SanDisk Technologies agreed to buy ~139M Nanya shares for ~$1.0B (~3.9% of Nanya) and Western Digital is selling 5,821,135 SanDisk shares in a $545 secondary ($>3B transaction); SanDisk gave no guidance update. BofA also reiterated a Buy, citing hyperscaler and AI inference demand.

Analysis

Memory market tightness is shifting bargaining power toward firms that can lock supply or embed themselves into the upstream capacity chain; ownership stakes or long-term offtakes function like quasi-fab insurance and can be worth multiple quarters of gross margin preservation when chips are rationed. That dynamic raises the option value of incumbent suppliers with scale and secured supply versus pure-channel distributors — think of it as a scarcity premium that compounds across revenue, inventory turns, and gross margin recovery.

Hyperscaler prepayments and large-volume deals shorten visibility but increase concentration risk: revenue booked today is less fungible if a single cloud buyer delays AI projects. Conversely, engineering-level productivity moves (cache optimizations, quantization, model sparsity) are real and can reduce per-model memory intensity over 6–24 months, creating a non-linear demand path where short-term pricing holds but long-term growth is uncertain.

Key reversals to watch are supply-led (new capacity coming online within 6–18 months), architectural (widespread adoption of memory-light inference stacks within 12–24 months), or capital-market driven (secondary issuance/insider liquidations that force rebalancing of free float). Macro/trade-policy tail risks — export controls or a demand shock in hyperscalers — would compress multiples rapidly, so any long position needs explicit liquidity and volatility hedges.

Net: position size should be driven by conviction in supply persistence versus the risk of architectural demand erosion. Prefer asymmetric instruments that capture re-rating if scarcity persists (calls, pairs where the long has secured supply), and avoid naked exposure to names with large potential free-float increases or highly concentrated buyer risk.

Key Decisions for Investors

• Long SNDK Jan-2027 LEAP calls (~0.5% of fund NAV): asymmetric upside if scarcity-driven re-rating continues; target +60% in 9–18 months, hard stop -30% from entry; hedge with 25% notional in out-of-the-money puts if implied volatility is cheap.
• Pair trade — Long MU stock (1% NAV) / Short SMCI (0.7% NAV) over 3–9 months: capture semiconductor raw-material pricing power (MU) versus system integrator margin compression (SMCI) if component scarcity persists; unwind if MU underperforms DRAM spot by >15% or SMCI reports order-book expansion.
• Relative-value — Long SNDK / Short WDC (equal notional, 3–12 month horizon): play secured-supply premium against potential free-float/overhang at the former-parent; expect spread capture of 30–50% if market re-prices scarcity, stop-loss: 20% adverse move in either leg.
• Volatility hedge: buy 3–6 month straddles on key names before major earnings/guidance windows (size 0.2–0.5% NAV): protects against downside from earnings misses or sudden demand guidance cuts while retaining upside participation if shortages drive beat-and-raise prints.