2 Super Semiconductor Stocks to Buy Hand Over Fist in 2026

Nvidia CEO Jensen Huang forecasts data-center operators could spend up to $4 trillion annually by 2030 to support AI workloads, underpinning demand across the semiconductor supply chain. Corning reported Q3 FY2025 revenue of $4.27 billion (up 14% YoY) with optical communications revenue of $1.65 billion (+33% YoY) and enterprise optical up 58%; optical net income was $295 million (+69% YoY) and Corning trades at a TTM adjusted P/E of 35.9. Micron posted record FY2025 revenue of $37.3 billion (+49% YoY) with Cloud Memory revenue up 257% to $13.5 billion, said its FY25 non‑GAAP EPS was $8.29 (P/E ~27.3), and noted HBM3E supply is nearly sold out while HBM4 sampling is underway — signalling durable, broad-based AI-driven demand for memory and fiber-optic components. Investors should view these metrics as bullish sector fundamentals that could re-rate select chip and components suppliers if AI CAPEX continues to accelerate.

Analysis

Market structure: Winners are Nvidia (NVDA) as the GPU monopolist, memory suppliers (MU) for HBM3E/HBM4 and fiber/optical vendors like Corning (GLW); enterprise data-center operators face rising capex. Corning’s optical comms growth (+33% revenue, enterprise +58%) and Micron’s Cloud Memory +257% to $13.5B signal strong demand-driven pricing power for select suppliers, while legacy copper/low-bandwidth suppliers and commodity DRAM commodity players could be hurt.

Risk assessment: Key tail risks are (1) export controls/geopolitical restrictions on advanced chips, (2) an HBM/DRAM oversupply cycle leading to >30% price declines, and (3) NVDA concentration causing single-point demand shocks. Time horizons: immediate (days–weeks) volatility around Micron’s Dec 17 report; short-term (3–6 months) supply-chain tightness; long-term (3–5 years) uncertainty on Huang’s $4T forecast and power/grid constraints that could raise data-center OPEX by 10–25%.

Trade implications: Tactical capital should overweight Micron (MU) and Corning (GLW) and use capped option exposure to NVDA to avoid single-name delta. Expect memory tightness to support MU margins into 1H26; if MU reports Cloud Memory growth >30% YoY and bookings “sold out” language persists, reallocate incremental risk budget to MU. Watch commodity signals (HBM spot pricing, copper down >10%, silica/fiber price moves) as entry triggers.

Contrarian angles: Consensus assumes durable supercycle — missing risks include rapid capex-induced overcapacity (historical DRAM cycles 2018–2019) and utility constraints; if HBM pricing falls >25% YoY or Micron delays HBM4 ramp, upside compresses quickly. The market may be underpricing GLW’s optionality to multi-rack GPU nodes but overpricing long-duration growth for AMD/other high-P/E peers relative to MU/GLW.

Key Decisions for Investors

• If Micron (MU) beats on Dec 17: establish a 2–3% long position in MU within 3 trading days, target +40% upside over 12 months, set a stop-loss at -20% or exit if Cloud Memory growth < +25% YoY on the report.
• Allocate 1.5–2% long to Corning (GLW) via shares or 12–18 month LEAPs (buy calls) to capture optical conversion; add another 1% on any pullback >10%; target 25–35% upside over 12 months, stop-loss -15%.
• Deploy a tactical 0.5% NAV NVDA bullish call spread (expiration ~Jan 2026) to ride continued GPU tightness while capping premium; unwind if NVDA announces material supply acceleration or if implied vol collapses >30% post-earnings.
• Initiate a relative-value pair: long MU (2%) vs short a broad legacy networking/copper-exposed name (1%) or commodity copper miner (1%) to express fiber/HBM secular share gains; rebalance if copper futures rally >15% or if HBM spot prices fall >20%.