Semiconductors

The semiconductors sector encompasses companies designing, fabricating, and manufacturing integrated circuits, including fabless AI accelerators, pure-play foundries, and memory producers. It matters now due to explosive AI-driven demand, with the global market reaching $792B in 2025 (up 25.6% YoY) and projected to hit $975B by 2026. The structural growth driver is the AI accelerator and high-bandwidth memory supercycle, fueling 13% projected CAGR through 2030 with acceleration from data center buildouts. The key tension is extreme concentration of value in a handful of leaders (NVDA at ~80% AI GPU share, TSMC at 60%+ advanced foundry) versus cyclical risks and geopolitical supply chain vulnerabilities. Bottom line for investors: the sector offers asymmetric upside for winners with wide moats in AI, but requires selective exposure as small-scale players lag.

13% projected CAGR with acceleration from AI, supported by $975B SAM by 2030 and recent 25.6% industry growth.

High in AI/logic with NVDA dominance and foundry leadership by TSMC, though memory shows more balanced competition.

Elevated due to Taiwan concentration and ongoing US export controls impacting the supply chain.

Rapid innovation in AI accelerators, process nodes, and memory creates continuous disruption and performance leaps.

Highly attractive for selective AI-exposed winners (composite scores 92/85/78 for NVDA/TSMC/MU) despite concentration risks.

The semiconductor sector is positioned for explosive multi-year growth as artificial intelligence infrastructure drives unprecedented demand for advanced logic and high-bandwidth memory chips, with the global market reaching approximately $792 billion in 2025 sales and forecasted to approach or exceed $975 billion to $1.3 trillion in 2026. NVIDIA's dominance in AI accelerators, combined with leading-edge manufacturing from TSMC and memory leadership from Samsung and Micron, underscores a winner-take-most dynamic where cutting-edge nodes and AI-specific technologies command premium pricing and margins. The key tension lies in balancing massive capital investments and geopolitical supply-chain risks against the debate over whether AI-driven hype will sustain long-term growth or lead to overcapacity in non-AI segments.

• Companies primarily engaged in the design, fabrication, or manufacturing of integrated circuits and semiconductor devices
• Pure-play foundries providing contract wafer manufacturing services
• Memory producers specializing in DRAM, NAND, or high-bandwidth memory for computing and AI
• Fabless or IDM designers of GPUs, processors, FPGAs, or analog/mixed-signal semiconductors
• Firms focused on advanced nodes or specialty semiconductor solutions supporting data centers, AI, and high-performance computing

• Semiconductor capital equipment manufacturers and materials suppliers
• Broad electronics or consumer device assemblers that do not primarily derive revenue from semiconductor production

The SAM represents the global semiconductor market accessible to the 8 public companies in the universe (primarily fabless, IDM, and equipment-adjacent players focused on logic, memory, connectivity, and AI-driven segments). It aligns with the consensus 2026 industry sales forecast, excluding highly captive or China-internal production not captured in public revenue pools.

TAM reflects the total global semiconductor market, including all ICs (logic, memory, analog, discrete, opto), wafers, and related value (standard sales reporting plus McKinsey adjustments for captive/hyperscaler-internal production and packaging value). Excludes pure semiconductor equipment and materials markets. The wide range is driven by traditional sales-based forecasts (SIA/Deloitte/Statista ~$0.9-1T near-term) versus McKinsey's higher value-capture methodology that highlights underestimation in AI-era segments. Key caveat: significant divergence between sources on 2024/2025 base (~$630-792B) and long-term trajectory; AI concentration creates volatility risk. Public company universe revenue (~$237.7B) represents a plausible fraction of accessible market.

• NVDA: Core long due to unmatched AI GPU dominance, explosive revenue, and high margins.
• TSMC: Foundational pick for AI supply chain exposure with strong growth and reasonable valuation.
• MU: Play on AI memory cycle with accelerating revenue and margin expansion.

• LSCC: Structural loser with stagnant revenue, margin pressure, and inefficiency at scale.

Exploit AI leader vs. non-AI laggard in semiconductors for sector-relative performance.

Memory AI beneficiary versus trailing foundry with negative growth trajectory.

• SMH
• SOXX