Why The iShares Semiconductor ETF Soared 112.8% In The First Half of 2026

iShares Semiconductor ETF (SOXX) is up 112.8% in 1H 2026 and nearly 300% over 5 years, reflecting investor bets that AI infrastructure demand is being amplified by memory chip shortages. The article highlights Micron’s stock up 314% YTD (as of July 10, 2026) alongside strength in other large AI beneficiaries like Intel and Nvidia. Despite the momentum, it warns semiconductors are historically cyclical and cautions against chasing after a >100% YTD surge.

Analysis

The key setup is not "AI is bullish semis"—that is fully owned—but that the current phase of the cycle is rewarding the most supply-constrained links in the chain while masking how late we may be in the broader basket. MU is the cleanest direct winner because memory pricing is still levered to tight inventory, but that is exactly the sort of margin peak that invites capacity additions, customer pushback, and a 2-4 quarter normalization window. By contrast, SOXX is a poor vehicle for this trade because it mixes the scarce winners with more mature names whose upside is now mostly multiple support, not earnings acceleration.

Second-order effects matter more than the headline AI narrative: if memory and HBM remain tight, the bottleneck shifts to server builds, advanced packaging, and foundry allocation rather than to GPU demand itself. That can cap the near-term unit growth of NVDA-adjacent hardware ecosystems even if order books stay healthy, while also creating a relative tailwind for equipment vendors and specialized suppliers that get paid earlier in the capex cycle. INTC remains more of a policy optionality story than a pure AI earnings compounder, so it should underperform any basket priced for scarcity-driven margin expansion.

The timing matters: momentum can persist for days to weeks because crowded winners tend to run until the first sign of inventory relief, but the 1-3 month risk is a sentiment reversal if guidance merely confirms that shortages are being met with supply response. Over 6-18 months, semis remain structurally attractive, yet index-level returns are likely to compress as the market shifts from scarcity pricing to digestion. The contrarian miss is that the shortage itself is often the best signal that forward returns are getting less attractive, not more.

Key Decisions for Investors

• Short SOXX on strength over the next 1-2 weeks, using a 3-6 month horizon; thesis is multiple compression as the basket moves from scarcity premium to peak-cycle risk. Falsify if SOXX re-rates higher on improving forward guide revisions rather than pure multiple expansion.
• Pair trade: long MU / short SOXX for 1-3 months. MU should retain the most direct pricing power if shortages persist, while SOXX is diluted by names with weaker operating leverage. Risk/reward improves on any market-wide semis rally because MU should outperform on spread widening.
• Avoid adding to NVDA here; if memory constraints become the binding bottleneck, upside shifts from accelerator demand to supply-chain friction. Use any post-earnings rally to trim exposure unless management explicitly raises shipment cadence and sees no packaging/memory constraint.
• Watch INTC as a relative-value candidate only on policy headlines, not as a clean AI alpha long. If AI capex broadens beyond GPUs into CPUs and edge inference over the next 6-12 months, it can work; otherwise it remains a lagging beneficiary versus MU and the equipment complex.
• Set an alert for any evidence of memory capex ramp or inventory rebuild; that would be the first sign the MU-led trade is getting crowded and the short SOXX thesis is maturing.