Semiconductors (SMH) vs 10Y Treasury Yield
SMH (VanEck Semiconductor ETF) tracks 25 of the largest US-listed semiconductor companies. Top holdings: NVDA approximately 20 percent, TSMC ~13 percent, AVGO ~8 percent, AMD ~5 percent.
Also known as: Semiconductors (SMH) (ETF_SMH, semiconductors, semis, chips) · 10Y Treasury Yield (10Y yield, 10 year treasury, TNX)
Why This Comparison Matters
SMH (VanEck Semiconductor ETF) tracks 25 of the largest US-listed semiconductor companies. Top holdings: NVDA approximately 20 percent, TSMC ~13 percent, AVGO ~8 percent, AMD ~5 percent. SMH price approximately $482 (April 2026, near recent highs). The 10-year Treasury yield (FRED DGS10) sits at 4.31 percent. Semiconductors are theoretically longest-duration tech sub-segment due to AI capex growth narratives with long payoff horizons. Rising 10Y yields should pressure SMH through duration framework. The 2024-2026 era has paradoxically broken duration framework: SMH +3x since November 2022 ChatGPT moment despite elevated 10Y range 3.6-5.0 percent. AI capex narrative ($400B+ hyperscaler annual run-rate) overwhelmed duration drag.
The April 2026 Configuration
SMH price approximately $482 (April 2026). 10Y Treasury yield 4.31 percent. NVDA ~20 percent of SMH, TSMC ~13 percent, AVGO ~8 percent, AMD ~5 percent. Combined top 4 ~46 percent of SMH (high concentration).
SMH has rallied substantially in 2024-2026. SMH +3x since November 2022 ChatGPT moment. AI capex narrative drove NVDA +12x peak-to-peak (2023-2025) plus Broadcom, AMD, ARM, ASML, Lam Research, Applied Materials.
NVDA Q3 fiscal 2026 results (released February 2026): revenue +62 percent YoY. Hyperscaler capex on AI infrastructure $400B+ annual run-rate (Microsoft, Meta, Alphabet, Amazon combined).
The combined April 2026 reading: SMH near recent highs despite elevated 10Y. Configuration confirms duration framework reasserts only in absence of major capex narrative. AI capex era (2024-2026) has overwhelmed duration sensitivity.
Forward-looking: NVDA next earnings May 2026 will signal AI capex sustainability. Hyperscaler Q1 2026 earnings (April-May 2026) will signal capex guidance. Sustained $400B+ run-rate supports continued SMH outperformance vs 10Y.
Why Semis Are Most Rate-Sensitive Tech Subsector
Semiconductors are theoretically the longest-duration equity sub-segment within tech. Drivers.
Growth profile: AI capex cycle has 5-15 year payoff horizon. Data center buildout (NVDA, AVGO, AMD) requires multi-year capacity expansion. Foundry capex (TSMC) requires 3-5 year construction timelines.
Multiple expansion: SMH forward P/E approximately 30x (April 2026) vs SPY 22x. Higher multiple reflects longer-duration cash flows.
Capital intensity: semis require substantial capex for fab construction. Higher discount rates compress NPV of future capex returns more than for software companies.
Empirical sensitivity (pre-2024): 100 basis point rise in 10Y yield typically associated with 8-15 percent SMH decline (over 60-90 day windows). Compare to XLK 5-10 percent decline; SPY 3-5 percent decline. Semis 1.5-2x duration sensitivity vs XLK.
2022 hiking cycle illustrated. 10Y rose 1.5 percent to 5.0 percent peak. SMH fell 38 percent peak-to-trough October 2022. Within semis, longest-duration names hit hardest: NVDA -65 percent peak-to-trough; AMD -55 percent. Established memory/legacy names (Intel, Texas Instruments) -30 to -40 percent.
2024-2026 era anomaly: AI capex narrative override. NVDA earnings revisions catalyzed SMH despite elevated 10Y.
Conditional Forward Response (Tail Events)
How 10Y Treasury Yield has historically behaved in the 5 sessions following a top-decile or bottom-decile daily move in Semiconductors (SMH). Computed from 1,238 aligned daily observations ending .
Following these triggers, 10Y Treasury Yield rises 0.65% on average over the next 5 sessions, versus an unconditional baseline of +0.59%. 123 qualifying events; 10Y Treasury Yield closed positive in 53% of them.
90-Day Statistics
Explore Each Metric
Related Scenarios & Forecasts
Get daily macro analysis comparing key metrics delivered to your inbox. Stay ahead of market-moving divergences.
Frequently Asked Questions
What are SMH and the 10Y Treasury yield?+
SMH (VanEck Semiconductor ETF) tracks 25 largest US-listed semiconductor companies. Top holdings: NVDA ~20%, TSMC ~13%, AVGO ~8%, AMD ~5% (top 4 ~46% high concentration). SMH ~$482 April 2026 (near recent highs). 10Y Treasury yield 4.31% April 2026. SMH +3x since November 2022 ChatGPT moment. AI capex narrative drove NVDA +12x peak-to-peak (2023-2025). Hyperscaler capex on AI infrastructure $400B+ annual run-rate (Microsoft, Meta, Alphabet, Amazon). NVDA Q3 fiscal 2026 revenue +62% YoY. SMH AUM ~$25B. SMH realized volatility ~28-40% annualized (vs XLK 18-25%, SPY 13-18%).
Why are semis most rate-sensitive tech subsector?+
Three drivers. Growth profile: AI capex 5-15 year payoff horizon. Data center buildout (NVDA, AVGO, AMD) multi-year. Foundry capex (TSMC) 3-5 year construction. Multiple expansion: SMH forward P/E ~30x (April 2026) vs SPY 22x. Higher multiple reflects longer-duration cash flows. Capital intensity: semis require substantial capex for fab construction. Higher discount rates compress NPV more than software. Pre-2024 empirical: 100bp 10Y rise = 8-15% SMH decline (60-90 day windows). XLK 5-10%; SPY 3-5%. Semis 1.5-2x duration sensitivity vs XLK. 2022 hiking 10Y 1.5% to 5.0%: SMH -38% (NVDA -65%, AMD -55%). 2024-2026 anomaly: AI capex override.
What is the AI capex boom?+
Largest tech investment cycle since dot-com era. Hyperscaler capex (top 4): Microsoft $80B+ FY26 (vs $50B FY24); Meta $60-65B 2026 (vs $40B 2024); Alphabet $75-80B 2026 (vs $50B 2024); Amazon $100B+ 2026 (vs $48B 2023). Combined ~$315-325B run-rate, up from $200B (2022). Plus Tesla, Apple, Oracle add ~$80B+. Total US AI capex ~$400B+ annual run-rate. Semi beneficiaries: NVDA H100/H200/Blackwell GPUs dominate AI training (revenue +62% YoY); AVGO networking + custom AI chips; AMD MI300 GPUs; TSMC foundry to all advanced AI chips; ASML EUV lithography monopoly ($200-400M per machine); Applied Materials, Lam Research fab equipment; Micron, SK Hynix HBM for AI accelerators.
Related Comparisons
Explore Across Convex
Data sourced from FRED, CoinGecko, CBOE, and other providers. This page is for informational purposes only and does not constitute financial advice. Past performance does not guarantee future results.