SMH vs QQQ: Are Semiconductor ETFs Entering an Independent Risk Cycle? How to Reassess Allocation Value

In June 2026, global technology stocks experienced a notable divergence. The semiconductor sector, represented by the VanEck Semiconductor ETF (SMH), faced a sharp sell-off, while the Invesco QQQ Trust (QQQ), which tracks the Nasdaq 100, also came under pressure but with significantly narrower losses. As of the week ending June 24, SMH had fallen over 5%, while the Nasdaq 100 Index dropped less than 3%. On June 23 alone, SMH plunged 7%, marking its largest single-day decline of the year, while QQQ closed down about 3% that day. This underperformance of the semiconductor ETF relative to the Nasdaq 100 was not merely a systemic market pullback but a concentrated release of structural risks within the semiconductor sector.

Both SMH and QQQ are exposed to the macro risks of the tech sector, so why did SMH suffer significantly deeper losses? Is this excess decline a temporary phenomenon, or is it an inevitable outcome of the inherent independent downside risk of semiconductor ETFs under specific macroeconomic conditions? By examining SMH’s concentration of holdings, marginal changes in AI spending expectations, and valuation levels, we provide a framework for analyzing the independent risk factors and allocation value of semiconductor ETFs.

Where Does SMH’s Independent Downside Risk Come From?

The fundamental difference between SMH and QQQ lies first in the concentration of their underlying assets. QQQ tracks the Nasdaq 100 Index, covering 100 non-financial large-cap tech companies, with industry distribution spanning semiconductors, software, internet services, consumer electronics, and other sectors. SMH, on the other hand, is a pure semiconductor-themed ETF, with holdings heavily concentrated in the world’s leading chip companies—including NVIDIA, TSMC, Broadcom, AMD, and Micron. This concentration is a source of excess returns during bull cycles but becomes a core mechanism for amplifying risk during downturns.

The underperformance of semiconductor ETFs relative to the Nasdaq 100 in June 2026 was driven by a resonance of multiple semiconductor-specific risk factors.

First, a marginal correction in AI infrastructure spending expectations. In early June, Broadcom released its fiscal Q2 2026 earnings, with its guidance for fiscal Q3 AI semiconductor revenue at $16 billion, falling short of Wall Street analysts’ consensus expectation of $17.2 billion—a gap of about $1.2 billion (approximately 7%). Broadcom also projected full-fiscal-year 2026 AI chip sales of $56 billion, again below analysts’ prior average estimate of $57.6 billion. As a major player in the AI chip space second only to NVIDIA, Broadcom’s guidance downgrade was interpreted by the market as a sign that AI infrastructure investment growth may be slowing. This expectation gap triggered a chain reaction in the semiconductor sector: NVIDIA fell about 6%, erasing over $300 billion in market cap in a single day; AMD dropped nearly 11%; Micron plunged over 13%. Broadcom’s own stock fell 11% to 13% in after-hours trading.

Second, cross-market contagion from a slump in Korean tech stocks. On June 23, South Korea’s KOSPI index experienced a sharp sell-off, with foreign institutional investors selling about $2.5 billion worth of KOSPI stocks. Samsung Electronics and SK Hynix—the two dominant players in the global memory chip market—fell sharply. This sell-off quickly spread to U.S. semiconductor ETFs, with SMH plunging 7% that day. The volatility in the Korean market reflected global capital’s concerns about the cyclical peak of the semiconductor sector, and SMH, as a collection of global semiconductor leaders, had almost no immunity to such cross-market risks.

Third, the amplifying effect of leveraged products. As SMH fell, the Direxion Daily Semiconductor Bull 3x ETF (SOXL) plunged about 23% on the same day, reflecting the amplification effect of daily reset leverage. Forced liquidations in leveraged ETFs further intensified the selling pressure on the semiconductor sector, creating a downward spiral. This downside pressure driven by the derivatives market was relatively milder in the broader tech sector represented by QQQ.

The superposition of these three risk factors explains why SMH’s single-day decline on June 23 (7%) was more than double that of QQQ (about 3%).

Why QQQ’s Diversification Advantage Acts as a Buffer

In contrast to SMH’s high concentration, QQQ’s industry diversification played a significant buffering role during this sell-off.

QQQ tracks the Nasdaq 100 Index. While the weight of semiconductor companies in its holdings has increased significantly in recent years due to the AI rally, it is still diluted by industries such as software services, internet platforms, consumer electronics, and biotechnology. When the semiconductor sector faces a systemic sell-off, non-semiconductor components in QQQ—such as Microsoft, Apple, Amazon, and Google—although also affected by market sentiment, have fundamentals less correlated with AI chip spending expectations, resulting in relatively manageable declines.

Looking at the data from June 23, the Nasdaq 100 Index fell 3.3% that day, while SMH plunged 7%. This means SMH’s excess decline was about 3.7 percentage points. Considering SMH’s weight in QQQ, this excess decline directly reflects the independent downside risk of the semiconductor sector relative to the broader Nasdaq 100.

Additionally, while QQQ’s options market also saw a surge in put volume imbalance—out of QQQ’s $3.7 billion in options trading volume, about $2.5 billion was in puts—the absolute scale relative to QQQ’s market cap was much less impactful than the effect of SMH’s options market on SMH itself. In SMH’s options market, totaling several hundred million dollars in premiums, the proportion of puts was abnormally high. This structural difference in the derivatives market further amplified SMH’s vulnerability during the downturn.

The Allocation Value of Semiconductor ETFs: Between Independent Risk and Long-Term Trends

The independent downside risk of SMH does not mean it lacks allocation value. On the contrary, understanding the nature of these risk factors is a prerequisite for formulating a reasonable allocation strategy.

From a long-term perspective, the structural drivers of the semiconductor industry have not disappeared due to the June sell-off. The AI super cycle remains intact—core demand drivers such as GPUs for model training and inference, HBM memory demand, advanced packaging and wafer foundry expansion (led by TSMC), and fiber optic network growth remain strong. SMH represents the infrastructure layer of the AI economy. While the growth rate of this layer may slow during the transition from the infrastructure phase to the deployment phase of AI, the absolute scale is still expanding.

Tom Lee, head of research at Fundstrat, noted on June 24 that historically, when SMH and SOXX experience such sharp single-day declines, the probability of a positive return over the subsequent month is as high as 88%. This statistical pattern has been verified across multiple market cycles, including the rebound after a 35% drop in the semiconductor sector in 2022 and the more than doubling of the sector in the 18 months following the 2020 pandemic. Lee believes that an 88% win rate suggests that sharp sell-offs often attract buyers—the market views such declines as overreactions and buying opportunities.

However, historical patterns do not guarantee future performance. The current valuation of SMH remains elevated. According to GuruFocus’s GF Value model, SMH’s current price is approximately $622.68, while its calculated intrinsic value is about $372.81, implying a premium of around 67%. SMH’s trailing P/E ratio (TTM) is about 15.2x, while its forward P/E ratio is as high as 40.71x. This huge gap in P/E ratios means that the market has priced in extremely aggressive future growth expectations—any signal falling short of expectations could trigger a valuation contraction.

Therefore, SMH’s allocation value needs to balance two dimensions: on one hand, long-term structural growth driven by AI; on the other hand, the independent downside risk brought by extremely high valuations and concentration. For investors seeking exposure to the tech sector, SMH provides the purest semiconductor beta, but the cost is bearing significantly higher sector-specific volatility than QQQ.

Conclusion

The divergence in performance between SMH and QQQ in June 2026 was not random market sentiment fluctuation but a concentrated reflection of risk factors unique to the semiconductor industry. Broadcom’s below-consensus AI revenue guidance triggered a reassessment of the sustainability of AI infrastructure spending; the slump in Korean tech stocks exposed the global linkage and fragility of the semiconductor supply chain; and the amplifying effect of leveraged products intensified the severity of the decline. Together, these factors constituted SMH’s independent downside risk relative to QQQ—a type of idiosyncratic volatility that is difficult to eliminate through diversification at the industry level.

For investors, understanding the risk-return differences between SMH and QQQ is fundamental to making allocation decisions. SMH is the purest beta tool for the semiconductor industry, capable of delivering significant excess returns during the upward phase of the AI super cycle—from the beginning of 2026 to June 3, SMH rose from $360 to $638, a gain of 77%. However, during downturns, its concentration and high valuation make it a concentrated carrier of risk release. QQQ, on the other hand, offers broader tech sector exposure, sacrificing some upside elasticity in exchange for relative downside buffering.

The two are not substitutes but serve different portfolio functions. SMH is suitable as a satellite allocation—playing an offensive role in the overall portfolio, with its weight needing to match the investor’s tolerance for semiconductor-specific risks. QQQ is more suitable as a core allocation—providing more stable tech sector beta and maintaining relatively balanced risk exposure across different market environments. At a time when the AI super cycle is still evolving but valuations are already high, this distinction in allocation logic may be more practically relevant than ever.

FAQ

Q1: What are the main differences between SMH and QQQ?

SMH (VanEck Semiconductor ETF) is a pure semiconductor-themed ETF, with holdings concentrated in global leading chip companies like NVIDIA, TSMC, Broadcom, and AMD, resulting in extremely high industry concentration. QQQ (Invesco QQQ Trust) tracks the Nasdaq 100 Index, covering 100 non-financial large-cap tech companies with a broader industry distribution across software, internet, consumer electronics, etc., offering significantly higher diversification.

Q2: Why did SMH significantly underperform QQQ in June 2026?

On June 23, SMH plunged 7% in a single day, while QQQ fell about 3%. The core reasons include: Broadcom’s AI revenue guidance falling short of expectations, triggering a revaluation of the semiconductor sector; a slump in Korean tech stocks spreading cross-market to U.S. chip stocks; and the amplifying effect of leveraged ETFs exacerbating the selling pressure. These factors had a much greater impact on the highly concentrated SMH than on the diversified QQQ.

Q3: Is SMH’s current valuation too high?

According to GuruFocus’s GF Value model, SMH’s current price is about $622.68, with a calculated intrinsic value of about $372.81, representing a premium of around 67%. Its trailing P/E ratio is about 15.2x, while its forward P/E ratio is as high as 40.71x, indicating that the market has priced in extremely high future growth expectations, posing a risk of valuation contraction.

Q4: Is the long-term growth logic of the semiconductor sector still intact?

The core drivers of the AI super cycle remain intact—long-term trends such as GPU demand, HBM memory, advanced packaging, and wafer foundry expansion have not changed. The market is experiencing a transition from the AI infrastructure phase to the deployment phase, with growth expectations normalizing from extremely high levels. The long-term structural growth logic has not been broken, but short-term volatility and valuation resets are inevitable.

Q5: How should investors allocate SMH and QQQ?

SMH is suitable as a satellite allocation, playing an offensive role, with its weight needing to match the investor’s tolerance for semiconductor-specific risks. QQQ is more suitable as a core allocation, providing more stable tech sector beta. The two are not substitutes but serve different portfolio functions—this distinction is particularly important at a time when the AI super cycle is still evolving but valuations are already high.