Semiconductor Earnings Soar 131% as SMH Drops 23%, AI Demand in Focus
Ted Hisokawa Sep 22, 2026 13:03
Semiconductor earnings surged 131% in Q2 2026, yet SMH fell 23%. AI demand drives growth, while geopolitics and capital cycles shape investor sentiment.
The semiconductor sector continues to deliver record-breaking growth, with industry earnings up 131% year-over-year in Q2 2026, according to VanEck data. Despite this, the VanEck Semiconductor ETF (SMH) plunged 23% over the same period, reflecting broader concerns about the durability of the current AI-driven boom.
Market data underscores the scale of this growth cycle. On September 14, Omdia reported global semiconductor revenue reached $425 billion in Q2 2026, a 31.4% sequential increase, and projected Q3 revenue to surpass $500 billion. Gartner’s August 2026 forecast is even more bullish, predicting $1.6 trillion in annual semiconductor revenue for 2026, nearly doubling 2025’s $809 billion. Much of this growth is tied to AI infrastructure demand, including high-bandwidth memory (HBM), advanced packaging, and cutting-edge foundry capacity.
Yet SMH's recent underperformance reflects a divergence between operational strength and investor sentiment. While companies are reporting strong fundamentals, the market appears to be pricing in concerns about valuation, cyclicality, and geopolitical risks. The ETF has delivered a robust year-to-date return of 82.3% as of June 30, 2026, but its Q2 drawdown signals skepticism over how sustainable the AI-driven surge is.
Geopolitics remains a critical factor shaping the industry. On September 17, Huawei unveiled new chip technologies, highlighting China's accelerated push for semiconductor self-reliance. This effort is driven partly by U.S.-led restrictions on advanced AI chips and manufacturing equipment. Meanwhile, the United States continues to incentivize domestic semiconductor manufacturing through federal subsidies and export controls, further complicating global supply chains.
For traders and investors, the key takeaway is the structural shift from traditional semiconductor cycles toward AI-driven demand. Memory pricing, particularly for high-bandwidth memory used in AI applications, remains a pivotal driver for future earnings. Companies involved in GPU production, data-center networking, and custom ASICs are expanding capacity aggressively, positioning themselves to capitalize on long-term AI trends.
As of September 22, 2026, the VanEck Semiconductor ETF (SMH) is trading at $559.34, up 4.93% in the last 24 hours. This rebound suggests that while the sector faces short-term uncertainty, investor appetite for AI-related growth remains strong. With Q3 earnings data and further geopolitical developments likely to drive volatility, the semiconductor sector is a critical space to watch in the coming months.
Image source: Shutterstock.jpg)