Global Semiconductor Latest Outlook

In 2026, the semiconductor industry is facing a risky paradox. The strong demand driven by artificial intelligence is pushing revenues to unprecedented heights, but this prosperity also comes with risks. The industry seems to have put all its eggs in the artificial intelligence basket, which might be justifiable if the boom continues. However, the industry should also consider how to deal with a slowdown or contraction in artificial intelligence demand. Current Market Situation Global semiconductor industry sales are expected to reach $975 billion by 2026, setting a historic new high, mainly due to the booming development of artificial intelligence infrastructure (Figure 1). In 2025, the growth rate of the semiconductor industry reached 22%, and it is expected to accelerate to 26% by 2026. Even if the growth rate slows down afterwards, annual sales are expected to reach $2 trillion by 2036. However, this record growth masks a significant structural difference. While high-value artificial intelligence chips currently contribute about half of total revenue, their sales volume accounts for less than 0.2%. Another difference is that while artificial intelligence chip development is booming, the growth rate of chips used in cars, computers, smartphones, and non-data center communication applications is relatively slowing down. The stock market is usually a leading indicator of industry performance. By mid-December 2025, the total market value of the top ten chip companies globally reached $9.5 trillion, a 46% increase from mid-December 2024 and a 181% increase from mid-December 2023. Additionally, this market value is highly concentrated, with the top three chip companies accounting for 80% of the total market value. By the time of publication, Deloitte predicts that by 2026, revenue from generative artificial intelligence chips will approach $500 billion, accounting for about half of global chip sales. Furthermore, AMD CEO Lisa Su has raised expectations for the potential market size of data center artificial intelligence accelerator chips to $1 trillion by 2030. It is expected that chip sales will reach 10.5 trillion units by 2025, with an average selling price of $0.74 per unit. Although artificial intelligence chips may account for about 50% of industry revenue in 2026, their production is less than 20 million units, representing around 0.2% of total sales volume. While global chip revenue is expected to grow by 22% in 2025, silicon wafer shipments are only expected to increase by 5.4%. In terms of major end markets, sales of personal computing devices and smartphones were originally expected to grow in 2025, but due to rising memory prices, a decline is expected in 2026. It is expected that storage revenue in 2026 will reach around $200 billion, accounting for 25% of total semiconductor revenue that year. The storage market has always been cyclical, as manufacturers seem to be cautious about overcapacity. Therefore, they are only moderately increasing capital expenditures, with most of it going towards the development of new products rather than large-scale production expansion. As a result, the demand growth for high-bandwidth memory 3 (HBM3), HBM4, and DDR7 memory for artificial intelligence inference and training solutions has led to a shortage of consumer-grade memory such as DDR4 and DDR5; the prices of these products increased by about four times between September and November 2025. Predicting the supply, demand, and prices of memory is very challenging, but some believe that the current shortage of consumer-grade memory may last for a decade. Price increases are...