Profit differentiation in the semiconductor industry by 2025

As the 2025 annual report disclosure season gradually peaks, many companies in the semiconductor industry have turned in impressive “scorecards.” According to data from Tushare, as of March 30 when the reporter’s copy was issued, among 173 semiconductor companies categorized under the Shenwan industry classification, 42 companies had officially released their 2025 annual reports. Of them, 29 companies saw year-over-year growth in parent-attributable net profit. A total of 117 companies released performance briefings, of which 76 companies recorded year-over-year growth in parent-attributable net profit. Overall, buoyed by the catalyst of a surge in demand from artificial intelligence (AI), the industry’s overall business sentiment is trending upward. However, behind the “spotlight,” the profitability of each link in the industrial chain shows a clearly differentiated, structural pattern. Industry experts say this differentiation is not simply a matter of cyclical fluctuations, but a vivid reflection of the AI-driven semiconductor industry shifting from “broad-based gains and broad-based declines” to “structural upturn.”

“Chip” opportunities emerge in storage and other sub-sectors

Benefiting from the acceleration of AI large-model commercialization, demand for computing power has grown exponentially, driving “multi-point blossoming” performance across certain sub-sectors. Among them, storage chips are seeing both volume and price rise—an encouraging “uptrend” in their outlook.

2025 performance at Beijing Unigen Storage Technology stood out particularly. The company achieved operating revenue of 11.302 billion yuan, up 68.82%; it recorded parent-attributable net profit of 853 million yuan, up 429.07%. Among this, revenue from AI emerging edge-side storage products was approximately 1.751 billion yuan, up significantly year over year. This strong growth momentum continued into early 2026. The company expects to achieve parent-attributable net profit of 1.5 billion to 1.8 billion yuan from January to February 2026, representing an increase of 921.77% to 1086.13% year over year; this figure has already surpassed the level of full-year 2025 net profit.

When speaking to reporters from Economic Reference News, Zhu Keli, founder and president of the Institute of New Economy Research, said that the simultaneous rise in volume and price of storage chips is mainly rooted in the explosive growth of emerging demand such as AI compute power and data centers. Demand for storage from AI servers is 8 to 12 times that of traditional servers. Coupled with leading companies shifting capacity toward higher-end products, this triggers structural supply-demand imbalance, pushing prices to historical highs. At the same time, the storage industry itself is in an upward cycle phase, further amplifying earnings leverage.

This compute-power-driven market demand also flows upstream to the semiconductor equipment sector. Based on performance briefing data that has already been disclosed, in 2025, 15 listed semiconductor equipment companies collectively generated revenue of 42.812 billion yuan, up 30.97%, and collectively achieved parent-attributable net profit of 6.305 billion yuan. Among all sub-sectors, their growth rate led the pack. Among them, Micro-Methods (Micro) Co., Ltd. ranked firmly among industry leaders with 12.385 billion yuan in revenue and 2.111 billion yuan in parent-attributable net profit. Luthai Technology’s revenue growth rate reached as high as 58.9%, and market demand for thin-film deposition equipment has been fully validated. Huawei Hai Qingke’s revenue grew 36.5% year over year, and the company has continued to maintain a competitive advantage in CMP equipment.

In the wafer foundry segment, leading companies delivered steady performance. According to annual report disclosures, Semiconductor Manufacturing International Corporation (SMIC) achieved operating revenue of 67.323 billion yuan in 2025, up 16.5%; and parent-attributable net profit of 5.041 billion yuan, up 36.3%. The company said that changes in operating revenue were mainly due to an increase in wafer shipments during the year. United Microelectronics Corporation achieved operating revenue of 10.885 billion yuan in 2025, up 17.69% from the same period last year; parent-attributable net profit was 704 million yuan, up 32.16% from the same period last year.

However, the industry’s improvement in sentiment is not distributed evenly across the supply chain. For example, from the performance briefings, many companies in the semiconductor materials segment saw “cooling” performance. Seventeen companies collectively generated revenue of 23.504 billion yuan, up 16.95%, but their combined parent-attributable net profit was -472 million yuan. In 2025, Husr Silicium Materials Technology reported losses of 1.476 billion yuan, Xi’an Yicai reported losses of 738 million yuan, and Tianyue Advanced shifted from profit to loss, posting losses of 208 million yuan, all facing loss pressure to varying degrees.

AI reshapes the industry and the value logic

If the core main line of the semiconductor industry in 2025 is distilled, it is undoubtedly “AI monetization.” Looking at the overall data, the company groups directly related to AI compute power and storage contributed the vast majority of the industry’s incremental profit.

As a benchmark enterprise in China’s domestic AI chip sector, Cambricon has experienced a “spotlight moment” since its listing on the STAR Market. According to its annual report, the company achieved full-year operating revenue of 6.497 billion yuan, up sharply 453.21% year over year; it recorded parent-attributable net profit of 2.059 billion yuan. This is the first time the company has achieved annual profitability since it went public in 2020.

Meanwhile, driven by factors including the development of the AI industry and strong demand in the high-performance GPU market, China’s domestic AI chip tier as a whole has shown a favorable earnings trend. The annual report shows that in 2025, Moxicis achieved operating revenue of 1.644 billion yuan, up 121.26% from the same period last year; parent-attributable net profit was -789 million yuan, narrowing its year-over-year loss by 43.97%. According to performance briefings, Moore Threads’ 2025 revenue was 1.506 billion yuan, up 243.37% year over year; parent-attributable net profit was -1.024 billion yuan, narrowing its year-over-year loss by 36.70%.

Overall, thanks to continuous technological iteration, many companies in the domestic AI chip field have successfully moved beyond the “technology validation” stage and entered the commercialization turning point of “batch shipments.” The deeper logic behind this shift lies in the large-scale implementation of AI application scenarios.

Regarding AI’s deep impact on the semiconductor industrial chain, Tian Lihui, dean of the Institute of Financial Development at Nankai University, pointed out that AI’s reshaping of the semiconductor industry is evolving from “single-point stimulation” to “underlying reconstruction.” On the demand side, AI is shifting from compute-power training to the inference stage, driving semiconductors from “general-purpose computing” toward “heterogeneous computing,” and exponentially increasing demand for high-bandwidth storage, advanced packaging, and application-specific chips. The essence of AI reshaping is to move semiconductors from “pricing by the number of transistors” to “pricing by system performance.” This is both a challenge and an unavoidable path for industrial upgrading. The industrial logic is moving from “process miniaturization” to “system-level innovation.” Companies need to build new “moats” through ecosystem integration and deep cultivation of use cases to move steadily through the computing revolution.

“AI’s penetration across the entire semiconductor industry chain is reconstructing, from the ground up, the industry’s competitive barriers and value-allocation logic—shifting competition from solely trading off one technology or capacity metric against another to a comprehensive contest of intelligent capabilities across the full chain and ecosystem coordination,” Zhu Keli added. He further said that overall, AI is making competitive barriers in the semiconductor industry more diverse, and value allocation is tilting toward segments and companies where “AI capabilities” are more prominent and ecosystem coordination is stronger.

The industry is moving toward “structural upturn”

Although overall industry performance is improving, beyond the AI “spotlight,” the semiconductor industry internally is showing a clearly differentiated trend.

A recent semiconductor industry performance tracking special report released by Dongguan Securities shows that, based on companies whose performance or performance forecasts have already been disclosed at home and abroad, the semiconductor industry’s overall business sentiment is rising, but sub-sectors are differentiating. Specifically, demand-side pull from AI in sub-sectors such as compute-power chips, storage chips, and wafer foundry is evident, while sub-sectors not closely related to AI are showing a more moderate recovery trend. In some sub-sectors—such as consumer electronics—business sentiment is under pressure overall because AI demand crowds out resources such as storage, leading to higher costs.

Regarding the phenomenon of performance differentiation, Zhu Keli said this is an inevitable outcome of the overlapping effects of industry cyclical recovery and structural demand reconstruction. At its core, as new and old growth drivers are transitioning, the supply-demand dynamics and value logic in different links of the industrial chain undergo fundamental changes. This differentiation is not merely a cyclical fluctuation, but a vivid manifestation of the AI-driven semiconductor industry shifting from “broad-based gains and broad-based declines” to “structural upturn.” The links led by emerging demand are capturing development dividends, while the links relying on traditional demand are experiencing the pains of transition.

In Tian Lihui’s view, performance differentiation is an inevitable result of the semiconductor industry’s cyclical recovery combined with structural change. The key reasons mainly concentrate in three aspects: differentiation in demand structure, misalignment in inventory cycles, and differences in capacity and pricing power. He emphasized that the differentiation is not a short-term phenomenon, but a new normal in industrial evolution. This is a process in which the industry moves from “universal growth” to “targeted prosperity,” where high-value segments seize resources and traditional segments speed up reshuffling—an unavoidable period of pain as China’s semiconductor industry advances toward high-quality development.

Zhu Keli believes that based on the 2025 annual reports and performance guidance for the first quarter of 2026, the pattern of earnings differentiation in the semiconductor industry will further intensify in the short term, and over the medium to long term it will gradually converge as some links complete their transformations and supply-demand is restructured. Its core reason is that the AI-driven structural demand dividend is still being released primarily, while traditional demand recovery is slow and cost pressure remains. As a result, development paces across different links cannot synchronize.