KPMG: AI brings structural opportunities, multiple real risks constrain cycle sustainability
The Zhitong Finance App learned that KPMG released the “2026 Global Semiconductor Industry Outlook” report. According to the report, AI is not spawning a short-term bubble, but rather a structural shift in the market, driving demand expansion throughout the entire industry chain. At the same time, the industry is also facing multiple realistic constraints. Tariff and trade policy risks have become the industry's primary concern. The power supply for advanced chip manufacturing, the high-end talent gap, and supply chain restructuring pressure have become the main variables limiting the development of the industry. Over 60% of the companies surveyed plan to expand capital expenditure, but they also listed improving supply chain resilience as a core strategic task for the next three years. At the same time, the report suggests that although market optimism is high, whether the supercycle can continue depends on whether enterprises can balance production capacity expansion, capital investment, and multiple risks such as geography, energy, and talent.
In the context of the supercycle reshaping the semiconductor industry, risks and opportunities coexist
For the first time, artificial intelligence became a major revenue engine. According to the survey, 73% of business leaders see artificial intelligence as a major factor in their revenue growth, up from 67% last year, and ahead of cloud computing and data centers. Demand for artificial intelligence computing has surged, driving up demand for memory: 67% of respondents believe that memory has the greatest potential for growth, which is roughly the same as 66% of microprocessors. This is the first time this has happened in the survey's history. This shows that artificial intelligence is driving the overall development of the market, and computing, storage, and mobile data chips are growing at the same time.
Semiconductor industry leaders have made supply chain flexibility a top strategic priority in dealing with geopolitical pressures and trade uncertainty. 45% of respondents said that for the first time in three years, the emphasis on improving the responsiveness and flexibility of the supply chain has surpassed talent development. Its development priorities are clear and clear, namely combining internal construction and outsourcing, implementing digital planning, and shortening the overall supply chain response cycle.
Business leaders are introducing artificial intelligence internally to address the talent gap. 66% of businesses plan to use artificial intelligence to improve employee productivity, including improving team skills, automating routine tasks, and freeing up scarce engineering and operations talent to focus on high-value work. These pro-growth measures are not only at the enterprise level, but also in the industry as a whole.
Although the risks have become apparent, confidence continues to rise. In 2025, the semiconductor industry confidence index rose from 59 in 2024 to 63, reaching the third highest level since we conducted this survey 21 years ago. Many businesses are developing growth plans, and most of them expect to increase capital expenses, headcount, and IT investments. This optimism stems from strong cross-quarter demand signals, although business leaders remain wary of geopolitics, trade frictions, and resource constraints.
Although research shows that artificial intelligence is beginning a semiconductor supercycle, many technology leaders still have concerns about demand and its sustainability, which are related to supply chain restrictions, energy supply, and geopolitical instability. To maintain resilience, semiconductor companies should build flexible business models, diversify supply chains to reduce geopolitical risks, and increase R&D investment to lead the next round of structural and application changes. At the same time, companies should also use artificial intelligence to empower employees so they can adapt to the way artificial intelligence works. In short, only with the right strategy can business leaders lay the foundation for the strong growth of the company's performance over the next ten years.
In the context of the intertwined global economy, artificial intelligence is leading an unprecedented boom, attracting huge investment, and promoting the rapid development of chip manufacturing and technological ecosystems. At the same time, traditional industries are facing challenges such as layoffs, trade blockages, and fiscal austerity, in stark contrast. The semiconductor industry is at the center of this transformation. Artificial intelligence has surpassed cloud computing to become the semiconductor industry's biggest revenue driver. However, leveraged investments, macroeconomic fluctuations, and global events have also raised concerns about the sustainability of demand. Research shows that there is real demand for artificial intelligence in the market, not short-term hype. The survey also showed that if business leaders can build flexibility, diversify supply, and use artificial intelligence to power operations, it will help the company gain a significant advantage over the next ten years.
Artificial intelligence becomes a major driver of revenue growth
73% of respondents believe artificial intelligence is their biggest driver of revenue growth, up from last year. The wide range of applications of artificial intelligence (mainly in scenarios such as defect detection, sales proposal generation, and autonomous driving) all rely on the high-speed data processing capabilities provided by semiconductors. As the “muscle and memory” of artificial intelligence systems, chips determine whether they can achieve efficient decision-making and large-scale implementation. As a result, chip selection is gradually becoming the starting point for many companies' product roadmaps and strategic planning. Artificial intelligence is rapidly and broadly driving the growth of the semiconductor industry. This far exceeds previous technological waves, and its importance has surpassed cloud computing. According to the data, the global artificial intelligence semiconductor market is expected to exceed 438.5 billion US dollars by 2029, with a five-year compound growth rate of 25.9%. Artificial intelligence is not only an additional function, but also a core revenue driver. The key challenge now is whether the supply chain has sufficient responsiveness.
The juxtaposition of memory and microprocessor has become the biggest growth opportunity point
In the latest semiconductor industry survey, memory solutions were at the forefront of growth for the first time, and were mentioned by 67% of executives, becoming one of the fastest growing technology fields, significantly narrowing the gap with microprocessors. This trend is mainly driven by artificial intelligence, as demand for high bandwidth and high capacity memory continues to rise. Tech giants such as Amazon, Google Cloud, and Microsoft Azure are speeding up the deployment of high-bandwidth memory, driving the industry towards advanced memory solutions. Some DRAM manufacturers are also shifting to producing high-bandwidth memory to meet market demand. This transformation not only exacerbates the tight supply of DRAM, but also marks a deep reshaping of the competitive landscape in the semiconductor industry chain. Artificial intelligence is no longer an add-on, but a major force that can generate revenue growth. As artificial intelligence drives overall demand growth in the industry, whether the supply chain can respond quickly becomes the next strategic issue.
Supply chain has become a top topic
For the first time in three years, the supply chain has become a top priority for the semiconductor industry, according to the survey. This shift reflects the real challenges companies face in facing the risk of exposure to the pandemic, geopolitical tensions, and difficulties in obtaining critical raw materials such as rare earths. 45% of respondents believe that increasing the responsiveness and flexibility of the supply chain has become a top priority, while geopolitical uncertainty is also viewed by 37% as the third biggest risk in the next three years.
Is energy becoming a bottleneck? As demand driven by artificial intelligence surges, energy supply has become an important supply-side issue. 34% of respondents are concerned that they will not be able to provide sufficient electricity for their own production equipment within the next three years, while the challenges facing data centers are even more serious. 58% worry that tech giants will not be able to obtain enough energy to support the expansion of AI infrastructure.
Collaborative technology and energy response: To address energy restrictions, companies are using artificial intelligence to optimize energy efficiency and increase data center capacity without increasing energy consumption. At the same time, the global semiconductor industry is highly dependent on a complex global network of specialized suppliers and manufacturing centers, and the government will intervene more in choosing the method and location of chip manufacturing. This complex situation also brings new opportunities and risks.
Resilience cultivation and regional diversification strategies: In order to enhance supply chain resilience, 54% of respondents plan to expand the regional distribution of the supply chain within the next 12 months to ensure that supply can still be guaranteed when problems occur in a certain region. Additionally, 36% of respondents plan to use generative artificial intelligence in procurement and supply chain management, using predictive analysis and automation tools to address shortages, adjust inventory, and ensure production capacity.
Artificial intelligence aims to empower, not replace, employees
Artificial intelligence is empowering rather than replacing employees, helping semiconductor companies address talent shortages and increase productivity. As human-robot collaboration deepens and the definition of competitive advantage continues to evolve, semiconductors are becoming an important pillar in the new world of artificial intelligence. Currently, within semiconductor companies, artificial intelligence is widely used in IT, R&D, supply chain, and marketing fields. 44%, 36%, and 30% of enterprises deploy artificial intelligence to optimize processes, predict demand, and accurate marketing, respectively. Only companies that pioneer changes in products, operations, and market strategies can lead the next ten years.
Artificial intelligence has not eliminated the need for human employees in businesses. Interviewees pointed out that talent has become a major challenge after tariffs, and cultivating and retaining talents is one of the top strategic tasks of enterprises.
Optimists are not afraid to take risks and move forward at full speed
In the face of geographical and resource challenges, semiconductor companies' confidence continued to grow, and the confidence index rose to 63, the third highest in 21 years. Market expectations are positive, and the global market is expected to reach 1 trillion US dollars in 2026. Since the end of 2022, sales have increased by nearly 50%, and more than half of companies expect revenue growth of more than 11% in the next year, and plan to drive growth through recruitment, IT upgrades, and mergers and acquisitions. Despite this, 58% of companies are still concerned about fluctuating customer demand and are wary of critical resource risks. To achieve steady growth, companies are optimizing orders, product portfolios and investment rhythms, expanding production capacity, strengthening talent, and hedging risks through diversified supply and prudent capital allocation, and striving to turn optimism into sustainable profits.
Figure: Global semiconductor market and industry confidence (unit: billion US dollars)
Source: Global semiconductor industry survey conducted jointly by KPMG and the Global Semiconductor Alliance, Q4 2025
Artificial intelligence is no longer an emerging opportunity for semiconductor companies, but rather the industry's single most important revenue driver. This transformation requires business leaders to rapidly adjust product, operational, and capital strategies around artificial intelligence, and adapt flexibly to regional dynamics. To this end, it is recommended to start with the following aspects:
Seize the growth opportunities where demand for memory and microprocessors go hand in hand: The surge in demand for memory products brings important opportunities. If business leaders can adjust products, customer goals, and energy strategies in a timely manner to meet emerging needs such as large-scale data centers, they will gain an advantage in continuous growth.
Increase supply chain flexibility to deal with geopolitical turbulence: In the face of trade barriers and resource risks, companies must enhance supply chain flexibility, reduce weak links, and improve responsiveness to maintain innovation and competitiveness.
Using artificial intelligence to enhance talent resilience and productivity: As generative artificial intelligence and proxy artificial intelligence reshape all functional areas of semiconductors, business executives must develop strategies that balance talent and technology to speed up decision-making, optimize processes, and drive innovation.
Maintaining strategic growth amid economic and geopolitical uncertainty: Despite high industry confidence and record high capital expenditure, business leaders still need to balance active expansion with risk control, avoid excessive investment, and ensure competitive advantage in a high growth cycle.
Li Jiming, KPMG's leading partner in China's semiconductor industry, said that as the world's largest semiconductor consumer market, China is entering the beginning of the “15th Five-Year Plan.” The plan emphasizes technological self-reliance and self-improvement, new quality productivity, modern industrial systems, and the resilience and safety of the industrial chain supply chain, and semiconductors are placed at the core. Currently, China's semiconductor industry is showing multiple forces in parallel: artificial intelligence computing power, smart cars, industrial automation, and consumer electronics demand; mature process expansion, advanced packaging and Chiplet exploration, and domestic verification of equipment materials are accelerated; at the same time, advanced manufacturing processes, high-end EDA/IP, and key equipment are still being constrained. China's policy orientation, capital expenditure, and domestic production substitution process will profoundly affect global supply and demand, technology routes, mature process competition, and the layout of multinational enterprises, and may also change the shape and pace of the “supercycle.”
Li Jiming believes that “the semiconductor industry super cycle has begun?” The answer is not a simple “yes” or “no.” A more accurate judgment is that it is currently in the early stages of the supercycle, but its development will be conditionally dependent, hierarchical, and regionally unbalanced. The effects of the implementation of China's “15th Five-Year Plan”, the return on investment in global trade and manufacturing, the return on investment in artificial intelligence, and the extent to which supply chain bottlenecks have been mitigated will determine whether this round of growth is a short-term rebound or long-term structural restructuring. The strategic choices made by business leaders now will not only determine the company's own development trajectory, but will also profoundly influence the future pattern of global technology development.