Top Three Sources' Views and TSO Validation Conclusion
Top Three Sources' Views and TSO Validation Conclusion: All three sources point out that the global semiconductor market is expected to reach a historic peak of $975 billion in 2026, driven by AI infrastructure. The core consensus is that the industry is in a conflict between high growth and structural divergence. The main disagreement lies in concerns about the sustainability of AI demand and growth forecasts for non-data center markets. TMT Outlooks emphasize the importance of risk management, suggesting the industry should consider scenarios where AI demand slows down and pay attention to potential disruptions from power and innovation.
Facts Confirmed by All Sources
According to Source 1 (Deloitte Insights), the global semiconductor industry is projected to reach annual sales of $975 billion in 2026, a milestone driven by the AI infrastructure boom. The industry is expected to grow by 22% in 2025 and accelerate to 26% growth in 2026. Furthermore, Source 1 notes that AI chips currently account for about half of total revenue but less than 0.2% of unit sales, indicating a structural divergence of high value and low volume. Additionally, Source 1 mentions that 10.5 trillion chips are expected to be sold in 2025, with an average unit price of $0.74.
Main Disagreements or Differences
Source 1 raises a warning about the risk of AI growth models, pointing out that if AI demand slows down or declines, data center projects might be canceled or postponed, negatively impacting chip sales. Source 1 further emphasizes supply chain risks, especially the shortage and rising price trend for key components like memory (HBM3, HBM4, DDR7), predicting a potential 50% price increase in the first half of 2026. In contrast, Source 1 also mentions that sales in non-data center markets like PCs and smartphones might decline in 2026 due to rising memory prices.
Background and Analysis
The core contradiction facing the semiconductor industry is the balance between "AI demand explosion" and "structural risks." Although AI chips are the main driver of current market growth, their proportion in unit sales remains very low, increasing the industry's sensitivity to demand slowdowns. Moreover, bottlenecks in power supply and infrastructure (such as the 92 GW of additional power required for AI data centers) and supply tightness in key areas like memory constitute structural challenges. Source 1 suggests that the focus for industry leaders should no longer be solely on chasing AI demand, but on managing the systemic risks of high-margin, low-volume paradigms, including component shortages and supply chain reshaping.
Summary of Three Sources' Views
Source 1 (Deloitte Insights): Sales reach $975 billion in 2026, but structural divergence is evident; focus on the risk of slowing AI demand; memory prices and supply tightness are key challenges; strategy should shift from chasing AI to risk mitigation and system architecture.* Source 1 (Deloitte Insights): Sales are expected to reach $975 billion in 2026, but there will be significant structural divergence; concerns about slowing AI demand; memory prices and supply shortages are key challenges; strategy should shift from chasing AI to risk mitigation and system architecture.
Source 2 (Deloitte Insights): AI chip revenue is projected to account for half of the revenue in 2026, but its share of unit sales is very low; concerns about ROI for data center projects; focus on power supply and efficiency gains from technological innovation.
Source 3 (Deloitte Insights): Emphasizes the need to plan for scenarios where AI demand slows down; points out slowing growth in non-data center markets (PC, mobile, automotive); memory supply and demand relationship may persist for a decade; improvements in AI model efficiency may lead to changes in chip demand.