Omdia has released its February semiconductor market snapshot. The report identifies three primary supply bottlenecks currently constraining production. The advanced packaging sector is facing a crisis. Taiwan Semiconductor Manufacturing's CoWoS capacity is expanding from 75,000 wafers per month in 2025 to 120,000-130,000 by the end of 2026. However, this increase remains insufficient to meet the surge in demand, ultimately leading to extended delivery times, price increases, and capacity allocation being skewed towards the largest customers. Concurrently, a global skills gap is evident: despite significant investments in Europe and the US, specialized packaging expertise remains scarce, severely hindering regional supply chain autonomy goals. Memory shortages are exacerbating a vicious cycle in the supply chain. By mid-2025, DRAM prices had increased by 50%; within this, DDR4/DDR5 prices saw quarterly increases of 30% to 50%, and the cost of memory modules had also risen by 50% by the second quarter of 2025. However, supply growth is limited, with DRAM expected to grow by only 16% and NAND by 17%, indicating a worsening supply-demand imbalance. There has been chronic underinvestment in capacity for certain mature nodes. As manufacturers shift focus to more profitable advanced processes, capacity for some mature technologies faces underinvestment and aging production line equipment. These less glamorous components from mature nodes are critical for automotive, industrial, and consumer electronics products. Years of underinvestment combined with lengthy product certification cycles have resulted in a lack of new capacity. Chip companies face high costs for redesigns and new tape-outs, further highlighting the fragility and imbalance of the supply chain.
February Snapshot: Semiconductor Market Update
NVIDIA Enters AI Laptop Chip Market: NVIDIA is re-entering the consumer PC market with AI-centric notebook processors. At least eight notebooks featuring its chips are expected this spring, primarily targeting the new generation of Windows systems. These processors integrate CPU, GPU, and AI acceleration engines, are compatible with Windows-on-Arm, and aim to enhance efficiency and performance.
Texas Instruments Acquires Silicon Labs: Texas Instruments announced the acquisition of Silicon Labs for $7.5 billion to strengthen its embedded wireless connectivity portfolio. The acquisition will enhance Texas Instruments' manufacturing capabilities and customer experience.
Dutch Court Investigates Nexperia: A Dutch court announced an investigation into the governance of Nexperia, limiting the control of its Chinese parent company, Wingtech. Political disputes between the Netherlands and China are disrupting Nexperia's supply chain, forcing the company to temporarily halt shipments.
Infineon's Strategic Shift to AI: Infineon announced that due to slow recovery in the automotive and industrial sectors, the company will increase investment in AI power capabilities and shift resources away from automotive and industrial areas. The company is converting IGBT capacity to MOSFET capacity to support AI applications, aiming to achieve full utilization of its 300mm wafer fabs.
Memory Market and Automotive Industry Recovery: Infineon addressed concerns about memory market tightness, stating its lower reliance on products with high DRAM content, such as smartphones and notebooks. Automotive industry inventory has normalized, and market visibility is improving, but order patterns remain cautious due to concerns about AI-driven capacity spillover effects.
Infineon Plans to Acquire ams Osram Sensor Business: Infineon plans to acquire the non-optical sensor business of ams Osram to strengthen its position in automotive, industrial, and medical applications. This acquisition aligns with Infineon's strategy of providing system-level solutions and diversifying revenue sources.
The 2026 Semiconductor Supply Chain Landscape
In January 2026, the World Economic Forum annual meeting in Davos, Switzerland, confirmed the growing importance of semiconductor supply chains. The meeting emphasized that the global semiconductor industry is experiencing unprecedented growth while facing severe supply constraints. Although the market size is projected to reach $1.06 trillion in 2026 (a 30.7% increase from $810.5 billion in 2025), this expansion masks a fundamental imbalance between surging AI demand and demand from traditional electronics sectors. Key market dynamics indicate the industry's trajectory is almost entirely dominated by AI applications, creating a polarized market structure. The second half of 2026 is seen as a critical period, with supply constraints expected to intensify significantly. By mid-2025, DRAM prices had surged by 50%, while prices for related equipment increased between 15% and 60%. This is not merely an issue of rising costs but is fundamentally reshaping the industry landscape, determining which products can be manufactured and which companies can secure critical components.
The Dominance of AI
The dominance of AI is radically reshaping the entire semiconductor supply chain. AI applications are fundamentally altering manufacturing priorities, sparking a zero-sum competition for limited capacity. High Bandwidth Memory (HBM) now accounts for 23% of total DRAM wafer output, a significant increase from the previous 19%, with demand growing year-on-year by 70%. Meanwhile, tech giants like NVIDIA and Apple have already secured large capacity quotas through the end of 2026, leaving second-tier customers scrambling for remaining supply. Micron's entire HBM supply for 2026 was already sold out, fully highlighting the严峻 challenges in capacity allocation. In this market structure, winners and losers are increasingly distinct: AI-centric companies are thriving, with NVIDIA's Q4 2025 revenue growing 20% sequentially; memory manufacturers Samsung, SK Hynix, and Micron also saw Q4 2025 revenue grow between 21% and 34% compared to Q3. In contrast, traditional sectors are generally struggling—Texas Instruments, Infineon, Sony Imaging, and onsemi all reported revenue declines; Intel's Q1 2026 revenue fell 11% due to memory shortages in the PC market; Qualcomm and MediaTek also experienced performance declines due to constrained memory supply for smartphones. The automotive industry is particularly vulnerable. With lower profitability compared to AI customers and chip certification cycles lasting up to a decade, automotive manufacturers are being systematically deprioritized in production. To mitigate risk, companies like Tesla have adopted a dual-foundry strategy with Taiwan Semiconductor Manufacturing and Samsung for sourcing AI chips to ensure the stability of their critical supply chain.
Comments