Skip to main content
GPUBeat Frontier Models AI Demand Drives Shift in Semiconductor…

AI Demand Drives Shift in Semiconductor Manufacturing Capacity

A report from M Stanley indicates a forthcoming upcycle in mature-node semiconductor processes, driven by rising AI infrastructure demands, creating new opportunities for second-tier foundries.

Near AI — ai-infrastructure — Near AI
AI Demand Drives Shift in Semiconductor Manufacturing Capacity Source: GPUBeat

The semiconductor industry is on the brink of a significant shift as demand for AI-related technologies continues to reshape manufacturing strategies. Major players like Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung Electronics are consolidating their mature-node capacities to focus on advanced nodes and packaging, responding to the surging requirements of AI infrastructure. This strategic pivot is expected to create new opportunities for second-tier foundries and specialty process players looking to capture spillover demand.

M Stanley's report highlights that the anticipated upcycle in mature-node processes is nearing, driven by an ongoing buildout of AI infrastructure that could last for another two years. Smartphone and personal computer manufacturers are expressing concerns about potential shortages in mature-node capacity. The report indicates that, given the current market dynamics, significant inventory corrections or order cuts for mature-node foundries are unlikely in the near future.

The implications of these developments are noteworthy. The total addressable market (TAM) for AI server power integrated circuits (ICs) is projected to reach $15 billion by 2026, with a compound annual growth rate (CAGR) of 50% over the next two years. AI power ICs are poised to play an important role in supporting the semiconductor sector. M Stanley suggests that this demand for AI power ICs should sufficiently offset the sluggish demand observed in consumer and smartphone chip markets. Recently, a major US power management integrated circuit (PMIC) customer of HUA HONG SEMI reported a tripling of wafer demand on its 55nm BCD platform, highlighting stable growth potential in the AI sector.

Additionally, M Stanley has revised its price targets for key market players. For SMIC (00981.HK), the target price has increased from HKD70 to HKD85, backed by an Overweight rating. The bank has also raised its earnings per share (EPS) forecasts for 2026 through 2028 by 11%, 14%, and 10%, respectively, reflecting a positive outlook driven by strong second-quarter revenue and gross margin guidance. This optimism stems from advances in the company’s advanced-node capacity and a favorable product mix, along with price increases on the BCD platform.

See also  Emerging AI Startup Gains Traction in Crypto Sector

HUA HONG SEMI has seen its price target rise from HKD88 to HKD118, while maintaining an Equal-weight rating. Although this year’s EPS forecast has been reduced by 7%, projections for the following two years have been increased by 15% and 20%, respectively, reflecting strong demand for AI-related power management semiconductors. The broker anticipates that while consumer products may face challenges due to rising memory prices, the demand for AI-related products will ease some of these pressures. Notably, HUA HONG SEMI is also experiencing positive momentum in its capacity and shipment growth, with ongoing expansion at Fab 9.

Looking ahead, the semiconductor market is likely to tighten in mature-node capacity by the second half of 2027. As major foundries prioritize advancements related to AI, it is crucial for industry stakeholders to adapt quickly to these changes to maintain a stable supply chain and take advantage of the emerging opportunities in the AI sector. The next few years will be vital for assessing how effectively the semiconductor industry can meet the increasing demands driven by the AI revolution.

GD

GPUBeat Desk

Desk · joined 2026

GPUBeat Desk covers AI infrastructure — chips, foundation models, inference economics, datacenter buildouts, and the geopolitics of compute.