Morgan Stanley: Asia Pacific 2026 Outlook - Prefer AI to Non-AI
Summaries are public. PDF access requires an active membership; all members have the same access. Sign in
Morgan Stanley released its Asia Pacific 2026 outlook, decisively favoring AI-related semiconductor plays over non-AI technology. The report identifies three primary long-term demand drivers: technology inflation, the AI crowding-out effect, and technological diffusion. Technology inflation encompasses rising wafer, back-end, and memory costs that threaten fabless margins. The crowding-out effect reflects how AI demand is monopolizing supply chain capacity, triggering shortages in components like T-Glass and memory. Technological diffusion suggests generative AI will expand into verticals like robotics and smart glasses. Within AI, Morgan Stanley holds an Overweight rating across leading foundries, back-end testers, and fabless names. Top AI convictions include TSMC, MediaTek, ASE Technology, and ASMPT. For memory, favored stocks benefiting from AI ripple effects include Winbond, Phison, and GigaDevice. The bank notes that China's DeepSeek is successfully driving inference AI demand, though domestic GPU supply limitations remain a key bottleneck. Nvidia H200 shipments could further dilute the domestic supply chain. Conversely, the non-AI sector faces demand softness and pricing pressures. Morgan Stanley maintains Equal-weight or Underweight ratings on several non-AI names, including United Microelectronics, Vanguard International, and Will Semiconductor. Meanwhile, it maintains Overweight ratings on specific non-AI beneficiaries like Realtek and US-listed Chinese semiconductor capital equipment leaders such as Naura and AMEC.