TSMC(US:TSM) 美股IPO·2025-11-12 10:19Core Viewpoint - Morgan Stanley's latest research indicates that despite ongoing concerns about an AI bubble, the semiconductor upcycle driven by AI is far from peaking, with its duration expected to extend beyond typical cycles until 2027, providing strong support for Asian tech stocks in 2026 [1] Semiconductor Revenue Growth - The firm has raised its forecasts, predicting global semiconductor revenue to grow by 18% and 11% in 2026 and 2027, respectively, driven by the early adoption curve of generative AI, strong capital expenditure from leading cloud service providers (CSPs), and conservative capacity expansion strategies in key supply chain segments [3] Asian Tech Market Dynamics - In 2026, the Asian tech market is expected to exhibit a "delicate" balance, where investor concerns about the cycle peaking coexist with continuous upward adjustments in earnings per share (EPS) driven by AI infrastructure development and price increases in certain components [4][9] Market Sentiment and Earnings Revisions - Morgan Stanley notes that while classic indicators suggest the current upcycle is nearing its late stage, this does not alter the upward trend. In the first half of 2026, Asian tech stocks are anticipated to have room for growth due to strong earnings from AI-leading companies [9] Supply Shortages and Price Increases - The strong demand for AI is "crowding out" supply across the tech industry, leading to shortages in various sectors, including advanced packaging, wafer foundries, and high bandwidth memory (HBM), with suppliers increasing capacity at a slower pace than normal, resulting in price hikes that will further boost corporate earnings revisions [11] Unique Characteristics of the Current Cycle - The current cycle is characterized by a "K-shaped recovery," with a divergence between AI and non-AI demand since mid-2023, where AI-related demand remains robust while other tech sectors face adjustments [12] - Generative AI adoption is following a steep S-curve, similar to the early stages of smartphones and public cloud, with expectations of 50%-60% year-on-year growth in 2026, indicating significant growth potential [13] - Leading CSPs are expected to maintain strong capital expenditure, with the top six CSPs projected to increase capital spending by 32% in 2026 after a 67% growth in 2025, demonstrating their financial capacity to support ongoing AI infrastructure expansion [16] Supply Chain Constraints - The current cycle has seen conservative growth in semiconductor capital expenditure, with supply concentrated among a few manufacturers. TSMC's capital expenditure is expected to grow by only 16% in 2026, while the DRAM sector's capital expenditure is projected to increase by 11%, indicating that supply shortages in critical areas will persist until 2026 [20] Investment Strategy Recommendation - Based on these insights, Morgan Stanley recommends a "barbell" investment strategy for 2026, allocating one end to leading AI enablers and the other to companies benefiting from price increases and margin expansion, with TSMC identified as a preferred stock [22]