Alla Harmsworth, Co-Head of Institutional Solutions at AllianceBernstein, believes data can help identify genuine skill in active fund managers. Her team uses a proprietary approach called Alphalytics to isolate "idiosyncratic alpha," which measures skill beyond market factors like value, growth, and momentum. This approach helps avoid chasing short-term winners whose returns may be driven by style tailwinds. Harmsworth emphasizes the importance of understanding how fund managers have achieved their outperformance and whether it is likely to persist.
Taiwan Semiconductor Manufacturing Company (TSMC) is on the cusp of becoming the first semiconductor firm to breach a $2 trillion market cap, driven by its dominance in the AI chip market and undervalued growth potential. As of August 2025, TSMC commands a market capitalization of approximately $1.197 trillion, reflecting its pivotal role in the global AI infrastructure boom [1].
TSMC's Q2 2025 results underscore its financial resilience. The company reported a 61% year-over-year surge in net income to NT$398.27 billion and revenue of NT$933.80 billion, a 38.65% increase, driven by robust demand for AI and high-performance computing (HPC) chips [2]. The HPC segment alone accounted for 60% of TSMC’s revenue, up from 52% in the same period in 2024, as advanced-node manufacturing (below 7nm) became critical for AI accelerators and GPUs [1]. For Q3 2025, TSMC guided revenue between $31.8 billion and $33.0 billion, with net income exceeding $10 billion—well above Wall Street expectations [3]. This performance positions TSMC to achieve a 30% revenue growth in 2025, fueled by AI infrastructure spending expected to reach $375 billion in 2025 and $500 billion in 2026 [3].
TSMC's strategic positioning in the AI era is bolstered by its leadership in advanced manufacturing and packaging technologies. The company holds a 100% market share in AI data center logic semiconductors, producing chips for NVIDIA, AMD, Intel, and custom accelerators for cloud giants like Microsoft, Amazon, and OpenAI [3]. Its collaboration with NVIDIA on the Blackwell AI chip, built using TSMC’s 3nm and 5nm nodes, highlights its strategic alignment with the AI era [3]. Additionally, TSMC’s $165 billion U.S. investment plan, including three new Arizona fabrication plants, ensures tariff exemptions and expands its capacity to meet surging demand [3].
Advanced packaging technologies like CoWoS further differentiate TSMC. These innovations enable the integration of multiple chiplets and high-bandwidth memory (HBM), essential for large language models (LLMs) and next-generation AI workloads [3]. With the global semiconductor market projected to grow to $1 trillion by 2030 [3], TSMC’s 35% share of the “Foundry 2.0” market and its mid-30% annual revenue growth in this segment position it to capture disproportionate value [5].
Despite its dominance, TSMC’s valuation remains attractive. A forward P/E ratio of 23.14 and a PEG ratio of 1.08 suggest the stock is fairly priced relative to its projected earnings growth [2]. In contrast, peers like AMD and Broadcom trade at P/E ratios of 96.7x and 105.6x, respectively [3]. Analysts argue that TSMC’s P/E is undervalued given its 45% CAGR in AI-related revenue and 20% CAGR in overall revenue through 2029 [1]. Morningstar estimates AI chip revenue could account for 50% of TSMC’s total revenue by 2029, driven by mid-40s growth rates in the segment [2].
The broader semiconductor industry’s PEG ratio of 0.55 in 2025 further underscores undervaluation relative to growth expectations [3]. With TSMC’s revenue expected to grow 38% in 2025 and 22% annually over five years [2], its valuation appears to lag its fundamentals. A 12% total return annually—combining 11% CAGR in market cap growth and a 1% dividend yield—could see TSMC reach $2 trillion by 2030 [1].
References:
[1] TSMC (TSM) - Market capitalization [https://companiesmarketcap.com/tsmc/marketcap/]
[2] TSMC's Recent Underperformance: A Strategic Buying Opportunity [https://www.ainvest.com/news/tsmc-underperformance-strategic-buying-opportunity-earnings-optimism-valuation-adjustments-2508/]
[3] TSMC: King Of Data Center AI [https://semiengineering.com/tsmc-king-of-data-center-ai/]
Comments
No comments yet