The stock market's new most-hated word is pummeling the AI trade

Core Viewpoint - The recent focus on depreciation concerns regarding expensive GPUs and semiconductor chips is causing significant anxiety among investors, particularly in the AI sector, leading to declines in major tech indices [2][4]. Group 1: Market Impact - The Nasdaq 100 has decreased by 6.3% in recent weeks, while the Technology Select Sector SPDR Fund has fallen over 9% due to depreciation fears [2]. - Notable short sellers, including Michael Burry and Jim Chanos, have highlighted depreciation as a critical reason for their skepticism towards the AI trade [2][3]. Group 2: Depreciation Estimates - Michael Burry estimates that Big Tech hyperscalers will understate depreciation by $176 billion from 2026 to 2028, predicting a two to three-year lifecycle for chips instead of the anticipated six years [3]. - Peter Berezin from BCA Research projects that hyperscalers will hold at least $2.5 trillion in AI assets by the end of the decade, leading to an annual depreciation expense of $500 billion, which exceeds their combined profits for 2025 [4]. Group 3: Future Projections - Kai Wu from Sparkline Capital suggests that annual depreciation values could increase from $150 billion to $400 billion over the next five years, indicating a significant financial burden on these companies [4][5]. - Wu argues that the current AI spending relative to GDP surpasses the peak of the Internet boom, although it remains below the railroad buildout peak, emphasizing the shorter useful life of AI chips [6]. Group 4: Industry Sentiment - The depreciation argument is not yet widely accepted on Wall Street or within the AI industry, with few strategists warning of such risks [6]. - Bernstein analyst Stacy Rasgon maintains that the depreciation accounting of major hyperscalers is reasonable, suggesting a divergence in views regarding the depreciation issue [7].