格隆汇11月11日｜电影《大空头》(The Big Short)男主角原型、传奇对冲基金经理Michael Burry在披露 大手沽空科技巨头后，再在社交平台发文称，科技巨头透过"人为"延长资产"有效使用寿命"来低估折 旧，从而抬高利润，这是"现代最常见的诈欺(common frauds)之一"。 Burry指出，科技企业大规模增加资本开支，采购英伟达晶片和服务器以扩充算力，这些产品周期通常 只有2至3年，但部分公司将折旧周期延长至6年，折旧年限不应被延长。不过，正是超大规模科技企业 如Meta及甲骨文都在这样做，据他估计，企业在2026年至2028年间将低估折旧1760亿美元。Burry估 计，到2028年，甲骨文盈利或被夸大26.9%、Meta则夸大20.8%，情况可能更差。然而，布瑞未有说明 详细原因，只表示更多详情将于11月25日发布。 此前，美国银行分析师Justin Post于9月曾发表报告指，谷歌母公司Alphabet、Meta和亚马逊资本支出在 2024年和2025年大幅增长，其折旧费用必然会在2026年后加速。推算到2027年，仅这3家公司折旧费 用，市场普遍预测就可能比实际情况低了近164 ...

Core Insights - Michael Burry, a legendary hedge fund manager and the inspiration for the protagonist in "The Big Short," has raised concerns about major tech companies manipulating asset depreciation to inflate profits, labeling it as a common fraud [1] - Burry estimates that companies like Meta and Oracle are extending their depreciation periods from the typical 2-3 years to 6 years, leading to an underestimation of depreciation by approximately $176 billion between 2026 and 2028 [1] - He predicts that by 2028, Oracle's earnings could be overstated by 26.9% and Meta's by 20.8%, indicating a significant potential misrepresentation of financial health [1] Company Analysis - Major tech firms, including Meta and Oracle, are reportedly increasing capital expenditures significantly, particularly in acquiring NVIDIA chips and servers to enhance computing power [1] - A previous report by Bank of America analyst Justin Post indicated that Alphabet, Meta, and Amazon are expected to see substantial growth in capital expenditures in 2024 and 2025, which will lead to accelerated depreciation expenses post-2026 [2] - The market consensus suggests that by 2027, the depreciation expenses for these three companies could be underestimated by nearly $16.4 billion, implying that their actual profitability may be much lower than currently perceived [2]

Core Viewpoint - Malaysia is not restricting the use of Nvidia chips or AI chips in servers under the Strategic Trade Act of 2010, but will cooperate with other countries on export control of sensitive goods [1][2] Group 1: Malaysia's Position on AI Chips - Malaysia's Investment, Trade and Industry Ministry states that the use of Nvidia chips in data centers is not classified as controlled goods under local law [1] - The country allows data centers to make independent business decisions while being supervised by relevant ministries [1] - Malaysia opposes any evasion of export controls or illegal trade activities and emphasizes compliance with international trade regulations [1] Group 2: U.S. Export Controls - Since 2022, the U.S. has intensified restrictions on AI chip exports to China, including a ban on Nvidia and AMD from selling advanced chips to the Chinese market [2] - The Trump administration initially imposed a ban on Nvidia's AI chip sales to China, which was later relaxed [2] - China's Ministry of Foreign Affairs has expressed that Chinese companies should operate within the law and opposes coercive restrictions on cooperation with China [2]