以下文章来源于孤独大脑 ，作者老喻的 孤独大脑 . 一个专注于"元认知"的跨学科思想平台。由得到App《人生算法》课程主理人、中信出版社年度作者老 喻原创，100万深度思考者订阅的决策指南。融合科学、哲学、与概率论，系统性拆解AI、投资与复杂 决策问题，与你一起探索理性边界，实现丰盛人生。 来源 | 孤独大脑 作者 | 老喻 预计阅读时间：5分钟 现在各种成功学方法论里，也强调各种杠杆的作用。 那么，杠杆这把"放大器"，是捷径还是陷阱？ 01 杠杆的"魔咒" 为何不是点金石？ 杠杆本身不创造价值，更非"低风险高收益"的套利手段。 1.数学迷思：每日复利的"波动损耗" 杠杆ETF通过每日重设仓位机制，确保当日回报维持预设杠杆倍数。这导致了"波动损耗"（Volatility Drag）或"复利衰减" 。 若标的先涨10%再跌10%，初始100元变为99元； 而3倍ETF则变为91元 。 简单计算如下： （1+30%）✖（1-30%）=0.91 此现象源于几何平均回报低于算术平均回报，且杠杆会不成比例地放大此效应。 也就是说，即使市场随机波动，盈亏之间是不对称的。 而加上杠杆，放大了这种不对称性——进一步扩大了对几 ...

Group 1 - The article discusses the potential risks and pitfalls of using leveraged ETFs, particularly those that aim to provide three times the daily return of the semiconductor index [1][2][3] - Leveraged ETFs do not create value and are not a low-risk, high-reward arbitrage tool, as they amplify both gains and losses [2][30] - The concept of "volatility drag" is introduced, explaining how daily resetting of positions can lead to significant losses over time, especially in volatile markets [3][6][24] Group 2 - A mathematical example illustrates that a 10% gain followed by a 10% loss results in a net loss for both the underlying asset and the leveraged ETF, highlighting the asymmetrical nature of returns [4][5] - The article emphasizes that leveraged ETFs have a "path dependency," meaning their final returns are highly influenced by the specific daily price movements [7][19] - The extreme risk of leveraged ETFs is underscored by the fact that a single-day drop of over 33.33% in the underlying index could theoretically render the ETF worthless [8][9] Group 3 - Historical performance data shows that while a specific three-times leveraged semiconductor ETF has had impressive annualized returns over various time frames, these returns are not guaranteed to continue [19][20][35] - The article warns that the high volatility and potential for significant drawdowns, as seen in 2022, make these products unsuitable for long-term holding [22][36] - It is suggested that such leveraged products are more appropriate for short-term trading or very small allocations due to their heightened risk compared to broad market indices [25][34] Group 4 - The article draws parallels between physics and finance, stating that relying solely on leverage for excess profits is akin to believing in a perpetual motion machine [13][15] - The high management fees associated with leveraged ETFs, such as a 0.75% fee for SOXL compared to 0.03%-0.2% for regular ETFs, further erode potential returns [17] - Regulatory warnings indicate that leveraged ETFs are not suitable for long-term investment strategies [18][34]

最近跟着某大V买入两倍做空恒生科技ETF 07552作为港股的对冲仓位。对于杠杆ETF的磨损成 本之前有所耳闻，但具体多大没有量化过。今天拿两倍做空恒科和做多恒科画了一张图，简直震 惊了。 上图是17年至今这两个ETF的数据，发现平均值（绿线）是先跌后涨的，趋势基本和做多恒指 03037是一致的，这难道说明1倍做多恒指跟踪得好点，1倍做空恒指跟踪得差一点？ 最后，本文的目的是，提醒一下买带杠杆ETF的同学：阿祖，醒醒吧，别买了，损耗太大了。 拉格纳罗斯 美股的三倍做多纳指，tqqq。很多大v都在推，用来对冲之类的。但我就信一条。纳指这么大牛 市。他是三倍杠杆，市值竟然只有30亿。看曲线，也没看出来纳指的三倍牛市，咱也不懂，咋也 不问。但咱知道，不懂不做。 这次惊人地发现，1倍做空恒指07300和1倍做多恒指03037的每日平均值居然是上升的，由11.4 涨到了12.9，涨幅为13%，好嘛，买点做空的ETF对冲，除了实现对冲，该做空产品未来期望居 然还是涨的。 如果完全没有磨损的话，那么两倍做空07552和两倍做多07226的平均值那条线（即上图绿 线），就应该是一条笔直的水平线，但是过去一年，这个平均值居然由 ...

Core Viewpoint - Nvidia's stock has seen a significant decline of 23% over the past year, currently priced at $94 per share, raising concerns about its future performance amidst changing market dynamics and competition [3][4]. Group 1: Recent Developments - Recent reports indicate that Microsoft has slowed its data center expansion, which could negatively impact Nvidia's growth prospects [7][8]. - Microsoft's CEO expressed skepticism about the immediate financial benefits of AI, suggesting a potential slowdown in capital expenditures across the tech industry [8][9]. - Nvidia's current market capitalization of $2.5 trillion reflects expectations of sustained high sales, but any slowdown in revenue growth or profit margins could lead to significant financial losses for the company [8][9]. Group 2: Capital Expenditure Trends - Analysis of major tech companies' capital expenditures from FY16 to FY24 shows a marked decline in spending, particularly in FY20 and FY23, with a potential reversal in FY24 due to AI developments [9]. - The total capital expenditures for major tech firms in FY24 are projected to be $217.267 billion, a significant increase from previous years, but this trend may not continue if stock prices and financial performance decline [9][10]. Group 3: Market Sentiment and Risks - The recent downturn in large tech stocks, including Nvidia, reflects broader market trends, with the Nasdaq index dropping over 20% from recent highs [9]. - The potential for Nvidia's stock to face long-term challenges is highlighted by comparisons to Intel's historical performance, where revenue recovery did not lead to a return to previous market valuations [13][15]. - The influx of retail investors into leveraged ETFs related to Nvidia may indicate market exuberance, which often precedes market corrections [22][23]. Group 4: Future Outlook - Despite the current challenges, there are expectations for AI applications to grow exponentially, which could sustain demand for Nvidia's chips and maintain its competitive edge [23]. - However, the recent decline in Tesla's stock serves as a cautionary tale, emphasizing the volatility and risks associated with high-growth tech stocks like Nvidia [23].