Is the Fastest and Deepest Tech Stock Sell-off in History Nearing Its End?

marsbit发布于2026-07-19更新于2026-07-19

文章摘要

The article discusses whether the historically rapid and severe sell-off in tech momentum stocks is nearing its end. Citing analysis from Goldman Sachs and Morgan Stanley, it notes that the TMT momentum factor has plunged 40% from its peak in just 17 trading days, marking the fastest and deepest such decline on record. Key experts, including Goldman's Mark Wilson, attribute the crash primarily to extreme crowding, high leverage, and structural factors rather than a fundamental deterioration in the economy or corporate earnings. They point to strong underlying data, such as robust bank lending and raised guidance from companies like TSMC, which contrast sharply with the market's price action. While Wilson suggests the momentum unwinding process is likely "approaching its end," he cautions that immediate catalysts for a sharp reversal are lacking. The report highlights extreme internal market stress, with single-stock volatility far exceeding index volatility, and warns that despite the sell-off, positioning in momentum factors remains elevated. It concludes that the core market矛盾 (contradiction) lies in this disconnect between solid fundamentals and violent price corrections driven by positioning and leverage.

Source: Wall Street Insights

Tech momentum trading is experiencing its most severe unwinding in history. In just 17 trading days, the U.S. stock technology momentum factor (TMT MoMo) has plunged 40% from its peak, setting a record for the fastest and deepest drawdown on record, with contagion spreading comprehensively from semiconductors to hedge funds and credit markets.

Mark Wilson, partner and head of EMEA hedge fund business at Goldman Sachs, provided a systematic review of this "brutal rotation" this week. He noted that this sell-off is historically rare in both speed and depth, but its root cause stems more from non-fundamental factors such as crowded positioning and concentrated leverage, rather than a substantial deterioration in the economy or corporate earnings. He stated that the unwinding process for the momentum factor is "nearing its end," but lacks an immediate catalyst for reversal in the short term.

It is noteworthy that this momentum breakdown occurred against a backdrop of overall robust macro and corporate fundamentals—U.S. banks reported a 17% year-on-year increase in corporate lending, TSMC raised its 2026 revenue growth guidance to over 40%, and inflation data also came in moderately below expectations. This divergence between fundamentals and market price action is the core contradiction in the current market.

Tech Momentum Factor Suffers Historic Sell-off, Drawdown Speed and Depth Exceed Historical Median

According to data from Morgan Stanley's Quantitative and Derivatives Strategy team (MS QDS), this momentum factor drawdown has lasted 17 trading days, with a peak-to-trough decline of 28%. In comparison, the median momentum factor drawdown since 1999 has been 22%, lasting an average of 33 trading days.

This means the current decline has exceeded the historical median in both speed and depth, marking the most severe episode since the 29% drawdown from December 2022 to February 2023.

The situation is even more extreme in the tech sector. The TMT momentum factor (TMT MoMo) has fallen 40% from its peak. According to MS QDS data, this represents the fastest and deepest sell-off ever for the tech momentum factor.

Looking across various subsectors, South Korea's Kospi index is down 27% from its peak, U.S. AI tech beneficiaries are down 25%, global memory chip stocks are down 36%, and European semiconductors are down 23%. Among these, memory chip stocks account for about two-thirds of the overall decline, while the broader AI beneficiaries are down about 24% from their highs.

Low Surface Volatility Masks High Internal Intensity, Market Risk Structure is Unraveling

Price declines are only the surface manifestation of this turmoil; changes in the market's internal risk structure are equally noteworthy.

According to Goldman Sachs' volatility trading desk data, the current volatility of Goldman Sachs' High Beta Momentum portfolio (GSPRHIMO) is approximately 10 times that of the S&P 500's volatility. In the historical backtest over the past 20 years, such a stark volatility ratio has only been comparable to the situation during the pandemic shock in November 2020.

Simultaneously, the gap between single-stock volatility and index volatility has widened to historical extremes. Goldman Sachs data shows that the three-month average implied correlation among S&P 500 constituents fell to a record low of 0.14 this week, causing the S&P 500 index volatility to remain low, while the average implied volatility for single stocks is as high as 40%, 2.8 times that of the index's implied volatility, also setting a historical record.

Positions Remain Crowded, Risks Not Yet Cleared

Despite the recent historic drawdown in the momentum factor, hedge funds' net exposure to it remains high from a long-term perspective. J.P. Morgan data indicates that the combination of current positioning levels and the magnitude of the drawdown continues to make the momentum factor one of the core risks most deserving of caution in the market.

Meanwhile, the Goldman Sachs High Beta Momentum factor has fallen 33% since its June high, with its year-to-date gain plummeting from 60% to just 12%, a development also noted by Mark Wilson.

He cited signs of deleveraging in the Korean market as evidence: reports indicate that approximately one in every 30 South Korean adults had their stock margin accounts forcibly liquidated this week, showing that the deleveraging process has unfolded to a considerable extent.

Fundamentals Remain Sound, Risks Lie in Positioning and Structure

The peculiar aspect of this momentum breakdown is that it occurred against a backdrop of generally positive corporate fundamentals and macroeconomic data.

Mark Wilson pointed out that U.S. bank earnings this week presented an "unmistakably positive reading" of economic conditions: corporate lending grew 17% year-on-year, a record high, covering all economic sectors; U.S. consumer spending tracker growth is in the mid-single digits, with credit card spending up 6%; investment banking-related business lines collectively grew over 40%; large banks' tangible common equity return reached 19%, a post-financial crisis high.

On the tech capital expenditure front, TSMC raised its 2026 revenue growth guidance to over 40% (based on a revenue base exceeding $150 billion), while ASML's earnings report sparked market expectations for a 15% to 30% upward revision in its EPS over the next one to three years.

However, both companies' share prices fell following their earnings announcements, displaying a classic "sell the news" pattern. In contrast, IBM's stock price suffered its largest single-day drop in over 20 years due to large contract delays and underperformance in its consulting business.

Mark Wilson emphasized that this sell-off is "difficult to pinpoint to a clear signal at the fundamental level," reflecting more the influence of structural factors such as positioning, leverage, crowding, and concentration.

Rotation Nearing End, But Reversal Catalyst Still Awaited

Mark Wilson stated that he is inclined to believe the unwinding process for the momentum factor is nearing its end, but he also noted that there is a lack of immediate summer catalysts to drive a market reversal in the short term.

He also suggested that as efficiency and commercial viability improve, new market leadership will gradually emerge, and market breadth will expand accordingly—the Dow Jones Transportation Average breaking to new highs again this week serves as one example.

However, he also warned that the second derivative of earnings growth (i.e., the slowdown in the growth rate) will become increasingly important as the market digests Q2 earnings and enters the summer, while current valuation metrics across the board indicate that tech sector valuations remain elevated.

Furthermore, abnormal breakdowns in correlations are occurring both between traditional asset classes and within them; for example, the three-month correlation between gold and crude oil has fallen to an extreme inverse level in its 35-year history, further increasing the difficulty of risk management and portfolio construction.

相关问答

QWhat is the historical context of the current tech momentum factor sell-off in terms of speed and depth?

AAccording to the article, the current tech momentum factor (TMT MoMo) sell-off is the fastest and deepest on record, having fallen 40% from its peak in just 17 trading days. Historically, momentum factor drawdowns have a median decline of 22% over an average of 33 trading days.

QAccording to Mark Wilson of Goldman Sachs, what is the primary cause of this momentum factor sell-off?

AMark Wilson of Goldman Sachs attributes the sell-off primarily to structural factors like crowded positioning, concentrated leverage, and high congestion, rather than a deterioration in economic or corporate earnings fundamentals.

QDespite the market sell-off, what positive signals are mentioned regarding the underlying economic and corporate fundamentals?

AThe article highlights robust fundamentals: U.S. bank corporate loan growth at 17%, TSMC raising its 2026 revenue growth guidance to over 40%, and inflation data coming in softer than expected. Bank earnings also showed strong consumer spending and high returns on tangible equity.

QWhat does the data say about the current volatility and risk structure within the market during this event?

AThe volatility of Goldman Sachs's high-beta momentum basket is about 10 times that of the S&P 500, a disparity only seen during the November 2020 pandemic shock. Additionally, the correlation among S&P 500 stocks is at a record low, causing single-stock implied volatility to be 2.8 times higher than index volatility.

QWhat is Mark Wilson's outlook on the momentum factor sell-off and the potential for a market reversal?

AMark Wilson believes the momentum factor unwinding process is 'nearing its end.' However, he notes a lack of immediate catalysts for a sharp reversal in the near term (summer), and warns that valuations remain elevated and that the second derivative of earnings growth (a slowdown in growth acceleration) will become more important.

你可能也喜欢

Claude杀入全美教室:一周干49小时的老师,终于等来免费AI助教

近日,Anthropic公司正式推出“Claude for Teachers”,面向全美通过身份验证的K-12教师免费提供其高级AI功能。该工具旨在深度融入教师工作流程,解决他们长期面临的工作负担过重问题——数据显示,美国教师平均每周工作达49小时。 Claude for Teachers接入了覆盖全美50州的学术标准系统,能够自动生成严格对齐本地课标的教案,并提供分层教学设计以满足不同水平学生的需求。同时,它还能帮助教师分析学生数据,快速定位学习难点。工具集成了多个教育平台,并支持定时自动执行重复性任务,如每日复盘学生课堂反馈并调整次日教案。 值得注意的是,该产品定位为教师的辅助工具,学生无法直接接触。Anthropic明确表示不向18岁以下用户开放账号,此举旨在规避围绕学生使用生成式AI产生的争议,如作弊和认知依赖等问题。公司强调其重点是减轻教师行政负担,使其能将更多时间投入个性化教学与师生互动。 为确保在教育场景安全落地,Anthropic做出了多项数据隐私承诺,包括数据不用于模型训练、符合《家庭教育权利和隐私法案》(FERPA)并与教师工会合作制定隐私标准。尽管如此,社区中仍存在对教育过度自动化及数据安全的担忧。此次发布也引发更深层思考:当AI全面辅助教学后,未来一代能否为AI无处不在的世界做好充分准备。

marsbit刚刚

Claude杀入全美教室:一周干49小时的老师,终于等来免费AI助教

marsbit刚刚

冷门的哲学,成了“治”AI的热门

在2026年世界人工智能大会(WAIC)上,一个引人注目的现象是哲学与人工智能议题的深度融合。哲学教授出现在AI论坛的巅峰对话席,相关蓝皮书也将深度思考与AI治理置于聚光灯下。这背后反映出AI能力飞速发展后所面临的本质性哲学追问:模型流畅回答是否代表真正理解?AI能否成为自主发现知识的“科学家”?对于人类无法解释却能验证的AI发现,我们该不该相信? 复旦大学教授孙宁在演讲中回归智能的源头,指出深度智能需要扎根于身体、环境、他者与历史构成的“世界”。论坛议题随即从“何为智能”延伸到AI具体应用带来的挑战。例如,自进化模型已能自动生成并通过同行评审的学术论文,这引发了关于AI科研能力评估、成果归属及责任划分的新问题。在数学与生命科学领域,AI能极大扩展探索空间,但最终仍需人类定义价值、机器严谨验证,并与真实世界反馈形成闭环。 当AI开始模拟社会时,问题变得更为复杂。社会模拟平台虽能高效推演政策长期影响,但其结果可能掺杂数据偏见,且虚拟角色的行为未必反映真实社会规律。专家提醒,切勿将模拟结果直接等同于社会声音。AI融入社会后,治理挑战超越技术精度,涉及攻击性使用、心理依赖、责任界定以及“一管就死、一放就乱”的监管困境。对此,《2026人文社会科学智能发展蓝皮书》提出了可操作的治理框架,强调研究过程需保留完整证据链,在公共决策中确保人类拥有实质性的介入权与解释权。 治理问题最终触及产业落地。在生物制造等领域,AI需跨越数字与物理世界的鸿沟,与自动化实验设备形成闭环。然而,技术加速可能遭遇监管、产业惯性和利益分配等新瓶颈。AI推动产业从过程交付转向结果交付,也催生了重新定义“生产关系”的需求。 综上所述,哲学在AI圈的“升温”,标志着AI发展已进入深水区,其问题与治理、伦理、公共政策紧密交织。在上海形成的青年科学智能生态中,跨学科学者正共同探索如何让AI既走得远,又能稳健地融入现实。当AI越来越善于回答,人类更需要深思:我们该问什么、相信什么,以及将技术引向何方。

marsbit41分钟前

冷门的哲学,成了“治”AI的热门

marsbit41分钟前

32亿募资画饼六年后九成补流:露笑科技碳化硅"产业化"的虚与实| 钛媒体深度

露笑科技于7月7日公告终止2021年度定增中的两大碳化硅募投项目,并将剩余12.17亿元募资永久补充流动资金,股价随后大幅下跌。自2020年高调跨界第三代半导体以来,公司通过两次定增合计募资约32亿元,但实际投入碳化硅项目建设的资金不足3亿元,近九成资金通过补流、还贷等方式“变通”使用。 公司规划的百亿碳化硅产业园一期项目进展缓慢,原计划年产24万片衬底片,但实际产能成谜,对外披露信息存在矛盾。2022年至2025年,相关主体合肥露笑累计亏损3.44亿元。在行业已从6英寸转向8英寸乃至12英寸衬底、竞争格局初定之际,露笑科技宣布将以自有资金转向8英寸和12英寸研发,被指“永远在赶晚集”。 此外,公司核心技术人才——首席科学家陈之战的现状在信披中语焉不详,其近年学术产出寥寥,且通过关联方与公司存在交易,引发外界对其技术竞争力的疑虑。 复盘公司历史,其自2011年上市后多次“追热点”转型,涉足机电、蓝宝石、新能源、光伏、登高机等领域,但胜率低。累计直接融资超70亿元,但合计盈利仅2.21亿元。与此同时,实控人家族自2018年以来累计套现约19.93亿元,并于近期再次抛出减持计划。露笑科技的案例,成为一段以融资和追热点为核心的资本运作样本。

marsbit1小时前

32亿募资画饼六年后九成补流:露笑科技碳化硅"产业化"的虚与实| 钛媒体深度

marsbit1小时前

交易

现货
活动图片