抛售在 2026 年 1 月 28 日 Q4 财报季开始后加速，Microsoft 和 Amazon 各自下跌超过 16%，Alphabet 较近期高点低约 11%，而 Meta 在财报驱动的跳升后下跌 13%。合计来看，Microsoft、Amazon、Meta 和 Alphabet 的市值蒸发近 $1.5 trillion，并推动 Nasdaq 100 在 2026 年转入负值区间。核心压力点是支出：预计这 4 家公司在 2026 年将投入超过 $600 billion 的资本开支；同时，AI 颠覆交易也打击了私募信贷公司、软体与游戏股、保险商、财富管理公司与物流企业，其中有一例由一家季度营收不足 $2 million 的公司触发。
数据显示出一种厄运循环动态：投资者惩罚潜在的 AI 失败者，同时也惩罚最大的 AI 支出者，尽管这两种看法难以调和。UBS 将科技板块下调至中性，并估计超大规模云服务商的资本开支可能吸收几乎 100% 的营运现金流，而 10 年平均为 40%，这意味著更依赖债务或股权融资。反方观点仍具分量：先前的 AI 恐慌（包括 2025 年的 DeepSeek 冲击）之后曾反转；AI 采用仍不均衡；而 Nvidia 和 Micron 等受益者在 3 年期股价上涨强劲，因此近期基准情境仍是高波动，而非已定型的长期盈利结果。
By February 15, 2026, US equity sentiment around AI had split into 2 conflicting fears: AI will rapidly destroy incumbents’ business models, and AI leaders are overspending with uncertain near-term returns. In the prior upswing, Meta rose nearly 450% from end-2022 to early-2026 and Alphabet gained more than 250%, but over roughly the last 2 weeks this narrative reversed as broad AI-related volatility spread from mega-cap tech to multiple non-tech sectors.
The selloff accelerated after Q4 earnings season began on January 28, 2026, with Microsoft and Amazon each down more than 16%, Alphabet about 11% below a recent peak, and Meta down 13% after an earnings-driven jump. Combined, nearly $1.5 trillion of market value was erased from Microsoft, Amazon, Meta, and Alphabet, helping push the Nasdaq 100 into negative territory for 2026. A central pressure point is spending: those 4 firms are expected to deploy more than $600 billion in 2026 capex, while AI-disruption trades simultaneously hit private-credit names, software and gaming stocks, insurers, wealth managers, and logistics firms, in one case triggered by a company with under $2 million in quarterly revenue.
The data suggest a doom-loop dynamic: investors punish potential AI losers while also punishing the biggest AI spenders, despite those views being hard to reconcile. UBS cut tech to neutral and estimated hyperscaler capex could absorb almost 100% of operating cash flow versus a 10-year average of 40%, implying heavier reliance on debt or equity financing. Counterpoints remain material: prior AI panics, including the DeepSeek shock in 2025, later reversed; AI adoption is still uneven; and beneficiaries such as Nvidia and Micron have seen strong 3-year share gains, leaving the near-term base case as elevated volatility rather than a settled long-run earnings outcome.