AI Slowdown Call Triggers Sharp Derating in Tech
Leaders of the biggest artificial intelligence companies warned of potentially existential risks and called for a slowdown in AI development, triggering a sharp selloff in AI-linked equities across Asia and the United States.
The calls came over the weekend and into Monday, Sept. 14, 2026. AI-linked Asian stocks slumped after the statements from top lab CEOs, according to Reuters. Top AI industry CEOs called for a slowdown in AI development, according to CNN.
In Japan, shares in SoftBank closed nearly 11% lower. SoftBank is invested in OpenAI, the creator of ChatGPT. In South Korea, the Kospi sank 3.3%.
The selling carried into U.S. hours. Technology stocks wilted on Monday as investors focused on the ripple effects of any new AI controls, according to The Washington Post. The price action was broad across megacap tech and AI supply chain beta. The selling was concentrated.
This is not the first violent derating in the AI trade. Magnificent Seven stocks saw their worst drop since the launch of ChatGPT in July 2024, with $1.7 trillion in value erased in two weeks, according to MarketWatch. MarketWatch later framed the sentiment cycle with a Nov. 29, 2025 piece titled "As ChatGPT turns 3, here's what's crashing the party," according to MarketWatch.
Factor performance this year provides additional context. Global technology stocks started 2026 with one of the worst periods of underperformance relative to non-tech sectors since the early 1970s, according to Goldman Sachs. That rotation left valuation dispersion unusually wide inside large-cap growth.
By late August, Micron Technology (NASDAQ:MU), PDD Holdings (NASDAQ:PDD) and Adobe (NASDAQ:ADBE) were identified as the three cheapest stocks on the Nasdaq-100, according to Yahoo Finance. The mix is instructive. Memory, Chinese e-commerce and application software carry very different duration, margin and regulatory exposures.
The broader context here is a market repricing two linked risks at once. The first is earnings duration. Hyperscaler capex and model training spend pulled forward a large share of expected AI cash flows, leaving multiples sensitive to any extension of payback periods or throttling of deployment. A voluntary or mandated slowdown compresses the near-term revenue opportunity for accelerators, memory and interconnect while leaving the capex base elevated. That combination produces rapid multiple compression, particularly in high-beta Asia proxies.
The second is regulatory overhang. Controls on frontier development would not fall evenly. Foundries, equipment, and balance-sheet funders with concentrated OpenAI and frontier-lab exposure face a different earnings path than cash-generative software with existing seat-based monetization. The August cheap screen reflects that split. Micron prices cyclical memory and utilization risk. Adobe prices seat expansion and pricing power risk. PDD prices a China consumption and delisting discount largely orthogonal to U.S. model policy. For portfolio construction, the implication is less about tech versus non-tech and more about which duration and policy beta investors hold inside tech.
