What's happening

Chip stocks experienced a broad global selloff on Friday, June 26, 2026, driven by growing investor concern over the escalating cost of artificial intelligence infrastructure. The declines were widespread across geographies and market segments, spanning U.S. fabless designers, Asian foundries and memory makers, and European semiconductor equipment suppliers. Among U.S.-listed names, Sandisk led losses at 10%, followed by Marvell at 5%, Micron at more than 5%, Arm at nearly 4%, and Intel at 3%. Nvidia, whose H100 and A100 data center GPUs sit at the center of AI infrastructure buildouts, and AMD, whose Instinct accelerators compete in the same market, were also caught in the broader sector pressure.

The selloff was not confined to U.S. markets. In Asia, South Korea's SK Hynix fell more than 8% and Samsung Electronics lost around 5%, while Japan's Advantest declined nearly 10% and Tokyo Electron was down 3%. Taiwan's TSMC, the primary contract manufacturer for chips designed by Nvidia, AMD, and Apple, sank slightly. In Europe, ASML declined 2%, Infineon fell 4%, ASM International dropped 4%, and STMicroelectronics shed nearly 4%, indicating that the pressure extended to semiconductor capital equipment and analog chip suppliers as well.

Why it matters for markets

The breadth and geographic scope of the June 26 selloff underscores how deeply AI infrastructure spending assumptions have become embedded in semiconductor valuations across the supply chain. AMD currently trades at a price-to-earnings ratio of 172.1, reflecting substantial growth expectations tied to AI accelerator demand, while TSMC carries a P/E of 37.5 and Nvidia trades at 29.5 — all elevated relative to broader market averages. A sustained reassessment of AI infrastructure cost trajectories could pressure these multiples, particularly for companies where near-term earnings may not yet justify current valuations.

The scale of capital at stake amplifies the significance of any shift in sentiment. Nvidia carries a market capitalization of approximately $4.66 trillion, TSMC $2.24 trillion, and AMD $850.49 billion — collectively representing trillions of dollars in equity value that are closely tied to continued AI infrastructure investment. TSMC's reported revenue of $4.10 trillion and Nvidia's $253.49 billion in revenue illustrate how central these companies have become to the global technology supply chain. If hyperscalers or enterprise customers slow AI infrastructure deployment in response to cost pressures, the revenue impact would ripple across foundry, chip design, memory, and equipment segments simultaneously.

The simultaneous declines in memory makers such as SK Hynix and Samsung, equipment suppliers such as ASML and Tokyo Electron, and chip designers across multiple continents suggest the market is reassessing the entire AI infrastructure investment cycle rather than reacting to company-specific news. This kind of synchronized cross-sector, cross-geography movement is characteristic of a macro-level repricing of a thematic trade rather than isolated stock-specific events.

Sectors and assets to watch

The primary tickers directly implicated in the selloff are Nvidia (NVDA), Taiwan Semiconductor Manufacturing Company (TSM), and Advanced Micro Devices (AMD), all of which are central to AI chip design and production. Nvidia's data center GPU franchise and AMD's Instinct accelerator line represent the leading edge of AI compute hardware, making both companies highly sensitive to any moderation in AI infrastructure spending. TSMC, as the contract manufacturer producing chips on advanced nodes for both Nvidia and AMD, faces correlated demand risk if orders from major fabless clients soften. TSMC's 52-week range of $221.18 to $476.79 and Nvidia's range of $151.49 to $236.54 illustrate the degree of price volatility these names have already experienced over the past year.

Beyond the primary tickers, the June 26 session highlighted vulnerability across the broader semiconductor ecosystem. Memory suppliers SK Hynix and Samsung Electronics, which provide high-bandwidth memory critical to AI accelerators, fell sharply. Semiconductor capital equipment companies — including ASML, Tokyo Electron, ASM International, and Advantest — also declined, reflecting the interconnected nature of the chip supply chain. Any sustained reduction in AI infrastructure investment would affect not only chip designers and foundries but also the equipment makers that supply the tools required to manufacture advanced semiconductors.

What to watch next

Market participants will be monitoring whether the June 26 declines represent a single-session correction or the beginning of a more sustained reassessment of AI infrastructure spending expectations. Key indicators to track include forward guidance from hyperscale cloud providers on capital expenditure plans, any commentary from Nvidia, AMD, or TSMC regarding order trends or capacity utilization, and macroeconomic data that could influence enterprise technology budgets. The behavior of high-multiple names such as AMD, with a P/E of 172.1, will be particularly telling, as any downward revision to growth expectations tends to have an amplified effect on richly valued stocks. Developments in semiconductor equipment order books — particularly at ASML and Tokyo Electron — will also provide early signals about whether foundry customers are adjusting their capacity expansion timelines.