U.S. semiconductor shares sold off sharply on July 27, led by a nearly 5% decline in Nvidia and a 5.8% drop in ASML, as a report on China’s domestic lithography efforts added pressure to a chip sector already weakened by heavy exposure to artificial-intelligence spending.
The Philadelphia Semiconductor Index fell 2.23% by the close after dropping more than 5% during intraday trading. The move stood out against a mixed broader market: the Dow Jones Industrial Average gained 0.51% to 52,210.08, the S&P 500 edged up 0.02% to 7,413.18, and the Nasdaq Composite slipped 0.18% to 24,932.08.
Nvidia closed 4.99% lower at $196.51, its largest one-day decline since June 5. The fall erased roughly $250 billion from the company’s market capitalization, leaving it valued at about $4.76 trillion. Apple regained the position of the world’s largest company by market value as its shares reached a fresh high.
ASML, whose lithography machines are central to global semiconductor production, fell as much as 8% before ending at $1,655.26. AMD dropped 5.17% to $494.95, while Applied Materials, KLA and Lam Research each lost more than 3%, extending the weakness from chip designers into the equipment companies that supply fabrication plants.
China lithography report hits equipment makers
Selling intensified after reports that a Shanghai state-backed entity had begun manufacturing immersion deep ultraviolet, or DUV, lithography systems designed for 28-nanometer chip production. The first systems were expected to reach Semiconductor Manufacturing International Corp., Hua Hong Semiconductor and ChangXin Memory Technologies in 2026, according to the report.
The proposed production schedule was relatively limited: about five machines in 2026 and 20 in 2027. Yet the development touched a sensitive part of the semiconductor market. Lithography equipment has been one of the largest technological bottlenecks for China’s domestic chip industry, particularly after export restrictions limited Chinese access to advanced foreign systems.
The reported machines would target mature-node production rather than the leading-edge chips used in the most advanced AI processors. A 28nm process can serve many automotive, industrial, consumer and communications applications, but it sits well behind the nodes used by Nvidia’s most powerful data-center products.
The report said most components were sourced in China, although several critical parts continued to rely on Japanese suppliers. It also said the systems lagged ASML equipment in throughput, overlay accuracy and long-term operating stability. Overlay accuracy refers to how precisely a lithography machine aligns one layer of a chip’s circuitry with another; small errors can reduce manufacturing yields.
That gap places the reported project closer to a gradual domestic substitution effort than an immediate challenge to ASML’s high-end business. ASML ships hundreds of DUV systems annually, compared with the reported five-machine Chinese target for 2026. The supplied figures also put an ASML DUV system’s price at roughly €200 million and ASML’s fiscal 2025 revenue at about €30 billion.
Mature-node progress remains relevant
China’s equipment progress at older process nodes has nonetheless become increasingly relevant for suppliers that depend on sales to Chinese foundries. Publicly available information cited in the original account said Shanghai Micro Electronics had delivered its SSA800 28nm immersion DUV tool to SMIC in volume, with yields above 90% and a localization rate exceeding 85%.
The same material said Naura had verified a dry DUV tool on SMIC’s 28nm production line, using self-aligned quadruple patterning. That manufacturing technique uses repeated patterning steps to create denser circuit features than a single lithography exposure could produce. The account said the setup had achieved an 85% yield and could support a patterning result described as “equivalent 5nm.”
Such claims should be read carefully because a patterning result does not automatically translate into the economics, power efficiency or production scale of a commercially mature 5nm process. Repeated patterning can add manufacturing steps, increase costs and create more opportunities for defects. It can still help foundries extend existing DUV equipment for selected products.
The reported distance between domestic and commercial extreme ultraviolet systems remains much larger. The material described a Chinese laboratory EUV source peaking at 5,080 watts, around one-fifth of ASML’s commercial standard. It also placed domestic mirror roughness at 0.12 to 0.2 nanometers, compared with 0.05 nanometers for mirrors produced by Zeiss for ASML systems.
EUV tools require extraordinary precision across optics, light sources, vacuum systems, masks and software. The report estimated that a complete machine requires more than 100,000 components and put domestic supply-chain completeness below 20%. Those constraints explain why China’s near-term progress is more likely to affect mature-chip equipment demand than the global market for leading-edge lithography.
AI concentration magnifies a sector retreat
The selloff also reflected growing sensitivity around the financial scale of the AI buildout. Nvidia’s credit default swap spread widened by 14 basis points in one day to a record level, according to the supplied account. A CDS spread is the annual cost of insuring against a borrower’s default; a higher spread usually signals that market participants see more credit risk.
The account linked the reassessment partly to two large financing-related commitments associated with Nvidia: supply-chain cooperation with SK Group valued at more than $500 billion, and a proposed compute-lease guarantee for OpenAI of up to $250 billion tied to a 10-gigawatt Ohio data-center project. Together, those figures exceeded $750 billion, although they represent arrangements and proposed commitments rather than a single direct cash outlay by Nvidia.
Chip shares have become unusually influential within U.S. equity benchmarks. The supplied data said semiconductors’ share of the S&P 500 had risen from roughly 8% several years earlier to more than 20%. It also said the S&P 500 gained 0.80% on the previous day when AI-related stocks were excluded, underlining how a small group of large technology names has shaped index-level performance.
The Philadelphia Semiconductor Index had already fallen more than 20% from its June high by mid-July, meeting the common definition of a technical bear market. Leveraged exchange-traded funds may have added to intraday volatility because they typically rebalance daily, requiring them to sell into declines to maintain their targeted exposure.
Memory stocks lead the deeper losses
Memory-related companies posted some of the sharpest declines. SanDisk fell 11.02%, while SK Hynix’s U.S.-listed shares dropped 7.47% to $143.02, slipping below their $149 offering price on their 12th trading day. Micron ended 2.25% lower after trading down more than 7% earlier in the session.
The uneven market performance showed that the move was concentrated rather than a broad liquidation of risk assets. China-linked equities rose, with a U.S.-listed index of Chinese stocks gaining more than 2.5%. Xiaomi’s American depositary receipts climbed 8.97%, while Tencent and Alibaba each rose more than 2%.
That rotation left semiconductor valuations exposed to two overlapping concerns: whether AI infrastructure spending can sustain its recent pace, and whether China’s push to localize chip equipment will gradually narrow a profitable market for foreign suppliers. The July 27 session did not resolve either question, but it showed how quickly both can affect a sector that has become one of the market’s largest and most crowded trades.
Worried about chip volatility and AI-driven swings? Learn how to read cross‑market signals in this technical analysis guide.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
