- Semiconductor stocks tumbled across Asia on Tuesday as signs of progress by Chinese chipmakers intensified concerns about competition in AI hardware.
Global semiconductor stocks came under renewed pressure on October 13 as investors reassessed the outlook for artificial intelligence hardware amid growing competition from China. South Korea’s Kospi fell 5.5%, while Samsung Electronics and SK Hynix suffered sharp losses after reports that Chinese memory-chip maker ChangXin Memory Technologies (CXMT) had advanced a new memory architecture, according to the Associated Press.
The selloff extended beyond South Korea. Japan’s Nikkei 225 declined 1.8%, Taiwan’s Taiex dropped around 2%, and shares of Taiwan Semiconductor Manufacturing Co. (NYSE: TSM) also weakened. In the US, chipmakers including Intel Corp. (NASDAQ: INTC) and Micron Technology Inc. (NASDAQ: MU) fell in premarket trading as the global retreat continued, according to The Wall Street Journal.
The losses reflect two concerns facing the semiconductor industry: Chinese companies are developing more competitive alternatives to established suppliers, and investors are becoming more selective about businesses whose valuations depend on sustained AI spending.
The sector has benefited from heavy investment in data centers, advanced processors and memory used to train and run AI models. However, as semiconductor earnings have outpaced broader market growth, expectations for continued profit expansion have risen, leaving stocks vulnerable to any sign that competition or weaker spending could erode future returns.
China’s CXMT and Huawei Intensify Competition
CXMT’s reported progress on memory technology has raised concerns for Samsung and SK Hynix, two of the world’s largest memory-chip producers. The Chinese company is reportedly working on a new memory architecture intended to improve storage density, with a server-memory product expected by the end of 2026, according to market reports cited by Bloomberg.
The development could increase competition in a market where high-bandwidth memory and other advanced products have become essential components of AI infrastructure. Yet technological progress does not automatically translate into commercially competitive production. Manufacturing yields, performance, reliability and the ability to supply large customers at scale will determine whether CXMT can materially challenge established producers.
The concern extends beyond memory chips. Nvidia Corp. (NASDAQ: NVDA) has also lost ground in China’s data-center AI accelerator market as US export controls restrict access to some of its products and domestic suppliers expand shipments.
Nvidia held 34% of China’s AI accelerator market in the first half of 2026, down from its previous position as Chinese competitors gained ground. Nvidia shipped 678,000 accelerators during the period, a 44% year-over-year decline.
Huawei shipped 572,000 accelerators, giving it a 28% share of the market. Other Chinese suppliers, including Alibaba’s T-Head, Cambricon and Baidu’s Kunlunxin, also recorded growth in shipments.
The IDC figures show that China’s AI hardware market is changing, although Nvidia remains a major supplier. The data also indicate a shift toward specialized processors designed for particular AI workloads, rather than relying exclusively on traditional graphics processing units.
That shift is important for investors because the AI chip market is not a single, uniform category. Different processors serve different tasks, and companies may compete through price, power efficiency, software compatibility and the ability to connect large numbers of chips into computing systems.
AI Valuations and Spending Come Under Greater Scrutiny
Chinese competition is only part of the pressure on semiconductor stocks. Investors are also questioning whether the extraordinary growth expectations built into AI-related share prices can be maintained as the industry moves into another earnings season.
Broadcom Inc. (NASDAQ: AVGO) Chief Executive Hock Tan added to market concerns after warning that some investors were approaching AI investments with excessive enthusiasm. His comments came as investors weighed the cost of building AI infrastructure and the time required for companies to generate returns from those investments.
The concern is that rising spending does not guarantee proportionate profits for every company in the supply chain. Chip designers and memory manufacturers may benefit from strong demand, but customers building data centers must also pay for electricity, networking equipment, cooling systems and other infrastructure. Financing costs can further affect the economics of these projects.
The recent pullback also follows a period of strong gains in AI-related shares. As the concentration of market gains among AI and semiconductor leaders has increased, a change in investor expectations can have an outsized effect on major indexes and technology-heavy markets.
Upcoming earnings reports could help investors determine whether demand remains strong enough to support current valuations. Results from Taiwan Semiconductor Manufacturing Co. are particularly important because the company manufactures advanced chips for major technology companies and provides insight into demand across the semiconductor supply chain. Its third-quarter earnings outlook is likely to draw attention to AI-related sales, profit margins and capacity investment.
For now, the selloff signals growing uncertainty rather than proof that the AI investment cycle is ending. Chinese companies still face challenges in matching the manufacturing scale and performance of leading global suppliers, while demand for AI computing continues to support parts of the industry.
The key question for investors is whether established chipmakers can sustain their profit growth as competition expands and customers scrutinize spending. Strong earnings may support share prices, but companies will increasingly need to demonstrate that demand, pricing power and investment returns can justify the market’s expectations.