SK Hynix has set up a dedicated team to search for growth opportunities beyond its core memory business and has put a senior group executive in charge, a move aimed at turning the company's lead in high-bandwidth memory (HBM) into a wider role in AI infrastructure.

Australia has tightened scrutiny of Chinese-linked holdings in Northern Minerals, requiring government approval before certain shareholders can sell their stakes and stipulating that the buyers cannot be related entities. The move, reported by Bloomberg and Reuters, extends Canberra's effort to limit Chinese influence in a rare earths supply chain seen as strategically important.

Samsung Electro-Mechanics has again delayed its semiconductor glass substrate commercialization schedule after a prototype reportedly failed a customer's reliability review. The setback has pushed back the construction timetable for GlaSSEM, its joint venture with Dongwoo Fine-Chem for glass core production in Pyeongtaek, South Korea, and has made equipment ordering dates difficult to pin down.

Quanta Computer reported record second-quarter results on August 13, driven by strong demand for AI servers and notebooks. The Taiwanese electronics manufacturer said consolidated revenue, gross profit, operating profit, and earnings per share all reached quarterly highs in the second quarter of 2026.

Chinese AI startup DeepSeek has officially launched its V4 Pro model with significantly enhanced agent capabilities, while sharply increasing API prices as it seeks to expand beyond its reputation for low-cost AI.

Semiconductor Manufacturing International Corporation (SMIC) is considering disclosing revenue from artificial intelligence (AI)-related "peripheral chips" as a standalone line item, possibly from the third quarter of 2026 or later, executives said on the company's second-quarter earnings call.

Alibaba Cloud has begun commercial service for its Zhenwu M890 supernode in Ulanqab, Inner Mongolia, giving global customers an early look at how AI infrastructure is shifting toward larger, faster, and more tightly connected systems. The launch could influence future cloud costs, model access, and enterprise AI deployment well beyond China.

Chinese printed circuit board (PCB) makers are stepping up investment in higher-value markets including AI servers, high-performance computing (HPC) and optical communications. Following similar expansion in semiconductors and display panels, Chinese companies are rapidly adding PCB capacity as they seek to catch up with South Korean suppliers that have established an early presence in the high-end substrate market.

China's largest contract chipmaker has spent most of the past two years increasing volume without much to show for it in terms of margins. The June quarter broke that pattern. Semiconductor Manufacturing International Corporation (SMIC) reported revenue of US$3.01 billion for the three months ended June 30, 2026, up 20.0% sequentially and 36.1% from a year earlier, and gross margin of 25.3% against 20.1% in the March quarter. The margin, not the revenue line, is the news: it is the first quarter in this cycle in which SMIC has been able to convert a full order book into price.

The AI wave is sweeping through the chip industry, but hallucinations, loss of control, and cybersecurity risks remain unresolved even as efficiency leaps. Youngsik Kim, head of the Design Technology (DT) team at Samsung Electronics' System LSI division, has pointed out that software has already leveraged AI to cut headcount and generate significant gains, but hardware has yet to see clear benefits.

SK hynix has pushed construction of its semiconductor campus in Yongin, Gyeonggi Province, South Korea, to 87% completion as it targets February 2027 for the start of operations at its first cleanroom. Power and water supplies have already been secured, according to Maeil Business Newspaper.

Rapidus expects interposers for high-performance chips to reach eight-reticle scale by around 2030, as larger chiplets and high-bandwidth memory (HBM) push advanced packages toward sizes that are increasingly inefficient to manufacture on conventional 300mm wafers.

As the world enters an AI-powered era where the persistent need for more complex architectures is driven by rising demands for higher computing performance, one bottleneck confronting the semiconductor industry lies not only in technological advancement itself, but also in the capacity to deploy standardized manufacturing processes at greater scale and speed.