Semiconductor News 20260707

Today’s semiconductor news is driven by advancements in manufacturing process technology and large-scale investments, against the backdrop of explosive growth in AI demand. Intel, in particular, is accelerating technological innovation to improve chip density—such as considering the adoption of a dual-side power supply system for its 14A2 node—as it prepares to compete with TSMC and Samsung in the 1.4nm generation.Meanwhile, in the AI memory market, companies are locked in a fierce battle for dominance: SK hynix has announced a massive investment of 1,100 trillion won in South Korea, while Micron is investing $9.3 billion in its Hiroshima plant to expand HBM production capacity.Furthermore, fabless companies are increasingly adopting cutting-edge processes, as evidenced by Socionext’s initiation of AI chiplet development using TSMC’s A14 process.These developments demonstrate that building AI infrastructure—from data centers to edge AI—has become the semiconductor industry’s top priority, and that the miniaturization of manufacturing technology and the increase in memory capacity are set to reshape the foundation of the future AI society.

Main News

1. SK Hynix to invest a total of 1,100 trillion won in South Korea to expand its AI memory production base.
   SK Hynix has announced a medium- to long-term investment plan totaling 1,100 trillion won targeting Yongin, Cheongju, and the southwestern region to meet the rapidly expanding demand for AI memory. The plan will significantly increase production capacity for HBMs, server DRAM, and enterprise SSDs, while simultaneously developing AI data center infrastructure. By capturing the structural changes in memory demand accompanying the practical application of generative AI and strengthening its production bases in South Korea, the company aims to secure leadership in the global AI memory market.
2. Intel is considering introducing a bifacial power supply (BIF) method in its 1.4nm generation 14A2 node.
   Intel is considering adopting a bifacial power supply architecture for its next-generation 1.4nm process, the 14A2 node, to compete with TSMC and Samsung in the miniaturization race. Initially, only a back-side power supply network (BSPDN) was envisioned, but the need to improve transistor density and suppress pattern errors in the lithography process has forced a reconsideration of the power supply method. This technological innovation is planned to enhance Intel’s competitiveness against the most advanced nodes of its competitors.
3. Micron is investing $9.3 billion in its HBM plant in Hiroshima, Japan, aiming for mass production in 2028.
   Micron has decided to make a massive $9.3 billion investment in its Hiroshima, Japan plant, aiming to expand its market share in the AI ​​memory market. This investment will significantly increase its HBM production capacity, with the goal of starting mass production shipments in 2028. Amid rapidly increasing demand for AI servers, this strategy positions Japan as a key production base and strengthens its global supply chain.
4. Socionext develops high-performance AI chiplet using TSMC’s A14 process
   Fabless semiconductor manufacturer Socionext has begun developing a high-performance compute chiplet using TSMC’s state-of-the-art A14 process as part of its custom SoC development for AI data centers. In September 2026, tape-out of multicore devices is planned, utilizing this chiplet as a platform to verify the scalability of CPUs and AI processors, accelerating development for next-generation AI infrastructure.
5. Kim Jong-ho, the “father of HBM,” talks about the importance of memory in the AI ​​era.
   Kim Jong-ho, who contributed to the development of HBM, pointed out that the essence of AI lies not only in computing power but also in memory. Storage chips are reshaping the AI ​​architecture itself, and memory bandwidth and capacity are decisive factors that determine the performance of AI systems. He emphasized that as AI models evolve, memory technology has gone beyond a mere auxiliary role and is now driving technological innovation as the core of AI infrastructure.
6. Shenzhen-based semiconductor substrate leader aims to break through 2nm technology and list in Hong Kong.
   It has been reported that a major semiconductor package substrate company in Shenzhen has achieved a breakthrough in 2nm process technology. The rapid increase in AI-driven computing demand is strongly driving the company’s technological development, and leveraging this technological strength, it aims to list on the Hong Kong Stock Exchange. Advancements in cutting-edge packaging technologies are essential for improving AI chip performance and are drawing attention as a symbolic move to strengthen China’s domestic semiconductor ecosystem.
7. Samsung, SK hynix, and Micron Face Intensifying Competition in the HBM4 Market
   As demand for custom chips for AI grows, competition among the three major memory manufacturers over the HBM4 generation is intensifying.Each company is strengthening its ability to provide customized solutions tailored to customer needs, shifting from a simple race for capacity to memory designs optimized for specific AI accelerators. As they vie for market leadership, the miniaturization of manufacturing processes and the advancement of packaging technologies are becoming critical factors that will determine the outcome.
8. TAI has completed evaluation of its edge AI test chip “Sting Ray” and is moving to mass production.
   Tokyo Artisan Intelligence (TAI) has completed evaluation of its reconfigurable AI semiconductor chip for edge AI, “Sting Ray,” which employs a proprietary architecture. Manufactured using UMC’s 40nm process with the cooperation of Oppstar, this chip will be succeeded by the next-generation mass-production model, “Manta Ray.” Going forward, TAI aims to commercialize the chip in physical AI fields such as railways, infrastructure, and robotics.

News List (199 items)

This article is delivered every morning via an automated collection and AI-generated system. Number of items collected: 199 | Last updated: July 7, 2026

Editor’s Note: This article utilizes AI to summarize and organize news content. While every effort has been made to be as accurate as possible, it may contain errors in background explanation or interpretation of causal relationships. Please always check the source article for details and accurate context.