Semiconductor News 20250929

Main News

1. AMD’s Instinct MI450X Impact on NVIDIA Rubin Chip Design AMD’s Instinct MI450X has reportedly prompted NVIDIA to redesign its upcoming AI chip, Rubin, as a competitor. This has forced NVIDIA to make adjustments to Rubin, including increasing TGP (Total Graphics Power) and memory bandwidth, suggesting increased competition between the two companies in the AI accelerator market.
2. TSMC ‘s 2nm Manufacturing, Expansion Strategy from Hsinchu to Arizona TSMC is facing a “gamble” over the deployment of its 2nm manufacturing technology, expanding its production focus from Hsinchu, Taiwan to Arizona, USA. This strategy is an important and closely watched move by the company to diversify its next-generation semiconductor production globally and diversify its risk exposure amidst the interplay of geopolitical factors and technological challenges.
3. D2D Interconnect Change in AMD Zen 6 CPU s AMD plans to convert its die-to-die (D2D) interconnect technology from traditional SERDES to a “Sea-of-Wires” approach in its next-generation Zen 6 CPUs. This change is intended to provide significant power savings and lower latency, and is an important architectural advancement in next-generation CPU design that is expected to contribute to improved performance.
4. Intel Panther Lake CPU and First Challenge with 18A Node Intel’s upcoming CPU, Panther Lake, is positioned as Intel’s first “challenge” as it will be the first product to use the company’s highly anticipated state-of-the-art 18A manufacturing node. The industry is watching closely as the yield and performance of the 18A node will be key to Intel’s ability to once again establish its dominance in the semiconductor market.
5. Tokyo Electron Revises Profit Outlook in Response to Slowing Sales in Chinese Market Tokyo Electron Limited (TEL), a leading global supplier of semiconductor production equipment, has revised its full-year profit outlook in response to slowing sales in the Chinese market and volatile market conditions. This decision reflects the current situation where technology friction between the U.S. and China and regional demand fluctuations in the semiconductor cycle are having a direct impact on the performance of major suppliers.
6. Qualcomm Launches Initiative to Accelerate Edge AI Deployment at Beijing Summit Qualcomm launched its AI Acceleration Initiative at the Beijing Summit. It aims to leverage the capabilities of the Snapdragon platform to accelerate the deployment of edge AI (AI processing completed on the device) and promote its broad application, including in industry.
7. Southeast Asia Emerges as Key Hub in Fluctuating PCB Supply Chain As the global PCB (printed circuit board) supply chain is being restructured, the Southeast Asia region is rapidly emerging as a new key manufacturing hub. This is driven by companies seeking to streamline manufacturing costs and diversify geopolitical risks concentrated in existing manufacturing hubs, and is indicative of the changing structure of the global electronic component supply chain.
8. Unresolved Tar iff Negotiations and Possible Application of 100% Tariffs on Pharmaceuticals With tariff negotiations still unresolved despite government efforts, former President Trump has warned of the possibility of imposing 100% tariffs on items including pharmaceuticals. This hard-line stance suggests that we are at a tipping point of excessive free trade and that protectionist tariff policies are likely to continue in the next U.S. administration.
9. SSDs Rewrite the Rules of Storage as AI Workloads Increase In response to the enormous data access speeds and parallel processing capabilities demanded by AI workloads, solid state drives (SSDs) are pushing the limits of traditional spinning rust (HDDs) and rewriting the rules of data storage. They are rewriting the rules of data storage. The demand shift to SSDs is accelerating, especially in the field of AI computation, where high-speed access is essential.
10. Hanmi Semiconductor Expands into Advanced Chip Packaging to Meet the AI Era Hanmi Semiconductor is aggressively expanding its business domain into advanced chip packaging to meet the demands of the AI era, which supports high-performance computing (HPC). This is intended to increase competitiveness within the growing AI semiconductor ecosystem by manufacturing equipment for advanced integration technologies, including High Bandwidth Memory (HBM).

News List