Semiconductor News 20260604

Today’s semiconductor news shows that strategic moves are accelerating throughout the supply chain against the backdrop of exploding demand for AI. In particular, TSMC is introducing NVIDIA’s AI technology into its manufacturing process to improve production efficiency, while AMD has announced a $10 billion investment in its Taiwanese ecosystem, and other massive investments in building AI infrastructure continue. In addition, amid rising geopolitical risks, semiconductor exports from South Korea to China are surging 243% y/y, with Beijing’s AI drive a strong driver of memory demand. On the technology side, companies are focusing on pushing the performance envelope, with Samsung introducing its Heat Path Block thermal management technology for next-generation HBM5 and Intel Foundry redefining HPC scalability with its packaging technology. In addition, rising demand for AI servers is also manifesting itself in rising component costs, with Hua Hsin Technology raising the price of passive components, and the industry as a whole is moving into a new AI-driven growth phase.

Main News

1. AMD Invests More Than $10 B illion in Taiwan
   AMD has announced an investment of more than $10 billion in the Taiwanese ecosystem to meet the growing demand for next-generation AI infrastructure. The investment is intended to strengthen strategic partnerships and expand advanced packaging manufacturing capabilities for AI. In response to rapidly growing global demand for AI, AMD aims to maintain and enhance its market competitiveness by strengthening its supply chain with Taiwan’s manufacturing base at its core, and by establishing mass production capacity for cutting-edge technologies.
2. Semiconductor Exports from South Korea to China Up 243% YoY in May – Beijing’s AI Push Boosts Memory Demand
   South Korea’s semiconductor exports to China recorded a significant 243% YoY increase in May 2026. Behind this surge is strong demand for memory chips in line with accelerated AI development in China. This has led to a dramatic improvement in Korea’s trade balance with China, from a deficit in December 2025 to a surplus of $1.1 billion, highlighting how AI-driven demand is strongly supporting semiconductor trade between the two countries, even amid geopolitical tensions.
3. Following Chips, Passive Components to Increase in Price: Hua Hsin Technology Launches First AI Server Price Hike in 2026
   In response to a sharp increase in demand for AI servers, passive components giant Hua Hsin Technology has raised the prices of resistors and some capacitor products effective June 1. In addition to soaring raw material costs, NVIDIA’s next-generation AI platform, Vera Rubin, requires a much larger number of MLCCs than the existing GH200 and B200, which has led to a serious supply crunch. This price increase symbolizes that the cost increase associated with the construction of AI infrastructure has begun to spread to peripheral components other than semiconductor chips.
4. Tail Risk of “China’s Invasion of Taiwan ” Poses Reality of Global AI Semiconductor Business that Cannot Exist without TSMC
   While the global AI semiconductor business is extremely dependent on TSMC’s manufacturing capacity, the geopolitical risk of “China’s Invasion of Taiwan” has emerged as a serious tail risk. The current supply chain, which could not exist without TSMC, is once again highlighted as a vulnerability.
5. Marvell Wins More, Following NVIDIA CEO’s $1 Trillion Praise ; Google Offers Custom Network Chip for TPUv8e on Intel’s 18A/18AP Process
   Marvell has followed up on its high praise from NVIDIA’s CEO with an order to design a custom network chip for Google’s order to design a custom network chip for the next-generation TPUv8e. The chips will be manufactured on Intel’s state-of-the-art 18A/18AP process, which is becoming increasingly important for network chips that synchronize the vast number of GPUs and ASICs in AI data centers, and the combination of Marvell’s design capabilities with Intel’s manufacturing process will accelerate the efficiency of AI infrastructure. The combination of Marvell’s design capabilities and Intel’s manufacturing processes will accelerate the efficiency of the AI infrastructure.
6. Samsung Gives Glimpse of HBM5 with New “Heat Block Path” Technology, Reminiscent of SK Hynix’s iHBM Cooling Strategy
   Samsung plans to introduce a new thermal management technology, Heat Path Block (HPB), in the development of its next-generation memory standard, HBM5 This is a new thermal management technology introduced by SK Hynix. This is similar to the cooling strategy introduced by SK Hynix for iHBM and is intended to solve the heat generation problems associated with high-load operations in AI data centers. Memory manufacturers are competing with each other to solve the thermal bottleneck of AI performance, and competition for next-generation HBM is intensifying.
7. TSMC Expands Use of NVIDIA’s AI Technology Across Chip Manufacturing
   TSMC is further expanding the implementation of NVIDIA’s AI technology across its chip manufacturing processes. The company aims to maximize production efficiency of increasingly complex advanced semiconductors by accelerating yield improvement and process optimization through the use of AI in manufacturing. The deep integration of NVIDIA’s AI solutions by the world’s largest foundry will further enhance the availability and quality of AI chips and drive industry technology standards.
8. Synopsys and Samsung Foundry Expand AI-Driven Design, Test, and 3DIC Enablement for Advanced Nodes
   Synopsys and Samsung Foundry announced today that they are expanding their AI-driven design, test, and 3DIC (three-dimensional integrated circuit) technology offerings for advanced nodes AI-driven design automation tools enable faster chip development and optimized performance. The collaboration between the two companies will play an important role in dramatically increasing the design efficiency of next-generation AI processors, especially in the construction of complex 3DIC structures.
9. From HBM to Physical AI…Korean Companies Increase Their Presence at COMPUTEX 20 26
   At COMPUTEX 2026, Korean semiconductor companies showed a dominant presence in the fields of HBM and physical AI. Korean innovations are attracting attention not only in memory technology, which is the foundation of AI data centers, but also in the area of physical AI, which integrates the physical world and AI. In the global AI ecosystem, Korean companies were shown to be not only a source of hardware, but also in a position to lead innovation in next-generation AI.
10. How Intel Foundry Packaging Technologies Redefine the Scalability Limits of AI and HPC
    Intel Foundry has announced new packaging technologies that push the performance limits of AI and HPC (high performance computing). In particular, technologies such as “EMIB-T” enable the creation of large multi-die systems while reducing cost and manufacturing risk. This innovation breaks through the physical constraints that have prevented AI chips from scaling up and contributes to a higher-performance AI computing environment.

News List (236 items)

This article is delivered every morning by an automatic collection and AI generation system. Number of collected articles : 236 | Update date : 20260604

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.