Semiconductor News 20260615

Today’s semiconductor news highlighted a day where the evolution of AI infrastructure intersected with geopolitical industrial strategies. Of particular note is the fact that NVIDIA’s GB300 has recorded a 20-fold performance improvement over the Hopper generation for agent-based AI workloads, accelerating the evolution of AI computing.In response, collaboration between hardware and power infrastructure is deepening, as evidenced by Infineon joining NVIDIA’s MGX ecosystem to enhance power management solutions for AI factories.Meanwhile, the South Korean government has launched approximately $329 million in R&D investment for next-generation power semiconductors, embarking on efforts to cultivate a new revenue source to succeed DRAM.However, developments surrounding the semiconductor supply chain in Asia are complex, and NVIDIA CEO Fan’s decision to exclude Japan from his itinerary is being viewed as a warning sign regarding Japan’s competitiveness in AI.Furthermore, with the U.S. government tightening regulations surrounding AI sovereignty—such as export restrictions on Anthropic’s AI models—tension within the industry is rising on both the technological development front and in terms of compliance with international regulations.

Main News

1. NVIDIA GB300 Dominates Agent-Based AI Workloads with 20x Performance Boost Over Hopper; Rubin Launch Imminent
   NVIDIA’s next-generation Blackwell GB300 chip has achieved an overwhelming record on AA-AgentPerf , a new benchmark that measures the inference performance of AI agents.Delivering a 20x performance boost over the Hopper generation, it will powerfully drive the adoption of agent-based AI capable of autonomously handling complex tasks. With the Rubin architecture on the horizon, the performance standards for AI infrastructure are poised to be dramatically redefined.
2. Power semiconductors used in AI data centers could become the next “cash cow” on par with DRAM, following the launch of a $329 million R&D initiative in South Korea
   Anticipating increased power demand from AI data centers, the South Korean government has launched a 500 billion won R&D project aimed at mass-producing next-generation power semiconductors.The goal is to replicate the success model established by Samsung and SK hynix in the DRAM sector within the power semiconductor industry, with the national strategy focused on improving self-sufficiency and strengthening competitiveness in semiconductor technologies that form the foundation of AI infrastructure.
3. South Korea Launches 500 Billion Won R&D Initiative to Develop Next-Generation Power Semiconductor Technology
   As part of the “Ultra-Innovative Economy Project” led by the South Korean government , a massive amount of funding will be invested in the development of next-generation power semiconductors. The initiative aims to establish power control technologies essential for improving the efficiency of AI data centers and to expand market share in the global market.By providing comprehensive support ranging from technology development to the establishment of mass production systems, the goal is to diversify the semiconductor industry’s portfolio.
4. Infineon Joins NVIDIA’s MGX AI Factory Ecosystem
   Infineon Technologies has announced its participation in NVIDIA’s MGX architecture ecosystem.The company will provide 800 VDC-compatible power management solutions optimized for AI factories, enhancing the scalability and power efficiency of AI infrastructure. With an eye toward the era of agent-based AI, this is a significant partnership that will accelerate the standardization and modularization of the power supply infrastructure supporting high-performance computing.
5. U.S. Government: “Sanctions on Anthropic Will Not Be Extended to Other AI Companies”… Reason Cited: “Uncooperative Attitude”
   The U.S. government has clarified that it has no plans at this time to extend the export restrictions imposed on Anthropic’s AI model to other companies.It has been emphasized that these measures are a case-by-case response to a specific company’s uncooperative stance toward government security requirements. This situation raises questions about the nature of AI regulation and corporate compliance.
6. U.S. Imposes Export Controls on Cutting-Edge AI Models; Securing AI Sovereignty Is Now an Urgent Priority
   The U.S. government’s issuance of an executive order blocking overseas access to Anthropic’s latest AI model, “Mitos 5,” and others demonstrates that AI technology is at the core of national security.This is the world’s first export control on AI models, making it an urgent priority for nations to develop their own foundational models and strengthen security technologies to secure their AI sovereignty. The international competition in AI development has entered a new phase.
7. NVIDIA CEO Skips Japan: Partner’s Appeal Falls Short of South Korea and Taiwan
   The fact that NVIDIA CEO Jensen Huang visited Taiwan and South Korea but skipped Japan highlights the declining competitiveness of Japan’s semiconductor industry and its lag in the AI revolution.As Taiwan and South Korea build strong partnerships as core players in the AI ecosystem, Japan must transform its industrial structure and deepen collaboration to maintain and expand its presence in the AI supply chain.
8. Intel’s Canceled Arctic Sound Xe-HP GPU Resurfaces from a 2021 Roadmap, Featuring Dual Tiles and Quad HBM2E
   Images of Intel’s Arctic Sound 2T GPU—whose development was announced to have been canceled several years ago—have been released and are attracting attention.Based on the Xe-HP architecture and featuring a dual-tile configuration and quad HBM2E memory, this GPU was an ambitious product that Intel was developing at the time to target the data center market. It is now generating buzz as a valuable historical document offering a glimpse into the company’s development history.
9. A retail investor who once rose to the rank of department head returns to the U.S. market… “Snapping up” a 3x semiconductor ETF
   Spurred by the booming U.S. market, Japanese retail investors are accelerating their aggressive buying of leveraged semiconductor-related ETFs.In particular, capital inflows into ETFs aiming to track three times the movement of the semiconductor index are notable, reflecting high expectations for the growth of semiconductor stocks amid the AI boom. This aggressive investment stance, which embraces risk, is one factor supporting market liquidity.
10. Soaring Japanese Stocks: How Did Retail Investors React? The Truth Behind the “Bull Market” Revealed Through Mutual Fund Flows
    This report analyzes how retail investors moved their funds through mutual funds during the recent surge in Japanese stocks.Amid growing enthusiasm for semiconductor-related stocks, the report details the investor psychology and capital flows underpinning the bull market. The trends among retail investors serve as a crucial indicator that cannot be ignored when assessing market overheating and future prospects.

News List (110 items)

This article is delivered every morning via an automated collection and AI-generated system. Number of items collected: 110 | Updated: June 15, 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.