Semiconductor News 20260713

Today’s semiconductor news highlights the impact that the explosive growth in AI demand is having on the entire semiconductor supply chain. In particular, the shortage of TSMC’s advanced packaging technology, “CoWoS,” is worsening, leading to a structural shift in which orders are flowing to competitors such as Intel as a result.Furthermore, the shift by major tech companies toward in-house designed chips is accelerating, as evidenced by reports that Meta will begin mass production of “Iris,” an AI chip jointly developed with Broadcom, in September.There is a growing consensus in the market that memory bandwidth, rather than computing power, is the bottleneck for AI, leading to a surge in demand for related test equipment manufacturers and PCB suppliers.Meanwhile, the financial results of TSMC and ASML are being closely watched as key indicators that will determine the future course of the AI rally, and the importance of ripple effects on regional economies—as seen in the Kyushu Semiconductor Park initiative—and of public-private partnership investments is being reemphasized.With technological innovation and the expansion of supply capacity proceeding in parallel, the industry is entering an unprecedented phase of growth. In particular, it is significant that the expansion of AI inference is driving demand not only for HBM but also for DRAM, eSSDs, CPUs, power supply components, and advanced packaging.

Main News

1. Meta may begin production of its self-developed AI chip “Iris” in September.
   Meta’s first in-house developed AI chip, “Iris,” co-designed with Hakutsu and manufactured by TSMC, is expected to begin mass production as early as September of this year. This move aims to reduce reliance on NVIDIA and AMD GPUs, lower the cost of AI infrastructure, and improve training and inference efficiency. Meta is accelerating its investment in AI computing power, planning to double its computing capacity to 14 gigawatts by 2027. It has also begun offering an API for its proprietary large-scale language model, “Muse Spark 1.1,” further diversifying its AI business.
2. TSMC Can’t Keep Up with CoWoS Demand; Advanced Packaging Orders Are Shifting to Intel and Other Companies
   As demand for AI and HPC chips reaches unprecedented levels, TSMC’s supply capacity for its advanced packaging technology, “CoWoS,” is reaching its limits.This supply shortage has led some customers to shift their orders to competing foundries such as Intel. Although TSMC boasts industry-leading technological capabilities, it is struggling to meet the surging demand for AI chip packaging on its own, highlighting once again the critical importance of the packaging process in semiconductor manufacturing.
3. Computing power is no longer the bottleneck for AI. Memory is—and suppliers recognize this.
   There is a growing consensus that, in the evolution of AI, memory bandwidth and capacity—rather than computing power—are the biggest bottlenecks.Suppliers are keenly aware of this situation and are rushing to develop next-generation memory technologies, such as HBM (High-Bandwidth Memory), and strengthen their supply chains. As AI models grow in size, memory performance has become the key factor determining overall system performance, and industry investment is shifting from computing chips to memory-related technologies.
4. TSMC and ASML, the global stock market, predict earnings; if semiconductor demand is strong, AI rally could resume
   The earnings announcements of TSMC and ASML, which have a significant impact on global stock markets, have attracted attention. Semiconductor stocks were in a correction phase due to short-term overheating, but capital is returning due to expectations for AI growth. If strong demand is confirmed in the earnings reports of both companies, a stock price rally centered on AI-related stocks is expected to resume, with major indices such as the Nikkei Stock Average aiming for higher prices once again. The market is closely watching the outlook for both companies, which serve as leading indicators for the semiconductor industry.
5. Apple’s chipset development is unparalleled, and they are already working on the M8.
   Apple has established a dominant position in developing its own chipsets, and it has been reported that it has already begun development on the next-generation M8, in addition to the M7. The M8 aims to significantly improve AI processing power and power efficiency on devices, clearly demonstrating Apple’s commitment to strengthening AI capabilities optimized for its own ecosystem. While other companies rely on general-purpose GPUs, Apple maintains its unique competitive advantage with its vertically integrated chip design.
6. Hermes Testing’s June revenue more than doubled amid surging demand for advanced chip testing
   Hermes Testing, a manufacturer of semiconductor test equipment, saw its June revenue more than double year-on-year, driven by the expansion of advanced testing needs driven by a surge in demand for AI chips. The increasing complexity of AI chips has increased the load and required accuracy in post-manufacturing testing processes, driving a surge in demand for the company’s testing equipment. These results indicate that the expansion of the AI semiconductor market is bringing significant benefits not only to manufacturing but also to suppliers involved in testing and inspection processes.
7. Taiwanese PCB giant launches funding round for AI data centers
   Driven by a construction boom in AI data centers, major Taiwanese printed circuit board (PCB) companies have begun large-scale fundraising. Advanced multilayer PCBs and high-speed transmission-compatible PCBs are essential for AI servers, and increasing production capacity to meet the surging demand is a matter of urgency. The companies will use the funds raised for capital investment and technological development, aiming to secure a market share in the expanding AI infrastructure market.
8. The “Dream” of TSMC Attracts Investment: A Future Where Kyushu Semiconductor Park Becomes a Mecca
   With TSMC’s entry into Kumamoto as a catalyst, efforts to revitalize Kyushu as “Silicon Island” are gaining momentum. As construction of the second plant progresses, the region faces the challenge of maximizing the benefits for local industries.Although the barriers to direct business with TSMC remain high, with the clustering of related companies and the development of talent gaining momentum, there are high hopes for a future in which Kyushu will establish itself as a global semiconductor supply hub. Joint public-private investment is beginning to function as a growth engine for the regional economy.
9. Public and private investment is focused on AI and semiconductors; instead, we need to provide stronger support to struggling foundational industries and reconsider the priorities of crisis management investments.
   As public and private investment in AI and semiconductors accelerates, strengthening the semiconductor industry as a foundational industry has become a crucial national strategic issue. The importance of continuing stable support even during market downturns is being emphasized, and there are suggestions to reconsider the priorities of crisis management investments from an economic security perspective. A long-term perspective is needed, one that looks beyond mere growth investment to strengthening supply chains.
10. Elon Musk accuses Sam Altman of stealing charity and Apple technology
    Elon Musk fiercely accused OpenAI’s Sam Altman of abusing the spirit of charity and stealing Apple’s technology. Musk argues that while SpaceX’s AI1 satellite, currently under development, will launch next year, Altman will likely be imprisoned by then due to legal issues. The conflict between the two symbolizes the industry’s tensions over AI development ethics and technology ownership, drawing attention to future legal developments.

News List (108 items)

This article is delivered every morning via an automated collection and AI-generated system. Number of items collected: 108 | Last updated: July 13, 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.