NVIDIA has announced the expansion of its open-source CUDA-Q platform with the introduction of the CUDA-Q Logical orchestration layer. This orchestration layer provides programmable and verifiable tools for the development of fault-tolerant quantum computing applications, supporting multiple application scenarios, including drug discovery. Fermi National Accelerator Laboratory has leveraged this orchestration layer to significantly reduce the development time of fault-tolerant algorithms from approximately 5 months to just 3 weeks, achieving an efficiency improvement of about 7 times. Meanwhile, QUOPS, a quantum computing benchmark developed by Sandia National Laboratories, has also been integrated into CUDA-Q to evaluate the progress of quantum computing hardware in practical fault tolerance across different platforms. Currently, this orchestration layer has been adopted by multiple quantum computing companies and research institutions.

On September 14, Hangzhou Changchuan Technology (300604.SZ) made a public announcement regarding its strategic move to issue H shares and seek a listing on the Main Board of the Hong Kong Stock Exchange. This initiative is primarily driven by the company's ambition to propel its globalization agenda forward, expedite the growth of its overseas business ventures, bolster its financial robustness and overall competitive edge, and ultimately set up an international capital operation platform. As of now, while the board of directors has given the green light to the pertinent proposals, other precise particulars are still under deliberation and have yet to be finalized.

On September 14, the three key US stock indices commenced trading with a downward trajectory. The Dow Jones Industrial Average dipped by 0.2%, the S&P 500 Index saw a 0.7% decline, and the Nasdaq Composite Index plummeted by over 1.2%. The AI sector felt the pinch as leading AI companies called for a deceleration in development pace. Consequently, AI chip stocks experienced a broad-based retreat. Nvidia's shares tumbled by 3.3%, Intel's stock took a 7% hit, and AMD's shares fell by 5.2%. In the memory chip arena, Micron Technology's stock price dropped by 6.7%, SK Hynix witnessed a 7.1% decline, and SanDisk's shares decreased by 5.8%. Furthermore, Oracle's stock price slipped by 5.2% amid reports that the company has embarked on a fresh round of layoffs.

Recently, the 22nd IEEE Conference on Electromagnetic Field Computation (CEFC 2026) took place in Thessaloniki, Greece. This conference served as a pivotal platform for sharing cutting-edge theories, methodologies, and applications across a broad spectrum of topics, including electromagnetic field computation, electromagnetic modeling and simulation, electromagnetic measurement and testing, electric machines, and energy systems.
Chang Chao, a doctoral candidate specializing in Instrument Science and Technology at the School of Automation Engineering, University of Electronic Science and Technology of China, delivered an insightful academic presentation at the conference. His talk, titled "PyECSim: A Python Library for Automated Eddy Current Non-Destructive Testing Simulation," introduced PyECSim, an innovative automated simulation tool for eddy current non-destructive testing developed by his team.
Built on the robust foundation of Python, PyECSim offers a user-friendly, high-level simulation interface. It streamlines the process of sensor, specimen, and defect modeling within Ansys Maxwell 3D, along with the programmatic setup of materials, excitations, meshing, and solution processes. Furthermore, PyECSim seamlessly integrates with computational frameworks like PyTorch and NumPy, facilitating parameter scanning, optimization calculations, and batch simulations.
The research findings demonstrate that PyECSim can significantly reduce the volume of modeling code required, enhance 3D simulation efficiency by up to 18-fold or more, and automatically generate natural defect models along with an extensible defect sample library. This advancement lays a solid technical groundwork for the intelligent optimization of eddy current sensors, the development of digital twins, and AI-driven electromagnetic simulations.
Following his presentation, Chang Chao actively engaged in fruitful exchanges and discussions with participating experts and scholars. The topics of conversation spanned automated modeling for eddy current non-destructive testing, strategies for improving electromagnetic simulation efficiency, and the latest trends in intelligent electromagnetic simulation.

On September 14, Huazhijie (603400.SH) disclosed its intention to acquire 100% of the equity in Suzhou Geliming Electronic Technology Co., Ltd. The acquisition will be executed through a combination of share issuance and cash payment, with plans to raise supplementary funds (Note: In financial contexts, 'matching 资金' is translated as 'supplementary funds' or 'matching funds'. Here, 'matching funds' is retained for precision unless a specific context necessitates an alternative). It is anticipated that this transaction will not qualify as a major asset restructuring, a restructuring listing, or a related-party transaction. Furthermore, it is not expected to lead to any alteration in the company's controlling shareholder or actual controller. Trading of Huazhijie's stock will be temporarily halted from the market open on September 15, 2026, with the anticipated duration of the halt not exceeding 10 trading days.

On September 14, during the pre-market trading hours for US stocks, the tech sector displayed a range of performances. Nvidia's stock took a significant hit, declining by more than 2%. Tesla and Amazon also experienced downturns, each dropping over 1%. In contrast, Meta and Google saw their stock prices rise by more than 1%, Microsoft's shares increased by 0.64%, and Apple's stock edged up by 0.16%. Meanwhile, memory chip stocks took a nosedive, with SK Hynix and Intel both falling more than 6%.

Data sourced from Tianyancha reveals that Caihong Display (Xianyang) Semiconductor Technology Co., Ltd. has been newly established, with Wang Wentao appointed as the legal representative and a substantial registered capital of 140 million yuan. The company's operational scope encompasses the production of display devices, semiconductor discrete devices, optoelectronic devices, glass, and technical glass products, among others. Its shareholders comprise Caihong Display Devices Co., Ltd., Xianyang High-tech Industrial Development Investment Co., Ltd., Xianyang Industrial Investment Group Co., Ltd., as well as Shaanxi Caitou Xianyang Angel Fund Partnership (Limited Partnership).

Awinic Technology (Shanghai) Ltd. has unveiled a cutting-edge 12V/1.4A stepper motor driver, which not only supports digital I2C control and 1/256 microstepping but also incorporates current and stall detection functionalities. This driver is ideally suited for precision equipment, including pan-tilt systems, optical telephoto lenses, and AR glasses.

Chipower Technology has revealed that, for the third quarter of 2026, it projects the net profit attributable to the owners of the parent company to reach approximately 21.5 million yuan, marking a significant year-on-year shift from losses to gains. The key factors driving this performance transformation include the cyclical resurgence of the semiconductor sector, a surge in demand for packaging and testing services, and a notable 61.26% year-on-year uptick in operating revenue. Moreover, the company's capacity expansion efforts have led to a reduction in unit manufacturing costs. Concurrently, the implementation of lean management practices, along with cost-cutting and efficiency-boosting initiatives, has further enhanced production efficiency. As per Xiaocai's analysis, the company's net profit for the third quarter is expected to hit 21 million yuan, up from 16 million yuan in the second quarter, indicating a projected 37% quarter-on-quarter increase in net profit for the third quarter.

Latest data from market research firm Omdia shows that global semiconductor industry revenue reached $425 billion in the second quarter of 2026, marking a new all-time high. Driven by the sustained surge in demand for enterprise-level AI hardware, quarterly revenue increased by 31.4% sequentially, far outpacing the industry's long-term growth patterns and breaking decades of market performance norms. AI demand is reshaping production priorities for memory suppliers, with DRAM, NAND, NOR, and other memory chips achieving record-high sequential growth rates in the second quarter of 2026. Meanwhile, other semiconductor categories also demonstrated growth surpassing seasonal cyclical trends, with sequential increases exceeding 10%. Omdia forecasts that single-quarter revenue will further expand to over $500 billion in the third quarter of 2026.

Earlier this month, the Party building department of the Tianjin Haihe Education Park Administrative Committee interviewed Dr. Wang Pan, co-founder and CEO of Speed Microelectronics, whose new headquarters is located in Jinnan District, Tianjin.

NVIDIA's CEO, Jensen Huang, is making proactive investments in emerging cloud service providers and AI firms in a bid to diversify and attract new clientele. Nevertheless, despite the company's impressive accomplishments, a significant portion of its sales revenue is becoming increasingly concentrated among a limited number of customers, with the possibility that some of these customers may opt to decrease their purchases in the foreseeable future.

Japanese NAND flash memory giant Kioxia is considering going public in the United States, planning to raise $10 billion, and has held discussions with Bank of America, Goldman Sachs, and JPMorgan Chase.

Recently, Loongson Technology made an announcement revealing that the container template server associated with the Linux Containers (LXC) virtualization toolkit has formally introduced LoongArch container templates. This development indicates that LoongArch has now garnered official endorsement from the LXC container template repository, making it the fifth architecture to achieve this distinction, following in the footsteps of x86-64, AArch64, AArch32 (ARMv7), and RISC-V.

In the second quarter of 2026, the landscape of global fabless semiconductor companies witnessed new changes. NVIDIA maintained its top position with quarterly revenue of USD 91.16 billion, a year-on-year increase of 99%, accounting for 64% of the total revenue of the top ten manufacturers. Broadcom's revenue reached USD 18.89 billion, propelling it to the second place, with a remarkable year-on-year growth rate of 112%. Through diversified Layout (strategic Layout translates better in context as 'strategic Layout ' or more naturally as 'diversified strategies'), it successfully expanded its market share in data centers and automotive electronics. AMD's revenue stood at USD 11.53 billion, surpassing Qualcomm to claim the third spot. Leveraging AI technology to upgrade its CPUs, AMD has consistently set new records in its data center business, while also increasing its market share in the X86 server market. Qualcomm slipped to fourth place, with its revenue of USD 8.5 billion affected by fluctuations in the smartphone business. However, its automotive chip business has maintained double-digit growth for 23 consecutive quarters. Currently, Qualcomm is restarting its data center business and actively exploring emerging fields. MediaTek ranked fifth with revenue of USD 4.81 billion. Affected by the slowdown in demand for smartphone chips, the company is increasing its investment in AI application-specific integrated circuit (ASIC) business. It is projected that its data center-related revenue will exceed USD 2 billion in 2026, while also raising its 2028 ASIC serviceable market size forecast to USD 80 billion.

Fujitsu is poised to commence the export of its 2nm process AI chip, Monaka, which was developed leveraging the technological prowess of the 'Fugaku' supercomputer, to the United States and various Asian markets as early as 2027. The company has already inked a supply agreement with Supermicro, a prominent U.S. server manufacturer, and is strategizing to extend its reach to major cloud service providers across the United States.

On September 14, JuCan Optoelectronics made an announcement stating that it has officially forwarded the inaugural batch of CPO optical chip samples to its principal clients for assessment. Furthermore, the prototype units have been duly delivered as planned, signifying the product's progression from the research and development phase to the customer verification stage.

According to media reports, the competitive focus in the semiconductor industry has shifted to talent acquisition. To maintain its market leadership, Samsung Electronics is strengthening its efforts to protect its core talent. Meanwhile, international competitors such as Micron and SanDisk are also entering South Korea to compete for HBM and NAND flash memory design talent. The ability to attract and retain key technical talent has become crucial in determining the outcome of the next-generation memory market competition. In response, Samsung Electronics has implemented measures such as increasing performance bonuses and expanding equity awards. In the first half of this year, despite a slight decrease in the number of employees, both total compensation and book compensation costs have increased.

An Omdia report reveals that in the second quarter of 2026, the global semiconductor revenue catapulted to over US$425 billion, hitting an all-time peak and registering a remarkable quarter-over-quarter (QoQ) growth of 31.4%. The burgeoning demand for artificial intelligence technologies, coupled with the robust pricing trends of memory chips, emerged as the primary market drivers. These factors collectively propelled the total semiconductor revenue for the first half of 2026 to an impressive US$752 billion. Notably, this 31.4% QoQ growth rate eclipsed the previous high of 29.2% recorded in the first quarter of the same year, underscoring the industry's exceptional momentum.

On September 14, at the 2026 mid-year performance briefing, Cheng Zhuo, Chairman of Canon Kejia, announced that the company anticipates a continued strong delivery of high-end PCB equipment orders in the latter half of the year. He also mentioned that advanced packaging equipment is now entering the verification and implementation stages. Looking ahead to next year, the expansion of AI servers, associated carrier boards, and advanced packaging technologies is projected to spur additional demand for equipment.

During the first half of the year, Canon Kejia's Phase II plant was in the process of ramping up production, achieving high overall capacity utilization. The company expects to fully unleash the capacity of Phase II in the second half of the year. By leveraging modular production techniques and pre-stocking essential components, Canon Kejia is well-positioned to continue accepting orders for high-end PCB and advanced packaging equipment.