Frequent Dynamics of Global Semiconductor Manufacturers, Accelerated Evolution of the Industrial Landscape

Recently, global semiconductor manufacturers have been constantly making moves in aspects such as technological cooperation, product manufacturing processes, and market layout, demonstrating the rapid development and transformation of the industry.

In terms of technological cooperation, NVIDIA has reached an important agreement with Fujitsu of Japan. According to the news from the “Science and Technology Innovation Board Daily” on October 5th, the two sides will jointly build a full-stack AI infrastructure integrating AI agents, focusing on developing industry-specific AI agent platforms for fields such as healthcare, manufacturing, and robotics. Before 2030, they plan to connect their respective semiconductors on the same substrate through NVIDIA NVLink Fusion technology to achieve ultra-high-speed interconnection among multiple chips like GPUs and CPUs. Fujitsu is currently developing the “MONAKA” CPU based on the Arm architecture with a 2-nanometer process. Its goal is to achieve twice the power efficiency of other companies’ CPUs and plans to put it into practical use in 2027.

Regarding product manufacturing processes, TSMC is progressing smoothly. The company plans to start mass production of the N2 process in the fourth quarter of 2025, earlier than expected. Currently, its factories in Kaohsiung and Hsinchu have already entered the trial production stage. Apple, NVIDIA, AMD, Qualcomm, and MediaTek are among its major clients. In addition, TSMC has already started the construction of a 1.4-nanometer factory in Taiwan, with a plan to start mass production in the second half of 2028. In the second quarter of 2025, TSMC’s revenue reached a record high of $30.1 billion, with a year-on-year growth of 44%. It has also raised its revenue growth guidance for the whole year of 2025 from 25% to approximately 30%.

In terms of market layout, Dr. Lisa Su, the CEO of AMD, responded to the rumor that AMD might invest in Intel and adopt Intel’s chip foundry services. She didn’t directly deny the possibility of cooperation but emphasized that the company has a solid supply chain with TSMC and reiterated its strategy of giving priority to production in the United States. In the field of memory chips, according to the news from the “National Business Daily” on September 18th, DRAM manufacturers in South Korea and the United States have suspended quotations for corporate customers for one week. The DRAM quotations in the fourth quarter of this year may increase by more than 30%, and the price increases of some DRAM specifications may exceed 50%. Market surveys show that the supply gap of DDR4 memory will reach 10% – 15% in the next three quarters, and the back-end packaging and testing capacity has also reached a bottleneck. Additionally, according to the news from the “Cailian Press” on October 8th, Jensen Huang, the CEO of NVIDIA, confirmed that NVIDIA has made an investment in xAI, a start-up company under Elon Musk. The investment amount is as high as $2 billion, and this financing is linked to the NVIDIA graphics processing units that xAI plans to use in “Colossus 2”.

Furthermore, the policy environment of the global semiconductor industry is also changing. On October 3, 2025, the European Union officially passed the “Chips Act”, planning to invest 48 billion euros by 2030 to boost the local chip industry. The United States simultaneously announced the detailed rules for the $52 billion chip subsidy, clearly requiring the subsidized enterprises not to expand advanced manufacturing capacity in China within ten years.

As a key area in global technological development, the semiconductor industry has drawn much attention. The series of actions by various manufacturers will have a profound impact on the future technological industrial landscape.