China has entered a new era in artificial intelligence computing with the completion and connection of the Sugon Dawn 8000 Dengfeng cluster to the National Supercomputing Internet. This cluster is the first fully domestic super-AI fusion cluster capable of operating with 100 thousand cards.
Dawn 8000 utilizes a super-AI fusion architecture that combines computational precision from FP64 to INT8 in a single system. This eliminates the need for traditional separation between infrastructure for scientific computing and AI training. The cluster is built on high-performance Hygon processors and employs the scaleFabric network technology for efficient inter-node communication, as well as the ParaStor distributed storage for high data throughput.
Even after launch, the cluster was optimized for over 300 applications across more than 20 scientific and industrial fields. These include protein folding simulation, molecular dynamics of trillions of atoms, and turbulence modeling with a grid of 100 trillion. The significance of Dawn 8000 goes beyond raw computational power, as its architecture is designed for super-AI fusion, combining the accuracy requirements of traditional high-performance scientific calculations with the flexible precision requirements of AI model training and inference.
This project demonstrates the maturity of China's domestic AI chip ecosystem, enabling the creation of 100,000-card clusters exclusively using internal components—processors, networking equipment, and storage systems. Dawn 8000 serves as a reference architecture for future AI infrastructure projects, providing proven design templates for large-scale domestic computing. Furthermore, due to its integration with the National Supercomputing Internet, Dawn 8000 resources are accessible to users across China, democratizing access to advanced AI computing capabilities for universities, research institutions, and enterprises that cannot afford their own clusters of this scale.
Against the backdrop of global discussions about whether China is experiencing a deficit or surplus of computing power, China's semiconductor center in Zhangjiang, Shanghai, presents a clearer picture: stable and universal computing power remains in acute shortage, with orders for certain chips scheduled for next year and beyond.
Sun Guolian, Product Director at the GPU manufacturer MetaX (MXC series), noted that the shortage is observed regarding universal and reliable computing power. However, there is a surplus in non-universal architectures that can only operate with outdated or highly specialized models. Chips from this company's MXC600 series are supplied in large batches, and next-generation product development is in advanced stages.
Parallel to traditional electronic computing, Zhangjiang positions itself as China's national silicon photonics center. The Shanghai Zhangjiang silicon photonics cluster now unites over twenty companies covering the entire production cycle. Among them stands Lightelligence—the world's first company producing AI silicon photonic chips, which has been listed on the Hong Kong Stock Exchange.
Shen Yichen, founder of Lightelligence, who holds a Ph.D. from MIT and has previously published in the journal Nature, forecasts that silicon photonic chips will account for over 30% of the intelligent computing center market within five years, compared to less than 3% today. The company has already implemented hybrid optoelectronic solutions in three optical interconnect clusters of a thousand cards, enabling the operation of over 5000 GPUs for 44 corporate clients, including research institutions, internet companies, GPU manufacturers, and AI infrastructure operators.
The Zhangjiang 895 incubator has attracted several developing photonic startups, such as Qisuan Guangqi, specializing in optical chip R&D, and Guangjiu Technology, focusing on photonic technologies. The cluster also includes JCET (Changdian Technology) for advanced packaging and Suzhou Yite for silicon photonics testing, ensuring full coverage of the chain from design to testing.
The Zhangjiang model represents a dual approach to solving China's computing power problem: developing traditional electronic computing to ensure universal reliability while simultaneously creating a silicon photonics ecosystem. This ecosystem potentially offers a breakthrough in optical interconnects, allowing thousands of GPUs to function as efficiently as one chip by transmitting data via optical rather than electrical channels.
AIFU Semiconductor, based in Guangzhou, is increasing its production of BAW filter chips to one billion units. This allows it to enter a market previously controlled by Broadcom and Qorvo in the 5G radio frequency filter segment, utilizing patented aluminum nitride single-crystal technology.
AIFU Semiconductor is the only company in China and the third globally to master the complete production chain for BAW filter chips for medium and high-frequency 5G bands. The company aims to achieve an annual production capacity exceeding one billion chips, with a stable yield rate of 98–99%. The company's products are already integrated into the production lines of major terminal manufacturers such as Samsung and ZTE.
BAW chips serve as key elements in the radio frequency front-end systems for 5G and higher networks, directly influencing signal reception quality and the ability of smart terminals to resist interference. Previously, American firms Broadcom and Qorvo effectively monopolized the markets and patents for 5G filters. AIFU Semiconductor managed to overcome this barrier by developing its own technology based on aluminum nitride single crystal (ScAlN), bypassing patent restrictions related to polycrystalline aluminum nitride established by foreign giants.
Unlike the polycrystalline aluminum nitride used in the industry, aluminum nitride single crystal provides superior piezoelectric characteristics and acoustic wave transmission efficiency, making it an ideal material for high-performance BAW filters. The company built the world's first 8-inch mass production line for BAW chips in Guangzhou. In 2015, AIFU Semiconductor launched a pilot 5G SABAR BAW line with a monthly capacity of 500,000 units. By 2020, it expanded production to 30 million units per month. Today, the company maintains stable mass production of nearly 100 chip models.
Thanks to systematic technological breakthroughs in materials, simplification of chip manufacturing processes, and domestic raw material supply, the company has achieved a 30–40% reduction in overall production costs compared to foreign counterparts. Meanwhile, SABAR chips demonstrate an improvement in key performance indicators of more than 20%. One smartphone typically requires over 30 filters covering various frequency ranges, and AIFU Semiconductor's products now support frequencies above 10 GHz. The company has upgraded about 12 sets of semiconductor equipment to support its patented process, with some core components developed in-house. The company's 50 mu (3.3 hectares) headquarters in Guangzhou has completed construction and is undergoing cleanroom equipping.
In addition to mobile communications, AIFU Semiconductor's high-frequency filters are used in satellite internet, Internet of Things (IoT), and automotive communication. The company sees robotics as the next frontier for growth, as robots transition from pre-programmed control to autonomous environmental perception, which requires intensive interaction with wireless signals. Furthermore, the company plans to bring its SAMES sensors to the industrial inspection and automotive electronics markets, aiming to double revenue this year.
