Chinese semiconductor thread II

[tokenanalyst]

Lianhe Capital completes investment in semiconductor equipment company Chipone Precision.​

United Capital Fund III and IV recently completed an investment in Ningbo Xinfeng Precision Technology, a leading semiconductor equipment company founded in 2021. The round attracted participation from major investors, including Tuojing Technology and SMIC Capital. This funding will support the company’s technology iteration and capacity expansion as it solidifies its position in the market for 3D integration and advanced packaging equipment.

Focused on R&D and manufacturing ultra-precision processing equipment and consumables, Xinfeng Precision has developed a comprehensive product portfolio, including 12-inch wafer grinders, tremmers, and laser cutting machines. The company specializes in key areas such as micron-level thinning and is advancing capabilities for next-generation applications like 3D stacking, AI chips, and HBM. Its equipment, including ring slicing machines, has already achieved mass production and is shipping to leading domestic memory manufacturers, effectively enhancing chip yield and reliability.



It is reported that Chipone Precision has launched a 12-inch thinning film application and removal machine based on advanced design concepts to meet the needs of 3D IC processes. It supports ultra-thin wafers and high warpage processes, provides a variety of customized solutions, is compatible with multiple process routes such as DBG and TCB bonding, and provides customized upgrade services for Chip to Wafer bonding processes.

Chipone Precision's ring slicing machine is mainly used in the manufacturing process of AI-related chips such as 3D NAND, HBM, and DRAM. By performing ring grinding on the edge of the wafer, it effectively removes edge defects and significantly improves the yield and reliability of 3D stacked chips.

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

 

[tokenanalyst]

ScaleX HyperCluster Breaks Records​

The core node of the National Supercomputing Internet was recently launched for trial operation. According to reports, the node simultaneously deploys three sets of Sugon's ScaleX 10,000-card supercluster systems, becoming the largest domestically produced AI computing power pool in China to achieve a deployment of 30,000 cards and be actually operational. From its first real-world demonstration last December to its large-scale deployment, the 10,000-card supercluster was completed in less than two months.



With strong support from national ministries, Henan Province, Zhengzhou City, and other parties, this facility has been deployed as a key project. It has not only achieved many industry firsts, but also provided a national-level model for the construction of China's computing power network.

Highest integration: The scaleX super cluster adopts the world's first high-density single rack, based on ultra-high density blades, immersion phase change liquid cooling and other technologies, the computing power density of a single rack is increased by 20 times, and the PUE value is as low as 1.04.
Largest integration scale: ScaleX is the first domestic super cluster with 10,240 accelerator cards. Based on the high-speed network of ScaleFabric, it can achieve integrated deployment of 10,240 accelerator cards. The combined large cluster also reaches the largest scale of domestic computing clusters.
Most open: This cluster is designed entirely based on an open AI computing architecture, breaking the traditional closed-loop dependence of a single technology stack. It can support multiple brands of AI accelerator cards and is compatible with the mainstream AI computing software ecosystem, with extremely strong access capabilities to upstream and downstream industries.

The goal of building computing power networks is to break down the heterogeneous barriers of computing power infrastructure and solve problems such as uneven distribution and insufficient utilization of computing power resources in China. The scaleX 10,000-card super cluster, based on an open architecture, achieves the industry's largest single-unit computing power supply, undoubtedly providing a replicable, verifiable, and widely applicable model for the construction of computing power infrastructure.

Open computing architecture, based on broad compatibility, can scientifically combine different computing tasks to match the most suitable computing resources. On the one hand, it can optimize resource allocation and significantly improve overall computing efficiency and performance; on the other hand, it can also form standardized and commercial computing services, realizing the universal access to full-stack computing power.

The scaleX super cluster is an open computing resource pool that acts as a platform link between different AI computing powers and users. It also provides a computing platform that can be quickly deployed and launched for various AI application scenarios, supporting the rapid and seamless migration and in-depth optimization of AI models and business applications.

It is understood that after the core node goes online, it will become a computing power hub and scheduling center in central China, effectively linking green computing power resources in western China, undertaking the spillover of computing power demand from eastern China, and even undertaking the important mission of connecting east and west and radiating across the country.

Please, Log in or Register to view URLs content!

 

[tokenanalyst]

Panshi Technology completes RMB 100 million Series A financing to break the monopoly of maglev giants.​

Panshi Technology (Shenzhen) Co., Ltd. has secured a RMB 100 million Series A funding round led by Haitong Innovation, Shenzhen Capital Group, CMB International, and Shenzhen Guarantee Venture Capital, with Yushan and Cornerstone Capitals joining. This marks a major milestone in the development of domestic magnetic levitation technology.

In terms of innovation, Panshi Technology's core product, the magnetic levitation bearingless pump, replaces traditional mechanical bearings with magnetic levitation bearingless motor technology. The rotor rotates at high speed without contact with the pump casing, avoiding mechanical wear and contamination. Currently, it has taken the lead in overcoming the R&D and mass production challenges of the industry's highest power 6000W magnetic levitation bearingless pump, and has also completed the development and mass production of a full range of mainstream models from 30W to 6000W, meeting the customized needs of different application scenarios. Its products boast advantages such as ultra-cleanliness, frictionless operation, low shear force, and precise and stable flow, accurately adapting to the needs of applications in semiconductors, biomedicine, and fine chemicals. Its stable and reliable performance has gained widespread recognition from industry customers, breaking the long-standing monopoly of overseas giants.



With over 130 core patents and recognition as a National High-Tech Enterprise, "Landmark Product," and "Shenzhen Seed Unicorn Enterprise," Panshi Technology has gained strong market validation. Its products have passed certification by top semiconductor firms and are now widely deployed in critical cleanroom environments.

As global semiconductor equipment sales grow (China leading at $49.6B in 2024), domestic substitution remains urgent magnetic levitation pumps, as key components in wet process equipment, present a vast market opportunity. Panshi Technology aims to expand into more applications and become not just a supplier of core components but an enabler of high-quality, innovative industrial productivity in China’s advanced manufacturing sector.​

Please, Log in or Register to view URLs content!

Please, Log in or Register to view URLs content!

 

[dropout003]

Which SMEE machine is in action? (Video top left)

Please, Log in or Register to view URLs content!