Chinese semiconductor thread II

antiterror13

Brigadier

BYD unveils its self designed Xuanji-A3 ADAS chip + all the other 566 chips that it has developed. Most of them, it not only designs, but fabs them and do the packaging + testing.

TSMC is likely the foundry partner until BYD is banned by the bully. In fact there is a good chance will be banned as BYD is beating Tesla in most areas
Next year, there is a good possibility that SMIC is able to make 4nm, but probably 100% fully booked already
 

tokenanalyst

Lieutenant General
Registered Member

Beijing CETC Delivers Multiple 12-inch High-End Thinning Equipment to Industry Leader​


Several 12-inch fully automated ingot and wafer thinning equipment independently developed by the company for SiC substrate materials were successfully shipped and delivered to leading companies in the industry. This marks another high recognition of the company's technical strength, product stability and market service capabilities in the field of high-end semiconductor thinning equipment by top customers in the industry, and further consolidates the company's leading position in the domestic thinning equipment market.

The 12-inch fully automated SiC ingot and wafer thinning equipment is one of the key core process equipment in the semiconductor material processing industry chain. The industry has high technical barriers, stringent process standards, and significant R&D challenges. The two types of equipment delivered this time address different process difficulties in ingot and wafer thinning, achieving key technological breakthroughs : the ingot thinning machine innovatively adopts an automated gripping and adsorption dual-mode conveying system, ensuring stable and efficient transport of large-size ingots, significantly shortening the processing cycle, and adapting to the needs of large-scale mass production; the wafer thinning machine integrates independently developed ultra-precision air spindles and air-bearing wafer support stages, achieving internationally leading levels in indicators such as intra-wafer thickness deviation and local thickness deviation.

1780000956656.png

Please, Log in or Register to view URLs content!
 

tphuang

General
Staff member
Super Moderator
VIP Professional
Registered Member

my thread on how SRAM scaling stopped at 5nm (or around there) and since SMIC is also close to that point, it actually can continue to get SRAM clock speed & energy improvement from 3D folding. In fact, it can leap ahead of TSMC 2D SRAM here. And this is quite consequential to any application needing large L1/L2 Cache. I use Groq as an example.
 

tokenanalyst

Lieutenant General
Registered Member

The Zhicheng Semiconductor Wet Process Equipment Manufacturing Headquarters Project, with a total investment of 1.2 billion yuan, was signed.​


The signing ceremony for the Suzhou Zhicheng Semiconductor Wet Process Equipment Manufacturing Headquarters project was held.

It is reported that the Zhicheng Semiconductor Wet Process Equipment Manufacturing Headquarters project signed this time has a total investment of 1.2 billion yuan and a total area of 45 mu. After it is fully completed and put into operation, it will have an annual production capacity of more than 200 semiconductor wet process equipment and is expected to achieve an annual operating income of 3 billion yuan.

1780006575253.png

According to available information, Zhicheng Semiconductor has long been deeply involved in the field of semiconductor wet process equipment. Its products are widely used in core areas such as silicon-based semiconductor manufacturing, advanced packaging, and compound semiconductor manufacturing, comprehensively covering various semiconductor wet process scenarios such as front-end processes, large silicon wafer processing, and specialty processes.1780005913530.png
1780005899216.png
Please, Log in or Register to view URLs content!
Please, Log in or Register to view URLs content!
 

tphuang

General
Staff member
Super Moderator
VIP Professional
Registered Member
TSMC is likely the foundry partner until BYD is banned by the bully. In fact there is a good chance will be banned as BYD is beating Tesla in most areas
Next year, there is a good possibility that SMIC is able to make 4nm, but probably 100% fully booked already
do you understand how many Chinese chip designers in the auto sector use either TSMC or Samsung as their foundry and there hasn't been any complaint about it? ADAS is just not a big deal to America.
 

tokenanalyst

Lieutenant General
Registered Member

Shenzhen Technology plans to invest 1.47 billion yuan to expand its high-end memory chip packaging and testing capacity.​


Shenzhen Technology issued an evening announcement stating that its wholly-owned subsidiary, Perton Technology (Shenzhen) Co., Ltd., and its holding subsidiary, Hefei Perton Storage Technology Co., Ltd., plan to implement a high-end storage chip packaging and testing capacity expansion project.

This capacity expansion aims to meet the growing demand in the high-end memory chip packaging and testing market, deepen cooperation with strategic customers, and break through existing capacity bottlenecks.

The announcement shows that the project plans a total investment of 1.47 billion yuan, which will be used to purchase high-end chip testing machines, high-precision wafer grinding machines and other equipment, factory decoration and supporting power facilities. After the project is completed and reaches full production capacity, Shenzhen Peiton is expected to increase its monthly packaging capacity by 5 million chips and its testing capacity by 8 million chips; Hefei Peiton Storage is expected to increase its monthly packaging capacity by 28.8 million chips.​

Please, Log in or Register to view URLs content!
 

tokenanalyst

Lieutenant General
Registered Member

JCET Group: 3D heterogeneous integration reshapes the boundaries of advanced packaging innovation​


Driven by rapid advancements in artificial intelligence, high-performance computing, and 6G communications, JCET Group is positioning 3D heterogeneous integration and Chiplet technologies as essential pathways to overcome traditional computing bottlenecks. As modern systems increasingly combine components from different wafer fabs, process nodes, and functional domains, the industry faces a critical challenge: unifying fragmented design data and ensuring seamless compatibility across diverse manufacturing ecosystems. Recognizing that success in this space fundamentally depends on supply chain collaboration, JCET emphasizes breaking down cross-fab data barriers to enable true system-level performance optimization.

To tackle these complexities, JCET has developed an integrated technical framework built on data standardization, multiphysics simulation, and intelligent design automation. The company implements a three-step strategy that creates a unified data middleware layer, ensures cross-platform PDK interoperability across EDA toolchains, and establishes closed-loop DTCO/STCO processes to balance performance, signal/power integrity, thermal stress, and reliability. This is reinforced by a novel cross-scale collaborative simulation system that models physical phenomena from global wafer deformation down to submicron TSV stress concentrations, enabling early risk mitigation through a “left-hand, front-end” reliability methodology. Additionally, JCET is partnering with leading EDA vendors to embed AI as an intelligent co-pilot, streamlining the design and optimization of highly complex 3D architectures.

Looking toward the future, JCET envisions System on Wafer (SoW) as a transformative architecture that consolidates multiple functional units onto a single wafer to deliver shorter interconnects, higher bandwidth, lower latency, and unprecedented compute density for AI model training and scientific computing. Heterogeneous integration is rapidly redefining semiconductor packaging by bridging cross-fab data collaboration, multiphysics coupling, and AI-driven design into a cohesive system innovation strategy. Through an open, collaborative approach, JCET will continue advancing its full-chain service platform spanning design, simulation, manufacturing, and verification, while actively working with upstream and downstream partners to build a robust heterogeneous integration ecosystem that accelerates the development of next-generation intelligent applications.
Please, Log in or Register to view URLs content!
 
Top