OFweek electronic engineering network news in the 5/24-25 held in the Samsung Foundry Forum Event released a number of technology roadmap has a look at the next little news: Samsung's 28nm FD-SOI (Fully Depleted Silicon on Insulator) process RF (Radio Frequency) and eMRAM (Embedded Magnetic Random Access Memory) options where the risk production of eMRAM is 2018; 18nm FD-SOI process risk production will start in 2020, as well as RF and eMRAM Options.
At the Taiwan Semiconductor Manufacturing Technology Forum that was held in the United States in March, TSMC announced that in the second half of 2017, the eFlash (embedded flash memory) will be replaced by eReRAM (embedded resistance random access memory) in the 40nm process; under 2018 Half a year in the 28nm process with eMRAM instead of eFlash and eDRAM. Incidentally, in September 2016, GlobalFoundries also released news that it plans to provide eMRAM on 22nm FD-SOI process in 2017 and mass production in 2018.
Why does the wafer industry talk about the traditional process of 28nm or even 40nm when talking about the advanced weapons arms race below 10nm, even extending to 3, 4nm, and the ultimate weapons such as EUV (Extreme Ultra Violet)? The reason is that current IC accumulation is high and various functions - such as memory - must be integrated into a single IC. During the evolution of the process, some traditional embedded memories cannot be smoothly scaled back, or the performance cannot meet the emerging applications. Therefore, new embedded memories must be introduced to meet the requirements of continuous miniaturization and performance. In the wafer-generation industry, four rapidly growing areas, mobile, high performance computing (HPC), automotive, and Internet of Things (IoT), automotive and IoT MCU manufacturing processes are entering For these generations, the former required high-speed embedded memories, the latter required low-power embedded memories, and the emerging embedded memories were developed with the goal of satisfying these demands simultaneously. In particular, the Internet of Things (IoT) application, estimated by CAGR at 25%, is the fastest growing of all applications and is the battleground of the foundry industry.
Although the market has a strong demand for eMRAM technology, the time horizon announced by various foundries is somewhat delayed compared to the one or two years ago, and there is uncertainty. The main reason is that eMRAM is a technology other than the process development miniaturization research and development spindle, and it is not a subject of conventional research and development. On the other hand, wafer foundry is used to create a service industry and habits are led by customers. The development of forward-looking and big turns has always been somewhat hesitate.
Those who are familiar with the industry may give counter-examples: Has Samsung not long ago acquired the Grandis, which has developed MRAM as its core business, in 2011? Yes, but the main purpose of the initial merger and acquisition was for its memory division. The initial design after the merger was a DRAM-like product, although it was ultimately unsuccessful. It is precisely from this example that the differences between R&D thinking patterns of product companies and manufacturing service companies in the past can be seen. But the environment of the industry has changed, and ideas have to be changed. The unconventional competition axis of eMRAM gives us the opportunity to observe the changes in the industrial competition strategy.
4000Puffs Disposable Vape,Elf Bar Bc4000 Puffs Vape,4000Puffs E-Cigarette Vape Pod,Elf Bar Bc4000 Puffs Disposable Vape
Shenzhen Uscool Technology Co., Ltd , https://www.uscoolvape.com