AMAT VP: "Data is like crude oil in new tech era"..............................................................
"economic benefits generated by AI and data applications to grow exponentially"
"capital expense intensity will increase along with the growing complexity of chip design and fabrication."
"He went on to say that 25 times more of NAND flash memory, compared with the current output, would be demanded if the penetration of Level 4 smart cars reaches 1% by 2020 (just one of many bit intense NBTs)"
New materials and techniques:
"He concluded that as semiconductor process is advancing to 7nm, Applied Materials is ready to apply its cobalt material evolutions to help significantly refine metal deposition process, making it possible to achieve high-performance computing at low power consumption in the future"
slime street rotation out of Tech roundtrip contort n distort?
Facts by RegFD VP.
Data is like crude oil in new tech era, says Applied Materials VP Erix Yu
Rebecca Kuo, Tainan; Willis Ke, DIGITIMES
Wednesday 27 December 2017
After progressing from the PC and network generation in 2000-2010 to the mobile device and social networking generation during 2010-2020, the ever-advancing tech development will enter a new generation of AI (artificial intelligence) and big data from 2020 on, with economic benefits generated by AI and data applications to grow exponentially, according to Erix Yu, group vice president of Applied Materials and president of Applied Materials Taiwan.
Yu, while speaking to Digitimes in a recent interview, likened data to crude oil. He said data must undergo further processing before they can be utilized to generate tremendous economic values regardless of the sources they are collected from, just like underground oil that has to be exploited, extracted and refined into fuel oil or plastic materials to generate huge rewards.
In the past, semiconductor technologies, as well as hardware and software support, could not handle massive data processing and computing. But constant semiconductor technological advancements have now enabled storage of massive data and boosted computing performance, helping realize previously unachieveable concepts.
Such advancements will enable AI, big data and even VR/AR applications to become the next wave of growth drivers, which will combine to bring drastic changes to human life and subvert the development of industries, Yu indicated.
In return, semiconductor technologies will also get a significant boost along with the advent of the AI and big data era, Yu said, reasoning that tremendous demand for data storage will drive rapid growth of the NAND flash market and AI computing will enable the wafer foundry market to sustain significant growth momentum.
Five major growth drivers
Yu disclosed that Applied Materials has identified five major business growth drivers in the AI and big data era. First, demand for 3D NAND memory chips for data storage will increase sharply. Second, AI and high-performance computing (HPC) chips will drive the growth of the wafer foundry sector. Third, patterning technology and material engineering will play an important role in securing semiconductor process development going in line with Moore's Law. Fourth, advanced display technology will lead to growing adoption of OLED panels for small-size mobile devices, while the increasing popularity of large-size TVs will power the development of 10.5G TFT-LCD panels. Fifth, China is fully gearing up to establish new panel and semiconductor fabs, which will greatly fuel related investments. All these have also been serving to support important developments of Applied Materials in recent years, Yu disclosed.
AI technologies and applications are expected to thoroughly alter the industrial and commercial sectors in the coming years, mainly driven by visual computing and AI, including AR/VR, machine learning, autonomous vehicles, big data, IoT and personalized medicine, according to Yu.
In addition, the semiconductor equipment market is expected to see sustained growth in the coming years, as capital expense intensity will increase along with the growing complexity of chip design and fabrication. For instance, Yu said, in terms of NAND production, advancing from 2D planar NAND to 3D 64-layer NAND will require a 60% increase in equipment investment, compared to a corresponding 40% increase for advancing DRAM production to 14-16nm from 25nm. For wafer foundry, making advanced 7nm chips is twice as expensive as making 28nm chips, and the smaller the nm size, the more challenging and expensive the development.
The same situation is also seen in investment in display panel production equipment. For the same 6G plant, the cost will shoot up 425% if the plant is turned into production of OLED panels from LCD.
Exploring immense business opportunities
In other words, investment required for developing advanced technologies will be much higher than before. Nevertheless, tech players are still keen on spending, mainly because of the immense business opportunities associated with the application of unimaginably huge data to either IoT or Industry 4.0 operations in the new era of AI and big data, Yu explained.
Citing an example to support his explanation, Yu said that a smart city with a population of one million will generate up to 200 million GB in daily data volume by 2020, including data collected from smart buildings, smart plants, public safety, smart cars, smart airplanes and social networks, among others. How to collect and process these data and where to store them involve extremely huge business opportunities, Yu stressed.
He went on to say that 25 times more of NAND flash memory, compared with the current output, would be demanded if the penetration of Level 4 smart cars reaches 1% by 2020, as the data generation and the requirement to compute and store those data will be incredible under those circumstances, with the daily volume of data to be generated by each car estimated to reach 4TB.
But no companies can afford the huge expenses on building 25 times more NAND flash fabs, and the best alternative is to improve fabrication process to meet demand for future production capacity, Yu opined, adding that innovative computing solutions are badly needed to support the process improvement.
Meanwhile, big data will bring more innovations into cloud and terminal devices, and material engineering innovations will play a crucial role in this aspect. Yu indicated that data will become new oil in the AI era, but this barrel of oil must integrate accurate and brand-new computing solutions before turning into a barrel of gold.
He concluded that as semiconductor process is advancing to 7nm, Applied Materials is ready to apply its cobalt material evolutions to help significantly refine metal deposition process, making it possible to achieve high-performance computing at low power consumption in the future