Monday, April 17, 2017
P>The global reaction to the aggressive investment strategy adopted by China-based Tsinghua Unigroup in the semiconductor industry is mixed, and the role being played by its subsidiairy Yangtze River Storage Technology is also unclear to people outside China.
Digitimes recently conducted an exclusive interview with Charles Kau, former chairman of Inotera Memories, who joined Tsinghua Unigroup over a year ago and now serves as executive VP at the company as well as acting chairman of Yangtze River Storage, to hear his own account on why he decided to join Tsinghua Unigroup, and the goals he wants to accomplish at the memory storage company.
Q: What prompted you make the decision to join Tsinghua Unigroup?
A: I was not acquainted with Tsinghua Unigroup chairman Zhao Weigou previously. He approached me in 2015 when he planned to acquire Micron Technology in order to build up a national-scale memory company in China. We hit it off on our ideas on how to develop the storage industry, and we also shared a consensus of the need to reach a point where we could compete effectively with Samsung Electronics in the memory market.
Samsung's domination in the global memory chip market is in fact a global issue, and it appears that China can serve as an emerging force to balance the power of Samsung. As a result, I decided to join Tsinghua Unigroup, with an aim to help build up China's memory industry.
China's National Integrated Circuit Industry Investment Fund originally aimed to put together resources to support one company to implement the memory-chip investment project. In addition to Tsinghua Unigroup, the Hubei Provincial Government and Wuhan Xinxin Semiconductor Manufacturing Corporation (XMC) were also interesting in undertaking the project. However, China's policy makers wanted people with different ideas on how to run the memory business to join the project. Through matching by the national investment fund, Tsinghua Unigroup and XMC agreed to merge, while I was invited to execute a project to build up a complete lineup of NAND flash technology for enabling mass production.
Q: What role does Yangtze River Storage Technology play in the memory-chip project, and what is its current status?
A: Yangtze River Storage was officially set up on July 16, 2016, and immediately acquired a 100% stake in XMC, making it a development center for memory products. Tsinghua Unigroup currently owns 51% of Yangtze River Storage, with the national investment fund taking another 25%, and a number of concerns from Hubei took 24%.
XMC currently has a total of 1,200 employees, while Yangtze River Storage has another 700, of which 500 are R&D personnel, including 50 from Taiwan.
In terms of manufacturing capacity, Yangtze River Storage has two 12-inch fabs locating in Wuhan and Nanjing, respectively, each having a production capacity of 300,000 wafers a month.
Yangtze River Storage will begin to sample 32-layer 3D NAND chips at the end of 2017, signifying a milestone in the development of related technologies in-house. Then, we will leapfrog to 64-layer which will substantially ramp up our production. We believe that the pace of development of 3D NAND chips will then slow down, enabling Yangtze River Storage to narrow its technology gap against leading players, including Samsung, SK Hynix and Toshiba, to one generation.
Q: People outside China have been skeptical about your company's plans to build foundry houses, wondering if the moves are just blind investments or simply fluffing?
A: The 12-inch fabs in Wuhan and Nanjing were planned earlier, and Yangtze River Storage will use these two fabs to fabricate memory chips. There should be no more fabs for memory.
We absolutely will not conduct blind investments and will not enter mass production until our technology is mature and competitive. We will not rush into volume production of memory chips. Yangtze River Storage will definitely develop relevant technologies in-house, and the rooting of technology should be done step-by-step.
Q: What is Yangtze River Storage's relationship with Micron?
A: While we aim to develop related 3D NAND and DRAM technologies in-house, we still maintain a certain level of flexibility and welcome cooperating partners to join us. I personally believe that Micron is the best partner for Yangtze River Storage to develop DRAM and 3D NAND chips. But now the political atmosphere is unfavorable to us, as the US is worried about the threat of the rise of China's semiconductor industry, and therefore our talks with Micron for cooperation will be prolonged.
Given that Korea currently accounts for 80% of the global DRAM market and 60% of the NAND flash market, I am worried that Korea-based companies will take control of 90% of the DRAM market in the short run, and that is terrible.
On the other hand, Micron has seen its market share decline to 18% from 28%, and its power to counter Korea-based vendors is diminishing. Cooperation between Micron and Yangtze River Storage will be the best partnership, as we have the support of 12-inch fabs.
Q: Why does Yangtze River Storage insist on developing memory technology in-house?
A: If Yangtze River Storage decides to produce DRAM or NAND flash chips under licensing agreements, it will repeat the failure and painful process that Taiwan endured previously on the development of its DRAM industry. Taiwan failed to develop its own DRAM industry because a total of 7-8 companies were fighting over a pool if insufficient resources. China needs to concentrate its resources to succeed, while also building up its own technology know-how. We definitely will not steal related technologies from others.
Copyright © 2017 CST, Inc. All Rights Reserved