|
After the Chinese state-owned company Tianjin Haiguang Advanced Technology Investment Co., Ltd. obtained the X86 technology license for US$293 million, the news that "well-known graphics semiconductor solution company AMD plans to sell its headquarters building in San Francisco for US$95 million" made netizens exclaim that AMD is "crazy because of poverty."
If the sale is successful, the office space will be drastically reduced by 30% to 45%, and there may be a series of cost-cutting measures such as layoffs. Will AMD suddenly become so poor that it has to sell buildings to keep the company going, like Nokia did in the past?
How poor is AMD?
According to U.S. stock data, AMD reported a loss of US$109 million, with net losses for six consecutive quarters. The data shows that AMD's operations have entered a vicious cycle in recent stages. So, can we judge that AMD has really entered a difficult stage at the moment?
Looking around, as far as the development of X86 architecture processors is concerned, AMD, as the only company in the world that can compete with Intel, has been completely at a disadvantage in the competition with Intel in the past decade. It has almost no share in the server market and less than 20% of the market share in the PC market. Moreover, Intel's performance last year showed that more than half of its profits came from the server market. After gaining a monopoly market share in the server market, it continued to increase the price of server processors to earn huge profits. Subsequently, taking advantage of the huge profits provided by the server market, Intel continued to reduce prices in the PC market, forcing AMD to follow suit. The price reduction of PC processors caused AMD's profits to become increasingly slim and led to continuous losses.
At the same time, AU has been declining since the excavator series. And since NV's Kepler, although the A card is not much worse in performance, there is a certain gap in energy efficiency, which makes it difficult to see AMD in the mobile market, and it is even more difficult to see performance in this market.
Looking back, in October 2015, AMD began to lay off employees in order to save money due to two consecutive quarters of losses in financial reports. However, the layoffs caused unrest among employees, and key technical personnel began to gradually seek better jobs. Last September, Jim Keller, the chief architect responsible for the Zen architecture, left, and Phil Rogers, the software architect who had worked for 21 years, also confirmed that he would leave and join AMD's competitor Nvidia to seek other development. From another perspective, the serious loss of talent is undoubtedly another major proof of AMD's financial crisis.
AMD has difficulty making progress in the server market
Currently, according to the analysis of market research firm Mercury Research, Intel, AMD and VIA dominate the X86 chip market, with Intel's market share of 87.7% being an absolute advantage, while AMD and VIA account for 12.1% and 1% respectively.
From various sources, we know that Intel is the world's largest semiconductor company. It has gained a leading edge over most competitors by continuously upgrading chip design and improving semiconductor manufacturing technology, which has enabled it to lead the development of global semiconductor manufacturing technology and win more than 90% of the server market share with this absolute advantage. With the dual advantages of PCs and servers, an ecosystem has been formed, which has been suppressing AMD in both the server and PC markets.
On the other hand, AMD's market share is far behind Intel, with a market share of about 1% in the server chip market. However, this has little impact on AMD, because AMD's market share is mainly in the PC market. However, this year, as Intel's ecosystem gradually improves, AMD is also unable to compete with Intel in the PC market, resulting in losses becoming commonplace. In 2015, the company had a revenue of US$3.99 billion and a huge loss of US$660 million. It was under such continuous losses that it began to split and sell its semiconductor manufacturing business in 2009. As losses continued, it even sold the shares held by the split company.
In order to achieve a comeback under such continuous losses, AMD needs to spend more money on the Zen architecture of PC chips, which it values very much, and it is regarded as a hope to counterattack Intel. In addition, the GPU business, one of its important businesses, needs to invest in research and development. However, its continuous losses make it difficult for it to have funds to take care of the research and development of ARM architecture server chips. In addition, the outflow of talents has gradually made a comeback a theoretical possibility.
Various problems have hindered AMD's attempt to enter the X86 server chip market. After careful consideration, AMD announced in 2012 that it would use the ARM architecture to launch server chips, but the loss of talent had a great impact on chip research and development, and the development progress was far behind other companies. When AppliedMicro launched an ARM architecture server chip in 2014, AMD did not launch the chip until the end of 2015. At the same time, Qualcomm, the leader in the mobile phone chip market, launched a 24-core ARM architecture server chip in 2015, and MediaTek, the second largest company in the mobile phone chip market, is said to be interested in entering this market. AMD is still facing severe competition.
China and AMD collaborate to achieve their respective goals
This year, due to the dual impact of information security and information technology blockade, my country has raised its information security issues to a new level. It hopes to develop its own server chips, but is also worried that foreign countries will use their own advantages to restrict China's own development. Today, after years of development, China already has chip companies such as Huawei HiSilicon, Spreadtrum, Ingenic, and Zhongsheng Hongxin, which have invested in the current global mainstream chip architectures ARM, MIPS, and POWER architectures and have achieved certain results. It is worth mentioning that Huawei HiSilicon is close to the mobile phone chip giant Qualcomm in technology research and development, which provides a good foundation for China to develop its own server chips.
However, the world's mainstream server market still needs the X86 architecture. Since the global server market has been built on the X86 architecture for decades, and the PC market is also dominated by the X86 architecture, the technical wall formed by X86 makes it quite difficult for non-X86 architectures to enter the PC and server markets.
According to rumors, both Amazon and Google were interested in adopting ARM architecture server chips, but no concrete action has been taken so far. China's Baidu also used ARM architecture server chips to build data centers, but because ARM's performance was too low and it was difficult to be compatible with the existing PC and server ecosystem, it finally chose to return to using X86 architecture servers.
China is currently the fastest growing server market in the world and one of the largest server markets in the world, which provides market space for China to develop server chips. China's existing chip companies are also developed based on non-X86 architectures, which makes it very likely that China will develop X86 architecture server chips. Qualcomm saw this opportunity and chose to cooperate with Guizhou to establish a server chip company in the hope of finding development opportunities for its own ARM architecture server chips.
However, as mentioned above, since the existing PC and server ecosystems are all based on the X86 architecture, even if China wants to develop a non-X86 architecture, it must gradually shift. Now AMD licensing its X86 technology to China for the development of server chips is undoubtedly the best choice at the moment.
Because AMD and Intel have a cross-licensing relationship in X86 technology, and Intel has a more obvious advantage in it, AMD added a restriction when cooperating with Tianjin Haiguang - Tianjin Haiguang can only use this technology to develop and sell server chips in China. For AMD, at a time when the server market is already difficult to make progress, cooperating with China in server chips can be said to be of great benefit without affecting its PC business. The most obvious benefit is undoubtedly that it will receive a licensing fee of US$293 million, and will continue to receive sales royalties from subsequent server chip sales. AMD's first quarter performance report released this year showed that its first quarter loss of US$109 million was reduced by 39.4% compared with US$180 million in the same period last year. It can be seen that the licensing fees obtained from the cooperation with Tianjin Haiguang played a considerable role.
How do they bypass Intel's technology cross-licensing?
So far, Intel has not made any statement on the cooperation between AMD and Hygon. However, netizens are quite concerned about "how they bypass Intel's technology cross-licensing?"
Gong Liming wrote on Zhihu: First of all, I want to give a thumbs up to the wisdom (buyaolian) of AMD and the Chinese government. It should be noted that most people think that x86 belongs to Intel and no one else can make it. In fact, this is not the case. The US patent protection is only 20 years. x86 appeared in 1978, and now it has been 20 years. In other words, it is completely possible to use current technology to remake the previous 386 and 486 chips without violating any patent protection laws. It's just that the performance is too poor. However, if you independently develop higher-performance chips, you will encounter Intel's strong patent protection anytime and anywhere.
Therefore, Intel's strong patent protection makes it almost impossible for other companies to make commercial x86. Only AMD and VIA (formerly Cyrix) are left. These two are old rivals that have been competing with Intel since the 386 era.
After years of technology accumulation, we have a rich technical foundation. Although our patents may not be stronger than Intel, we will reduce the chance of touching Intel's patent protection in chip development.
The fundamental difference between patent licensing and technology transfer is that patent licensing is actually providing IP cores, just like what ARM does. IP manufacturers provide you with a hard core and a bunch of technical documents. The design cannot be modified, and the design details are not disclosed. The patents still belong to AMD. Technology transfer directly tells you all the design details and sells you the patents. The ownership belongs to the buyer, who can freely use these patents for design without paying any more fees. Although this is a big detour in the procedure, it makes it possible for AMD to use patents in China. However, according to the exposed agreement, it emphasizes that chips are only sold in China. Because chip design is a high-cost business, there is almost no precedent for manufacturers to say that they will do chip research and development but will not sell them to the international market. Because there is no contract with exposed details, the editor cannot speculate on more useful information here.
Can AMD’s “transformation” be successful?
After years of development, AMD, as a super enterprise, seems to be developing well. However, if you are familiar with AMD's development trajectory, you will know that AMD has been actively adjusting its development direction due to corporate competition over the years, hoping to make a comeback at a certain stage. Especially after CEO Lisa Su took office, she established AMD's technology route. The first of the three major strategies she proposed is to create great products.
Moreover, Lisa Su has a deep affection for the Chinese market, and strengthening the relationship with Chinese partners is another core strategy that she is striving to build. Sugon, with which AMD has a close relationship in history, is naturally still an important partner of AMD. This time, AMD announced its cooperation with Hygon. Although Sugon has not officially released any information, everyone knows that the relationship between the two cannot be avoided.
AMD recently released its first quarter 2016 financial results, showing revenue of $832 million, slightly higher than the $818 million expected by Wall Street analysts and down 19% from $1.03 billion in the same period last year; net loss measured in accordance with U.S. generally accepted accounting principles was $109 million, compared with a net loss of $180 million in the same period last year.
AMD's revenue situation is improving, and the cooperation with Hygon has been disclosed, which led to a surge of nearly 22% in AMD's stock price after the market closed on April 22. The importance of the Chinese market to AMD is self-evident.
However, if AMD had entered the Chinese market earlier and sought cooperation, there might have been better development. Intel, which entered the Chinese market a few years ago, cooperated with Tsinghua University and Lanqi Technology; Qualcomm cooperated with Guizhou Province; IBM cooperated with many domestic manufacturers by introducing OpenPower. This time, AMD's strategy is still in line with the situation.
AMD has been deeply involved in China for nearly 30 years. In October 2015, AMD announced the establishment of a joint venture with Nantong Fujitsu to integrate its two packaging and testing plants in Penang, Malaysia and Suzhou, China to the latter.
In general, regarding the matter between AMD and Hygon, AMD kept a low profile because there may still be uncertainties in the authorization with Intel, while Sugon remained silent, which may be due to its relationship with Intel. It may also be that they have their own long-term plans for chips, and as a listed company, it is not yet time to announce it.
Also, think about it, the Chinese Academy of Sciences and Alibaba Cloud are actively developing quantum computers. So in the future, will related Chinese chip, network, architecture, software companies and other industrial parties work together to create a new industry and position?
Among them, international giants may just play supporting roles. As a winner of Toutiao's Qingyun Plan and Baijiahao's Bai+ Plan, the 2019 Baidu Digital Author of the Year, the Baijiahao's Most Popular Author in the Technology Field, the 2019 Sogou Technology and Culture Author, and the 2021 Baijiahao Quarterly Influential Creator, he has won many awards, including the 2013 Sohu Best Industry Media Person, the 2015 China New Media Entrepreneurship Competition Beijing Third Place, the 2015 Guangmang Experience Award, the 2015 China New Media Entrepreneurship Competition Finals Third Place, and the 2018 Baidu Dynamic Annual Powerful Celebrity. |
