Yang Guanglei, an adjunct professor at National Taiwan University who once served as director of TSMC's R&D department and later became a senior technical advisor to Intel, said that there are inherent problems with Intel's development of wafer foundry business. Therefore, it was originally believed that Intel's foundry business had a 50% chance of success. After later going to the United States to discuss and communicate with Intel, the chance of success was reduced to 1%.
After Yang Guanglei gave a speech at SEMICON Taiwan, he was interviewed by the media and expressed his views on Intel's development of foundry business. He said that the current competition between TSMC and Intel is not the issue of not being advanced in process technology, because Intel's process. The technology is actually quite good, but due to its inherent problems, even if it develops advanced process technology and has a good yield rate, there will be no customers to use it, making it difficult to succeed in developing wafer foundry business.
Yang Guanglei believes that the first essential problem with Intel is that it does not have a customer-centric engineering mindset. He gave the example of TSMC. When Qualcomm, a major mobile processor manufacturer, was developing the first generation of processors, Yang Guanglei himself flew NB to San Diego, where Qualcomm's US headquarters is located, and stayed there for a week to communicate with Qualcomm. Later, they later switched to wafers produced by TSMC at 65nm. On the other hand, Intel is technology-oriented and does not have the mindset of serving customers. It is up to you whether customers want to use it or not. Therefore, the things developed are difficult to use and customers will not be willing to use them.
In addition, Yang Guanglei believes that the difficulty for Intel to develop wafer foundry now lies in the environment. He pointed out that the current employment focus of young people in the United States is in software, and they hope to work at Microsoft, Google, and META. There are almost no advanced semiconductor R&D and manufacturing talents. For the United States, the semiconductor industry is an immigrant enterprise. In the 1980s and 1990s, the American semiconductor industry was supported by Indians and Taiwanese with oriental faces. Now, after these oriental-faced talents will go to their original countries, it has become quite difficult for the U.S. semiconductor industry to continue to develop at this stage.
As for the US government's launch of the Science and Chip Act, which will provide US$53 billion in funds to encourage companies to invest in the United States to develop the semiconductor industry, Intel will also benefit. Yang Guanglei is not optimistic that this will succeed. The reason is that US$53 billion in funds is not as much as TSMC's annual revenue, and such subsidies must be allocated to each investing semiconductor company, so the amount is even smaller. There is really nothing that can be done with such funds. What's more, if the subsidy is provided this year, will it continue next year? This makes it very difficult to develop an industry relying on such "tonics" to have a chance of success.
台積電前研發處長楊光磊：英特爾晶圓代工產業成功機率為 1%
曾經擔任台積電研發處處長、後來也成為英特爾資深技術顧問的台大兼任教授楊光磊表示，英特爾要發展晶圓代工業務，有其本質上的問題。因此，原本認為英特爾晶圓代工業務有 50% 成功機會，在後來前往美國與英特爾進行討論與溝通之後，再將這成功機率降低到 1%。 楊光磊在 SEMICON Taiwan 發表演講後...
finance.technews.tw
