If you want to know the forefront of silicon-based semiconductors, you need to pay attention to TSMC's 5nm/3nm fab in Taiwan. If you want to know how the current third-generation semiconductor game is progressing, you can turn your attention to the wafer factory of Wolfspeed, the global leader in SiC substrates, located in New York, USA.
The substrate material, which is technically difficult and accounts for half of the device cost, is regarded as a key link in the silicon carbide industry chain. There is also a saying in the industry that "whoever has the substrate wins the world".
Domestic substitution: there is noise in the technical gap, and some projects are unfinished without suspense
The international leading 8-inch mass production is imminent, and the revenue is about to be realized. How is the progress of the domestic industry?
At present, all aspects of domestic silicon carbide have achieved the layout of the entire industry chain, and many representative enterprises have emerged. Substrate manufacturers include Tianyue Advanced, Tianke Heda, etc.; epitaxy manufacturers include Han Tiancheng, Dongguan Tianyu, etc.; design manufacturers include Shanghai Zhanxin Electronics, Shanghai Hanxin, etc.; IDM manufacturers include Sanan Optoelectronics, Times Electric, China Resources Micro, Silan Micro, etc.
The third-generation semiconductors have been given strategic significance and have quickly become a super outlet after being included in the 14th Five-Year Plan in 2021. However, even when the industry is beginning to take shape, there are still many noises when it comes to issues such as how big the gap between three and a half generations of domestic and foreign companies is. For example, it is often heard that the start-up time and technology gap between domestic and foreign three-and-a-half-generation enterprises are not large. What is particularly important is that some actions based on misunderstanding have sowed hidden worries for the industry.
In terms of starting time, the global leader in substrates has accumulated for more than 30 years. Infineon, the leader in power devices, has started research on silicon carbide since 1992, and became the first manufacturer in the world to launch silicon carbide diodes in 2001. Commercial. This also means that European, Japanese and American manufacturers have accumulated more than 30 years of patents.
In terms of technology gap, the overall gap at home and abroad is 5 to 8 years, and the most obvious indicators are process and cost.
The current domestic silicon carbide substrate is mainly 4 inches, and the international first-tier manufacturers use 6 inches as the mainstream. Many manufacturers can already supply 6 inches stably. A chip with a diameter of 4 inches can make 1,000 chips at a time, while a wafer with a diameter of 6 inches can make 3,000 chips at a time. It is difficult for a 4-inch product to be competitive.
Although domestic companies have made breakthroughs in the 6-inch field, the yield rate is difficult. Wolfspeed first demonstrated 6-inch silicon carbide substrates in 2015, and the yield reached 65% in 2017.
An industry insider revealed to "Ask Core Voice" that the yield rate of 6-inch substrates for domestic leading companies is 50% to 60%. Some manufacturers have a yield rate of 30%, and some manufacturers are not included in the statistics because the yield rate is too low. In terms of devices, there is a generational difference between the products learned by domestic companies and those being sold by first-tier manufacturers.
The three-and-a-half-generation concept continues to heat up, and the investment fever remains high. Respondents told "Ask Core Voice" that "there are so many new projects that it makes people numb" and "a new project starts every few days in three and a half generations".
In the secondary market, there is no shortage of companies such as Roshow Technology that have been chasing hot spots for a long time. After industrial hemp and sapphire, they have begun to bet on the third-generation semiconductors. From time to time, there are voices worrying about "three and a half generations of big refining" in the industry, and industry players often cite the blooming industrial parks and the panel industry with excess capacity.
Yu Kunshan, secretary general of the third-generation semiconductor industry technology development strategic alliance, publicly stated that while encouraging local governments and social capital to actively participate in the third-generation semiconductor industrialization, the high risk, high investment, and high cost of the semiconductor industry must be taken into account. Technology threshold, long cycle and other characteristics, and orderly promote the industrial layout.
Although there are many voices of early warning, the third-generation semiconductor industry, which has received a lot of policy resources, has already released the danger signal of low-level and repeated construction. According to domestic media statistics, the total investment in silicon carbide projects publicly reported across the country in the past ten years has exceeded 130 billion yuan.
An analyst who has been following three generations of domestic semi-industries for a long time told "Ask Core Voice" that some production lines may not be completed at all, and at least half of them are unfinished in the end. Some are because the source of the technology is not clear, some have problems with the equity in the middle, and some may be a scam at the beginning. He also pointed out that some companies that have been involved in three and a half generations are unheard of, and many production lines built in second- and third-tier cities will easily fail.
The analyst said that the overall investment threshold for silicon carbide projects is lower than that of silicon-based semiconductors, which is the reason why the future production rate is not high and the project is more likely to be unfinished. On the whole, the current domestic investment in new projects ranges from hundreds of millions to 10 billion, and the projects with the highest attention are concentrated in the scale of 10 billion.
Electric vehicles, charging piles, new energy power generation and other fields require the use of third-generation semiconductor technology. Yole predicts that the size of the silicon carbide SiC market is expected to exceed US$2.5 billion in 2025, with a compound annual growth rate of more than 30%.
According to TrendForce Consulting Research, the output value of third-generation power semiconductors will grow from $980 million in 2021 to $4.71 billion in 2025, with a compound annual growth rate of 48%.
Focusing on market prospects, major substrate manufacturers such as Wolfspeed and II-VI are gradually expanding their production capacity. In addition to production expansion, process conversion, mergers and acquisitions, and vertical integration have also become highly consistent actions by major substrate manufacturers.