Chinese chip companies still have a long way to go before entering the automotive supply chain on a large scale. But from a dynamic perspective, the participation of Chinese companies is increasing. It is only a matter of time before the proportion of Chinese companies’ automotive chip business equals or even overtakes that of the vehicle business.
Article/"Automobile Man" Huang Yaopeng
In the consumer market, automotive chips receive less attention than other automotive parts. The former are usually not within the scope of replacement after-sales maintenance. But the "core shortage" of the public in the north and south has appropriately ignited public interest. Enterprises worth hundreds of billions will also be forced to stop production by gadgets such as chips.
After Huawei was sanctioned, chips once again came into the public eye. But this time, it's just a partial application of the chip.
Already lagging behind
In 2019, the global semiconductor business scale was approximately US$500 billion, and automotive chips accounted for US$47.5 billion, accounting for less than 10%. However, as the "new four modernizations" become a definite trend, software and chips have quickly become the value-added part of automobiles. Starting from 2021, the annual increase will be at least 20%.
However, the scale of China's automotive chip industry is less than 15 billion yuan, accounting for 4.5% of the global share; while China's complete vehicles account for 30% of the global business. In contrast, automotive chips lag behind the entire vehicle business in terms of scale, technology, brand and category.
In general, Chinese chip companies still have a long way to go before entering the automotive supply chain on a large scale. But the situation here is different.
If roughly divided, automotive chips can be roughly divided into three categories: the first category is responsible for computing power, specifically processor and controller chips, such as central control, ADAS and autonomous driving systems, as well as engines, Chassis and body control, etc.; the second category is responsible for power conversion, used for power supply and interfaces, and IGBT power chips for EVs fall into this category; the third category is sensors, used for various radars, airbags, and tire pressure detection. kind. Of course, there are some miscellaneous items, which are ignored here.
From a global perspective, four European companies, NXP, Infineon, STMicroelectronics, and Bosch, account for 36.2% of the share; among U.S. companies, TI and ADI (the former acquired the latter, Currently one company) accounts for 15.7%; Japanese brands Renesas and Toshiba account for 14.3%. The above eight major suppliers account for more than 2/3 of the market share.
In reality, some Chinese companies have already figured out how to enter the supply chain. Driven by broad social awareness and huge investment, it is only a matter of time before the proportion of Chinese companies' automotive chip business equals or even overtakes that of the vehicle business.
Demand traction
As mentioned earlier, it is difficult for Chinese companies to enter the supply chain. The technical threshold is not an isolated factor. Because of the abnormal stability requirements (0PPM) of automotive-grade chips, Tier1 suppliers responsible for purchasing chips do not want new players to practice. If you are not an established company (in fact, the rise of the chip industry happened after the 1980s), you will not even have the opportunity to be a backup tire.
The problem is that if you don’t come in, it will be impossible to iterate on technology and keep up with technological development. The more you can't keep up, the harder it is to get in. It's become an endless cycle.
How to break the situation? BYD provides an idea, that is, demand traction. Aren’t OEMs and Tier 1 suppliers not allowed to enter? BYD itself is a vehicle manufacturer and Tier 1 supplier. If it acts as a customer and places orders for itself, wouldn't the demand problem be solved?
In the cost structure of new energy, the second highest is IGBT. IGBT, like other chips, is cut and etched from the wafer. BYD's current production capacity is 50,000 units per month, and it will double its production capacity next year to meet the demand for 1.2 million new energy vehicles. BYD has invested 1 billion yuan in Changsha wafer production line and has started construction, with an annual output of 250,000 8-inch wafers. Like batteries, in addition to producing and selling its own products, BYD also plans to become a supplier to other car companies.
The big players in IGBT are Infineon, Fuji Electric, ON Semiconductor and Mitsubishi. In 2019, Infineon supplied 628,000 sets of IGBT modules in China, with a market share of 58%; BYD supplied 194,000 sets, with a market share of 18%.
BYD uses independently developed and produced IGBTs on many models such as Qin and Tang, but it was not announced until 2018. By the end of 2018, BYD had 175 patents on IGBTs.
However, in the field of automotive 600-1200V applications, Infineon has an overwhelming advantage; in Japan, Mitsubishi and Fuji Electric divide the market, and the IGBTs used in Toyota hybrids are all completed in-house.
China's Star Semiconductor (listed company) is more advanced in IGBT technology and has developed to the sixth generation (BYD is the fifth generation). Last year, Star supplied 17,000 units, with a market share of 1.6%, far less than BYD, which shows the traction effect of the OEM.
From a global perspective, IGBT has developed to the 7.5th generation (Mitsubishi), and there is a big gap between Chinese companies, but this is already a subfield in which Chinese companies have been relatively successful in the chip business. As long as the demand side continues to exist, it will only be a matter of time before we catch up with the advanced level.
A new track
Among automotive chips, computing chips are considered a “new track”. With the rise of the concept of smart cars, the demand for on-board computing power has surged. New demands promote new investments, and old revolutions encounter new problems.
The eight major suppliers of automotive chips have no outstanding advantages in this field. In this sub-field, there is no monopoly brand that dominates the sky, and all manufacturers are in the testing stage. Chip companies such as Samsung, Nvidia, Intel, Xilinx, and Qualcomm are all involved in the field of vehicle computing power.
Traditionally, OEMs do not directly purchase chips, but computing power chips are an exception. Tesla, NIO, Xpeng, and Uber (which just sold their ADAS business) have all invested in the research and development of their own computing chip business. Of course, they are at best chip designers and must find another foundry.
Even so, this also breaks the dilemma of host manufacturers only providing requirements and mission statements. General-purpose CPU control chips cannot meet the needs of ADAS, prompting OEMs (mostly new forces) to design dedicated chips for autonomous driving.
Although OEMs bear risks when developing their own chips, investors like to hear this kind of news. Moreover, OEMs do not want a large part of profits to fall into the pockets of chip suppliers in the future.
However, similar investments need to be used on a large scale to be profitable. If US$150 million is spent on the design of a dedicated ASIC chip and the annual output is 400,000 vehicles, the investment will be recovered within 4 years. At present, only Tesla is expected to reach this passing line.
The remaining opportunities belong to suppliers. Although Huawei has repeatedly reiterated that it will not build cars, it will not reassure OEMs because it has always been a provider of a package of solutions: cloud, 5G infrastructure, Internet of Things, vehicle operating systems, and V2X communication chips. If you add in For computing power chips, Huawei will have enough ambition to become a new Tier1 supplier.
As early as October 2018, Huawei released the computing platform MDC6000 that supports L4 autonomous driving capabilities. The platform has eight Ascend 310 AI chips launched by Huawei, and also includes computing and image processing modules. Audi China becomes Huawei's first important customer for ADAS business.
Huawei claims that the MDC6000 has a computing power of 352 Tops and a power consumption ratio of 1 Tops/Watt, surpassing Nvidia.
What needs to be clarified is that supporting L4 autonomous driving does not mean achieving L4 autonomous driving. Huawei only provides the computing platform, and software implementation is completed by customers or designated third parties. Moreover, vehicle-level certification also takes a long time.
Tesla’s FSD has a computing power of 144 Tops, the Nvidia used by Xpeng has a computing power of 30 Tops, and the Mobileye Q4 used by Weilai and Ideal has a computing power of 2.5 Tops. This means that OEMs currently implementing L2 or L2.5 autonomous driving have a more realistic choice instead of using Huawei’s extravagant computing power and power consumption. Only when there is a realistic solution for L4 autonomous driving can Huawei really get involved.
Therefore, in the field of automotive computing chips, Huawei is targeting the future, not the present. The products of start-up companies such as Horizon, Black Sesame, Cambrian, Shenjian Technology, and Feibu Technology have been applied to independent vehicle products. In the field of vehicle computing power, Chinese companies are making partial breakthroughs. Of course, multinational Tier 1 companies such as Bosch have complete ADAS solutions. In the long run, Huawei's "omnidirectional" solution may be more suitable for future comprehensive requirements for the Internet of Vehicles, vehicle machines, new energy and ADAS.
Traditional Power
In the field of functional chips in computing chips, such as MCU (microcontrol unit, especially fuel vehicle power system control), and sensor chips, traditional giants still Firmly held, Chinese companies also have similar products, but the participation rate is very low, only about 4.5%.
There are no signs of breakthrough in this area for the time being. We can only start from small orders and spare tires, hone the level of vehicle-level stability, and gradually accumulate the capital to challenge the giants. However, Chinese companies are actually in a dilemma as to whether they should continue to invest in functional MCUs related to fuel vehicles. However, some technical reserves can still be used in the new energy era for long-term applications, such as the ESP (body stability control) chip that is out of stock this time. It is too crude to believe that the traditional MCU chip market is shrinking.
From a dynamic perspective, the situation is less worrying, because the participation of Chinese chip companies is now increasing. The "new four modernizations" wave is inherently friendly to new players, and Chinese companies cannot waste historical opportunities. (Text/Huang Yaopeng of "Autobot") Copyright Statement This article is an exclusive original manuscript of "Autobot", and the copyright belongs to "Autobot".
This article comes from the author of Autohome Chejiahao and does not represent the views and positions of Autohome.