This article was reproduced from the autocarweekly public account

Author: Financial Street Laoli

Many automakers and financial peers have reached a consensus: “Automotive chips will continue to be a hot topic within 5 years.” At the end of last year, after media exposed Volkswagen’s “chip shortage”, industry attention around the chip shortage continued to escalate. First, a power outage caused a Samsung factory to stop production, and then Changan Automobile delivered nearly 80,000 fewer cars in the first quarter due to the “chip shortage”. NIO also announced production adjustments and capacity reductions last week. Today, let’s talk about why global automakers lack chips, where the gaps in Chinese automotive chips are, and when automakers will develop their own chips.

Why global automakers lack chips

When it comes to chip shortages, people think of the US-China trade confrontation and the US’s chip sanctions against Huawei, but this shortage of chips for automakers has nothing to do with US sanctions. The chip shortage crisis is a global issue, not just for Chinese automakers, but also for foreign companies such as Volkswagen and General Motors. Even Toyota and Honda, which have always placed supply chain security at the core of their business, have fallen victim to chip shortages. The reasons for the shortage of chips are complex, but overall, the main causes are insufficient upstream wafer capacity, mismatch between supply and demand, and natural disasters.

In the semiconductor industry, the manufacture of chips must rely on wafers, and the applications for 8-inch and 12-inch wafers are the most extensive. 8-inch wafers are generally used in automobiles, but for upstream wafer manufacturers, the return on investment for building 8-inch wafer capacity is very low. In addition, the construction period of wafer capacity is relatively long. It takes approximately 36-48 months to build new production lines, similar to the design cycle for automobiles. Wafer companies cannot invest in capacity in the short term, so we have not seen wafer manufacturers increase capacity on a large scale after the global chip shortage crisis. Due to the long construction period of capacity, supply and demand balance between upstream and downstream is very important, but the epidemic has disrupted the rhythm.

The automotive chip industry chain is composed of wafer companies, chip companies, automotive parts companies, and car companies. When the epidemic occurred last year, major car companies lowered their annual targets, so upstream chip companies and wafer manufacturers followed suit and lowered their initial stockpiling to stabilize capacity. In the second half of the year, the domestic auto market grew rapidly, but upstream wafer factories could not expand capacity in the short term. The shortage of chips was exacerbated by some carmakers and distributors buying excessive amounts of chips offline and stocking up excessively.The automotive industry has been the fastest growing semiconductor application market in the past five years. Currently, the total cost of a single vehicle chip exceeds $500, and there are over 50 high, medium and low-end chips combined. With further development of intelligent electric vehicles, the application scenarios of automotive semiconductors will increase, and the demand for the quantity and functionality of semiconductors will also increase. However, the problem is that the supply chain of automotive semiconductors is basically not in China. According to data from the National Innovation Centre of China, over 85% of domestic automotive semiconductors depend on imports, and there are almost no locally developed chips, which presents a great challenge to the safety of the chip supply chain.

National Teams Needed for Automotive Chips

The current wave of automotive innovation is actually a competition between countries in terms of technology and supply chain. Enterprises and countries that master the core integrated capabilities of software and hardware will have advantages in the new round of competition. The digital economy is China’s future, and fields such as digital currency, smart cars, smartphones, and IoT will all have a great demand for chips. Therefore, it is crucial for both economic development and national security to establish a sound chip supply chain.

Just like cars, chips are also divided into low-end and high-end products. The current chip shortage crisis mainly lacks 8-bit MCUs for ECU and control systems. These products have low technical content, but are widely used in cars, from control systems to audio-visual systems. After the upstream supply is cut off, many automakers have to stop production and adjust their capacity. As the automotive electronic and electrical architecture evolves from distributed to centralized, the MCU chips for traditional control systems will decrease, and high-computing chips used in intelligent cockpits and autonomous driving will become mainstream products.

When it comes to national teams for automotive chips, the division of labor in the chip industry cannot be ignored. The chip industry is divided into three major links: design, manufacturing, and packaging.

Most industrial-grade chips use technology above 28nm, such as TVs and air conditioners. Ten years ago, China was completely blank in this field, but in recent years has made significant breakthroughs in materials and chip design. The development of consumer-grade and automotive-grade chips lags behind industrial-grade chips. CPUs, GPUs, and AI chips used in smartphones and automotive applications require high process and technology requirements. China still has significant shortcomings in the design and manufacturing of high-end products such as 7nm and 5nm.

Compared with consumer-grade and industrial-grade chips, automotive-grade chips have higher requirements and need to ensure stability in complex temperature and vibration environments.## Developing Self-made Chips: Will Chinese Automakers Follow in Tesla’s Footsteps?

Everyone may ask, what is the motivation behind automakers developing self-made chips? First is supply chain security, second is product differentiation. Huawei is currently developing the intelligent driving MDC chip to ensure domestic supply chain security. However, major automakers still tend to develop their own chips. In the long term, in the era of intelligent electrification, self-made chips will be the core competitiveness of automakers. Customized chips can better develop systems, have greater room for product iteration such as OTA, and create product and ecological differentiation. Developing self-made chips is only a matter of timing for automakers.

Tesla is a typical example of self-made chips. After splitting from Mobileye and Nvidia, Tesla developed the FSD chip on its own in three years. It uses 14 nm process and is manufactured by Samsung, with a single chip computing power of 72 TOPS. Tesla’s chip is not the top in terms of process or computing power, but self-made chips greatly improve Tesla’s synergy of “software + hardware” and significantly enhance the upper limit of its pure visual automatic driving.

A number of domestic enterprises, such as NIO and SAIC, have also attempted pure visual solutions, but due to the inability to achieve software and hardware synergy, the ceiling is very obvious and they have turned to adopt lidar. In addition to automatic driving chips, Tesla is also developing 7nm audio and video entertainment chips, which are manufactured by TSMC. Qualcomm 820A will also become part of Tesla’s history in the near future.

Domestic automakers are subject to technology and craftsmanship, and their self-made chip roadmap is different from Tesla’s. After Tesla completes the chip design, it only needs to find manufacturers such as Samsung and TSMC for manufacturing, wafer testing, packaging and testing before use. There are no such foundries in China due to neck bottlenecks.

In addition, the design capabilities of domestic automakers are not yet mature. Currently, most of them enter the chip field through cooperation. For example, Geely’s E-car and Arm China jointly established Xinqing Technology to build a chip industry chain. DearCC, the parent company of Leading Ideal, designed the “Lingchip 01”, which meets low-level automatic driving functions such as 360° panoramic views, automatic parking, and ADAS domain control.Software-defined car is a consensus across the industry. To carry software, cars must have good hardware. Mechanical architecture is already very mature, while the core hardware of electronic architecture is the control chip. Car companies are very aware of the importance of the control chip, but the timing is not yet mature. On the one hand, domestic high-end chip manufacturing solutions are not mature enough, and car companies cannot find suitable foundries. On the other hand, as intelligent electric vehicles have not yet reached a large scale, chip manufacturing cannot form economies of scale. In the short term, car companies still have to rely on companies like NVIDIA.

Currently, car companies such as Geely, BYD, and NIO have already carried out chip collaborations or self-research through various means. In order to make autonomous driving come faster and to increase the company’s market value, car companies must have their own “HiSilicon” through self-research of chips. Only in this way can car companies form a synergistic ability of “software + hardware” and break the so-called “Tesla moat”.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.