Is NIO entering the chip industry to follow the trend or because of internal demand?

Recently, Li Bin announced that NIO will develop its own driverless chips, and has established a hardware team codenamed SmartHW (Hardware). Whether it is a rumor to the media or the time has come for NIO to make chips, this news has become the focus of attention in the domestic automotive industry. As a representative of the new domestic car-making companies, NIO has overcome the initial uncertainty, and the positive gross profit margin has indicated that NIO can consider the long-term development of the company through self-sustaining while taking into account the current situation.

1. Self-developed chips are the key for car companies to expand the road of driverless driving         

Comparing the differences between Tesla and Volkswagen and Toyota, the former can surpass the latter in market value with a smaller size. The biggest reason is that Tesla is regarded as a technology company rather than a traditional company. To become a technology company, it is not enough to simply launch a few electric cars to convince investors. Breakthroughs in the underlying software and hardware technologies such as batteries, chips, and codes are the key to NIO winning the favor of investors and constantly attracting domestic and even global consumers to buy NIO models. This is also the best opportunity for emerging auto companies to widen the gap with traditional auto companies.

For the new four modernization technologies of automobiles, the competition of driverless technology has gradually become the key to the new four modernization technologies. The current range of electric vehicles is generally more than 500 kilometers, which can fully meet the daily commuting needs of ordinary consumers. Excessive pursuit of long range of electric vehicles is not the most economical solution for OEMs and consumers. In the future, the speed of battery potential tapping will definitely give way to the reduction of battery costs, and these are not urgent for OEMs like NIO to invest, and can be handed over to battery manufacturers like CATL to solve.

However, in the field of autonomous driving, especially the autonomous driving technology that reaches L4 level, the current competition can be said to have just begun. Most of the newly launched models in China are equipped with or at least the high-end models are equipped with driving assistance systems that reach L2 or L2+, and some models even call out L3 technology. However, the truly autonomous driving L4 level autonomous driving technology is still mostly in the laboratory or demonstration park roads. There are both problems of high costs, but the core is that there is no breakthrough in chips and algorithms. The code for autonomous driving has reached hundreds of millions of lines, and chip performance will be the key to determining whether L4 autonomous driving technology can be implemented in the future. Previously, Tesla dared to use a solution without laser radar, hoping to achieve perception of the surrounding environment through cameras, and the biggest confidence came from its own autonomous driving chip. Therefore, whether it is Weilai or other OEMs, if they want to win the autonomous driving competition in a few years, they must start from now on to develop exclusive chips.

2. 7nm chips will be the goal of NIO

At present, the mainstream in-vehicle driverless chips are all 14nm. For example, Tesla's FSD chip uses Samsung's 14nm, Nvidia's Xavier uses TSMC's 12nm, and Huawei's MDC's Ascend 310 also uses TSMC's 12nm. From the perspective of the chip development cycle, NIO can only start developing 7nm chips to ensure that the chips it develops will not be completely outdated in performance in two to three years. Because currently, the chip manufacturing processes of several major chip players in the field of driverless driving are 7nm. These include the chips developed by Tesla and Broadcom, Mobileye's next-generation chip EyeQ5, which was acquired by Intel, and Nvidia Drive Orin, which has just released a new generation of chips. If it weren't for 7nm, how could the power consumption be reduced by half and the computing performance increased by 6 times?

But the investment in chips is also huge. After completing the front-end and back-end design, for a company with no experience in chip design, the pressure is self-evident. According to relevant data, the design cost of a 7nm chip is about 2 billion yuan, and the later tape-out cost is about 200 million yuan per time. That is, each failure of tape-out means a loss of 200 million yuan. However, in addition to the high capital investment, talent is a headache for Li Bin. If you want to truly design a chip successfully, NIO needs to recruit a large number of senior experts from the United States. Just like Huawei, it is only possible to accomplish this by bringing together many scientists. But I don’t know if NIO, as a listed company, shareholders will give NIO the same strong support as Ren Zhengfei.

3. In addition to design, chip manufacturing issues also need to be considered

Chip design and development is certainly important, but chip manufacturing is equally important. This can be seen from Huawei. Huawei's self-developed HiSilicon is not inferior to Qualcomm's chip performance, but it is subject to the US export control from chip design to manufacturing, which makes it impossible for chip manufacturers such as Samsung and TSMC to manufacture the most advanced 5nm chips for Huawei, which directly leads to Huawei's inability to continue to produce high-end mobile phones. For NIO, if there is still no way to break through domestic lithography technology in the future, and at the same time NIO's self-developed chips can help it achieve performance that is not inferior to Tesla's driverless technology, then according to the current international situation, NIO needs to worry about whether it will follow in Huawei's footsteps and be suppressed by the United States. Because the current chip manufacturing is monopolized by companies such as TSMC and Samsung.

At present, NIO is going smoothly. On the one hand, terminal sales continue to hit new highs and gross profit margin has turned positive. Its brand and customer operation and maintenance have been fully accepted by some domestic consumers, helping it to gain a foothold in the luxury brand market; on the other hand, benefiting from the loose policies of the US capital market and the aftermath of Tesla's stock price surge, NIO's market value has also set new highs. Its total market value has even surpassed SAIC Group and ranked second among domestic vehicle manufacturers. It is in this context that NIO has a certain confidence and resources to start self-developing driverless chips exclusive to its models. But the road to chip research and development is full of difficulties. It is really not easy for NIO to succeed.