news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

21st Century Business Herald reporter Zheng Zhiwen reports from Shanghai

One day in 2018, in Silicon Valley, North America, NIO held an internal discussion lasting several hours. The topic of discussion was whether or not to develop its own operating system.

"Cross-domain operations will not work without the support of the underlying system," Wang Qiyan advised Li Bin, CEO of NIO. He believes that NIO must develop its own operating system. Wang Qiyan is a veteran R&D expert at NIO. Before joining NIO, he won the first place in the iCTF hacker competition and a Ford F150 Raptor. This is an international competition held for student teams majoring in information security around the world. After joining NIO, Wang was initially responsible for information security and functional safety-related businesses, and then he was promoted several levels and is now the vice president of NIO Digital Systems.

At that time, vehicle manufacturers rarely developed their own operating systems. The operating system they commonly used was AUTOSAR, which was born in 2003 and was jointly established by automotive parts manufacturers, vehicle manufacturers, and other electronics, semiconductor and software system companies. Its Chinese name is "Automotive Open System Architecture Alliance".

AUTOSAR can only be used on low-computing ECUs (Electronic Control Units). “It’s backward, and even the most advanced ones don’t work very well,” said Wang Qiyan. “In addition, car companies have to pay.”

The function of the self-developed vehicle-wide operating system is to integrate and dispatch computing power and manage all subsystems. Wang Qiyan helped Li Bin do the math and said that if a certain number of cars are sold, the R&D investment of NIO's self-developed operating system will be "worth it".

After a difficult 2019, NIO made up its mind to develop its own operating system in 2020. Since then, NIO has recruited teams around the world, spent four years on research and development, and invested more than 23,000 people.

Li Bin compared the self-developed operating system to "laying the foundation". When laying the foundation, it is necessary to plan clearly at the beginning, divide the various functions such as transportation, corridors, and pipe networks, and then build them in batches. The role of the foundation is "the better the effect and the greater the value will be in the future". If the foundation is well built, more upper-level applications will be needed.

At last year's NIO IN, NIO's full-domain operating system SkyOS debuted for the first time. This year, Li Bin introduced more features of SkyOS. He said that the system connects smart hardware, computing platform, communication and energy systems at the bottom layer to achieve unified management and coordination of full-domain applications such as vehicle connection, vehicle control, intelligent driving, digital cockpit, and mobile phone applications.

In Li Bin's view, applications can also feed back to the system. "Applications are running in the front. There is so much computing power and so many sensors on the car, the ceiling of applications is relatively high, and sometimes it will also bring good input to the operating system."

NIO became the first company among the emerging forces to release its self-developed global operating system.

Although Xiaopeng emphasizes "full-stack self-research" in terms of intelligence, including autonomous driving systems and smart cockpit systems, Xiaopeng's in-vehicle system is deeply customized based on Android, and the Xmart OS in-vehicle intelligent system is developed on it. It is not a full-domain operating system.

Li Xiang also firmly stated that he would develop a car operating system. In July 2022, Xie Yan, who was the chief architect of Alibaba's AliOS business group, joined Ideal Auto as CTO. He is considered to be the executive who is responsible for promoting the development of ideal operating system LiOS, autonomous driving chips and other technologies. However, Li OS has not yet taken shape.

In addition to its self-developed SkyOS, NIO has also developed its first intelligent driving world model NWM, successfully taped out the world's first 5nm intelligent driving chip NIO Shenji NX9031, and launched the new NIO Phone. In addition to user companies, NIO is trying to show the outside world its hard-core side.

"We insist on developing our own smart chips because we know very clearly that smart electric vehicles have once again become the commanding heights of technological innovation, and AI will become the core basic capability of smart electric vehicle companies," Li Bin explained. Ideal and Xiaopeng are also developing their own chips. According to media reports, Xiaopeng's self-developed smart driving chip has been sent for tape-out, "and it is expected to be returned in August." Ideal's smart driving will complete tape-out within the year.

Behind self-research is a huge investment in R&D.

The financial report shows that in 2023, NIO's R&D investment will be 13.431 billion yuan, exceeding Ideal's 10.586 billion yuan and Xpeng's 5.277 billion yuan. In the past eight years, NIO's R&D investment has exceeded 46 billion yuan. According to Li Bin, "60% to 70% (of R&D expenses) is in basic R&D."

However, NIO currently has a net loss of more than 20 billion yuan. Therefore, some people are questioning: When will the huge amount of research and development undertaken by NIO be able to turn a profit?

Weilai President Qin Lihong believes that Weilai is unlikely to be a company that gets rich overnight. "We think about problems in the long term and plan for the harvest in the long term."

In Li Bin's view, NIO is laying the foundation for a 100-story building. As long as the ground is above the ground, each floor can be built quickly. "Not every company is willing to do this, not every company has the ability to do this, and not every company can survive to see the results."

On the day of NIO IN, 21st Century Business Herald and several other media outlets had an in-depth discussion with Li Bin, Qin Lihong, and NIO’s R&D executives regarding the self-developed operating system, NIO’s input and output, and the industry’s competitive situation.

The following is an edited transcript of the interview:

“To lay the foundation for a 100-story building”

Q: Some people believe that with so much investment in self-development and self-production, NIO must create a particularly long-term business (model). Will this affect NIO's profit plan?

Li Bin: 60% to 70% of our expenses are now for basic R&D, not for a particular model or brand. The results of basic R&D investment can be used on all brands and models. Even if NIO sells 10 million units a year, the investment in basic R&D will not increase much. This is the benefit of platformization in the fields of intelligence and electric vehicles.

Take batteries as an example. NIO’s battery packs have been completely decoupled from the car. A brand and one or two generations of models use only one specification of battery pack. Even in the future, when the battery form and high-voltage platform are stable, four or five generations of models can use universal battery packs.

We need to lay the foundation for a 100-story building. As long as we get out of the ground, each floor can be built quickly. Not every company is willing to do this, not every company has the ability to do this, and not every company can survive to see the results. In products such as Ledao and NIO ET9, you will gradually see that it has begun to play a role.

Q: Some of the technological innovations that NIO has made will take many years and a lot of investment to see results. Why don’t you do some more obvious value creation? Do you feel like you’re at a disadvantage?

Li Bin: Yes. In terms of product definition, I think NIO has learned a lot of lessons. We should spend more energy on some functions and experiences that are highly perceived and valuable to users. If we don’t think comprehensively or understand deeply enough in some areas, we should change them. Ledao L60 has a refrigerator, a color TV, and a large sofa, which is safe, comfortable, and spacious. We listen to advice.

But explicit value creation and long-term technology investment do not conflict. If you don't make long-term investments in basic technologies, you won't spend money on some functions and configurations that users are highly aware of. Behind power consumption, cost structure, safety, and space efficiency, the competition is still about underlying capabilities. There is no shortcut to basic technology. NIO has insisted on forward-looking research and development since 2015, and the entire foundation is different.

Q: NIO has invested a lot in intelligent driving. What do you think about the "moat" of intelligent driving?

Li Bin: We are running a marathon, and there is still a long way to go. Even in the field of battery swapping today, I dare not say that we have a moat. The entire industry still has a very long way to go in terms of competition, and it is difficult to say who has a moat. For a long time, with such rapid technological changes and rapid iterations of various artificial intelligence technology routes, I think I can only go all out.

Q: What are the long-term plans and phased goals for NIO mobile phones?

Li Bin: Our investment in mobile phone R&D is basically a fixed cost, and we only make one model every year.

We still have pressure in making mobile phones. To make a flagship mobile phone, it definitely requires investment, and some investors think that the mobile phone business distracts the company's energy and resources.

But making a mobile phone today is different from making one ten years ago. The mobile phone industry chain is already very mature. 80% of our mobile phone team is working on software. I don’t expect to sell tens of millions of mobile phones a year, and I don’t need to compete on marketing expenses. NIO focuses on software experience, pure system, and driver connectivity. In terms of hardware, many capabilities are interoperable.

Q: What proportion does NIO hope intelligent driving technology capabilities will account for in its product value system?

Li Bin: Weilai's foundation is technology and research and development. I often say that a student who is good at Chinese is not necessarily bad at math, but may be well-rounded in moral, intellectual and physical development. Just like Weilai's service is good, it cannot be said that our technology is bad. Today, it is easy to be led by others.

Weekly sales data, no more than 10 people inside NIO know

Q: The NIO brand will not have any new models in 2024. Are you worried about the impact on sales?

Qin Lihong: Li Bin and I have always felt that NIO is unlikely to be the type of company that gets rich overnight. We think about things in the long term and plan for the harvest in the long term. Since May, NIO has maintained a monthly sales volume of 20,000 vehicles. At present, this level seems to be relatively stable. We do not make any form of sales forecasts and do not release any unaudited order information.

Three weeks ago, I attended an industry conference organized by the Automobile Association. The leaders said that your automotive industry has done a good job in high-quality development, but it has also done a lot of low-quality internal circulation. I raised my hand and said, can we not publish the weekly chart? First, the data is not serious. Second, our brand is not authorized. Why does another company use my brand to publish the weekly chart? Can we remove our brand? No more than 10 people in the company know Weilai's weekly sales data.

Q: Half of NIO’s replacement car owners are from BBA users. What do these half of users value about NIO?

Qin Lihong: Among the existing BBA gasoline car users, more than 30% of them may have already decided to try out new energy vehicles for their next car. This is a huge issue and a huge opportunity for us.

The characteristic of high-end brands is that they do not have a one-size-fits-all approach. The hardware must be strong enough, the software must be soft, the experience must be good, the service must be good, and the style, attitude, and all aspects of the entire company are very comprehensive requirements. Weilai's soft aspects such as aesthetics, humanity, and corporate attitude will make high-end gasoline car users feel that their spiritual consumption has not been downgraded when they switch to Weilai.

Q: If NIO makes driverless taxis, will it be more inclined to make money for car owners by utilizing existing car owners like Tesla does, or will it set up its own team to operate independently like the Radeon Sports Car?

Li Bin: NIO will never make driverless taxis. The value of intelligent driving is not to eliminate the hard work of today's private car drivers and taxi drivers. The real value is to reduce accidents and free up some of the energy of driving, making driving less tiring and safer. We believe that "freeing up energy and reducing accidents" is the real social value created by intelligent driving.

Traffic is a social issue, not a purely technical one. A city can only accommodate a certain number of operating vehicles, whether they are operated by people or not. This number means that driverless taxis will never become a business that is as scalable as software cloud services, with no boundaries. Wake up!

SkyOS is similar to battery swapping. It will be run first and then opened to external peers.

Q: What is the underlying difference between NIO's SkyOS system, Xiaomi's Pengpai system, and Huawei's Hongmeng system? Compared with mobile phone or PC operating systems, what is the difference in development difficulty and time for NIO's system?

Li Bin: Leading technology companies are increasingly aware of the importance of operating systems. I think each company has a different starting point. NIO is car-centric, so we consider stability, safety, car-based big data throughput, bandwidth, latency, robustness, etc. a lot. This is our car-making gene. We will also expand this to some terminals of panoramic interconnection, such as NIO Phone, to open up AI capabilities and information flow.

Q: NIO has cooperated with many car manufacturers in the field of battery swapping. Will the SkyOS system also cooperate with other car companies through some technologies?

Li Bin: The first step is to do our own work well and serve users of different brands, platforms and models within NIO. The second step is to open up the service to some parts partners that have a cooperative relationship with NIO. The third step is that if other vehicle companies are willing to use it, we are of course very willing to share it. Just like the process of swapping stations, it must be run through first, and coordinated internally before opening it to external peers.

Q: After SkyOS comes out, has NIO planned corresponding technology breakthrough points internally?

Li Bin: Some technological leaps, whether in language models, multimodal models, or world models, may bring many surprises in the next one or two years. It is definitely not possible to solve all problems, and there may be some new technical difficulties to be solved, but overall, progress will be very fast.