7 questions, 7 answers: Why did Cambrian enter the autonomous driving chip market?

Cambrian Technology, a "China Science" company, is an AI chip design company. Originally a "unicorn" that had nothing to do with the automotive industry, it called out the slogan of "entering the field of autonomous driving chips" at this year's World Artificial Intelligence Conference.

Just when people were making all kinds of speculations and even feeling confused, Cambrian’s official “Answers to Investors’ Questions” came.

Question 1: What is Cambrian’s logic behind making in-vehicle smart chips?

Cambrian: The technical path the company intends to take is to promote the coordinated development of vehicle-mounted smart chip-related businesses through high-level intelligent driving chips and an open and unified software ecosystem.

Based on its judgment of technology routes and mainstream industry trends, and with the help of its R&D accumulation in the field of cloud-based smart chips, the company is fully prepared for chip design and R&D for high-level intelligent driving application scenarios.

At the same time, based on the uniqueness of the automotive industry itself, the in-vehicle smart chips designed by the subsidiary Xingge Technology will pay more attention to functionality, safety and adaptability of the software platform. Therefore, it is also necessary to design chips that meet automotive-grade requirements based on existing chip technology components to build a unified intelligent ecosystem of "cloud-edge-car".

Question 2: What plans does Cambrian have for business development in the field of in-vehicle smart chips?

Cambrian: As for the in-vehicle smart chip-related business, it is still in the product definition stage, and the company intends to adopt a technical path consistent with its traditional main business. The company will first focus on launching in-vehicle smart chips that meet the needs of higher artificial intelligence computing capabilities and are oriented to high-level intelligent driving application scenarios.

The on-board intelligent chip will integrate automotive computing cores such as automotive intelligent processors and automotive CPU cores, as well as various automotive peripheral interfaces. It will be mainly used to process the perception data collected by various sensors such as vision of smart cars, and plan and control intelligent driving based on the processing results of the perception data.

At the same time, the company will make full use of its existing technological accumulation in the field of cloud-edge intelligent chips, and based on the company's basic system software platform, fully mobilize the linkage between cloud-edge and vehicles, maximize the performance of on-board intelligent chips, and build a new intelligent chip ecosystem.

Question 3: In addition to making in-vehicle smart chips, will Cambrian also be involved in smart cockpit and other aspects of the business?

Cambrian: The company's first important goal in developing in-vehicle smart chip-related businesses is to design and develop high-level smart driving chips. Other types of in-vehicle smart chips will be developed based on market demand and the company's business development.

Question 4: In addition to the underlying software support, does Cambricon have any plans for upper-level assisted driving (such as visual processing, etc.) related software algorithms?

Cambrian: The company is positioned as an independent chip design company and does not directly engage in the development and sales of AI final application products to avoid competition with its own chip customers. However, the company has in-depth research and understanding of various AI algorithms and application scenarios, and can support the continuous update and iteration of customer algorithms through an open software platform, effectively supporting the needs of high-level intelligent driving.

Question 5: What role can the experience and technology accumulated by Cambrian in the field of cloud-edge intelligent chips play in the field of intelligent driving?

Cambrian: Intelligent driving is a complex system, a systematic task that requires the four-in-one linkage of "cloud, edge, terminal and car" based on a unified software and hardware ecosystem.

Specifically, the on-board intelligent driving chip is responsible for handling driving tasks; vehicle-road collaboration requires edge intelligent chips on the roadside to realize real-time collection and low-latency transmission of interaction information between roads and vehicles, and between vehicles; many data collected by sensors will be transmitted back to the cloud, and then cloud or edge intelligent chips will be used to process complex training and reasoning tasks.

If the above-mentioned chips do not adopt a unified basic system software platform, the systems are incompatible, and there are development barriers between the cloud, edge and end, customers will need to spend a lot of time, money and manpower costs on tedious transplantation, and will not be able to exert the optimal performance of the on-board intelligent driving chips.

At the same time, with the in-depth expansion of intelligent driving application scenarios, high-level intelligent driving will inevitably generate higher demands for artificial intelligence computing. The company is currently one of the very few manufacturers that can provide a full range of cloud-edge intelligent chips and processor products and a unified basic system software platform, and has a certain advantage in core technology accumulation and brand recognition.

Question 6: What is the attitude of automobile manufacturers towards hardware pre-embedding? In the future, when Cambricon and automobile manufacturers are working together, what is the initial judgment on hardware pre-embedding?

Cambrian: At present, L2-level in-vehicle intelligent chips have been applied in some models of some car manufacturers, but they cannot support high-level intelligent driving at present; for high-level intelligent driving, a few high-end models have pre-embedded related hardware, and a few models are waiting to be launched on the market. The technical path of hardware pre-embedsion can leave sufficient room for the upgrade of algorithms and application software, and is expected to become the mainstream technical path for high-level intelligent driving in the future.

Question 7: What is the staffing situation of Cambrian's wholly-owned subsidiary Xingge Technology? What is the future plan for R&D team building?

Cambrian: At present, Xingge Technology has initially established a company team, and some key R&D personnel have already taken up their posts. The number of personnel in the business development team and functional departments meets the company's current operational needs. After Xingge Technology completes the capital increase, it will accelerate the establishment of an in-vehicle smart chip R&D and productization team.

In the future, Xingge Technology will introduce and absorb more outstanding talents who are familiar with the software and hardware of the automotive industry based on the unique technical characteristics of the automotive industry, and accelerate the development of related businesses of in-vehicle smart chips.