"According to the innovation diffusion theory, the early users of smart cars are mainly 'enthusiasts', but the real start-up needs to rely on the 'mass group'. To seize the golden growth opportunity of smart driving popularization, it is necessary to capture the differentiated needs of the two groups , understand changes in industry rules. At present, smart driving products must focus on functions, experience, and return to user value; smart driving chips must meet performance standards , cost controllable, reliable supply, convenient development, and timely response .”
Recently, Superstar Future Partner and COO Zhu Yuqi attended the Global Automotive Chip Innovation Summit (GTIC 2023) and delivered a speech "Energy Efficient Computing with Software and Hardware Collaboration, the Cornerstone of a New Generation of Intelligent Driving ", sharing his observations on the trend of the intelligent driving industry, and summarizing The new environment puts forward new requirements for upstream suppliers of the industrial chain and smart driving chips.
The audience is changing, and the landing of innovative products needs to cross the "innovation gap"
Zhu Yuqi pointed out that the smart driving market is returning to its basics and entering a more pragmatic and stable growth stage. The core logic behind it has changed from "function priority" to "cost reduction and efficiency increase" , requiring upstream suppliers to achieve target functions under limited cost conditions, or to make greater profit margins under the same functional performance.
The reasons behind this shift can be interpreted through the "innovation diffusion theory". In the early stage of the development of innovative things, users are mainly enthusiasts and fashionistas; as things become popular, the proportion of core users occupied by the mass group is getting higher and higher. The needs of these two groups for products are irrelevant, which also causes a "gap" to be crossed for the landing of innovative products.
Taking smart cars as an example, in the eyes of enthusiasts, it is a big toy, so it emphasizes functionality and despises cost, and the technical route is also a matter of contention; in the eyes of the public, it is a means of transportation, so economic benefits must be considered in addition to functions, causing Industry throes.
The intelligent driving industry chain presents "value return" in multiple dimensions
To cross the innovation gap and seize the golden growth opportunity of intelligent driving, it is necessary to correctly capture the user value behind it and understand the changes in industry rules. Zhu Yuqi uses six "returns" to interpret current challenges and opportunities.
-
Development route : Smart driving has returned to a progressive route dominated by L2+ from the previous L4-led leap-forward route.
-
Industry division of labor : In the past, OEMs tended to conduct full-stack self-development, but now they are returning to division of labor and collaboration. Especially for large chips like smart driving, the sales of a single car company can hardly support huge investment, so the end result of the industry must be a division of labor.
-
Discourse weight : In the past, technology companies and car companies competed for the right to speak. At present, from the perspective of model selection power, pricing power, and enterprise stability, the right to speak still lies with auto companies, especially traditional auto companies.
-
Ecological level: In the past, it was said that the supply chain relationship was reshaped, but at present, the "new Tier1" still exists. Firstly, it will be more convenient for car manufacturers to issue fixed-points, and secondly, Tier 1 is also forming pre-integrated products, which are now often referred to as "ecological partners."
-
Configuration preference : In the past, it was preferred to pre-embed stacked materials and then OTA upgrades, but now it returns to cost reduction and efficiency enhancement, and you get what you pay for.
-
Value placement : The last point is also the most essential. The goal of the entire intelligent driving domain is to return to the emphasis on functions, experience, and user value, and reflect it on the entire industry chain.
Provide purchase value to consumers and use value to developers
For the key components of intelligent driving, who is the direct user and who will pay the final bill, the relationship and difference between the two is a question worth thinking about.
Zhu Yuqi believes that the direct users of intelligent driving computing chips are B-end solution providers and car companies, and the end-customers who make the final payment decision should be C-end consumers . In the past when smart driving was dominated by POC projects, OEMs (especially research institutes) paid the bill; now, the person who judges the value of smart driving products is the car owner, so apart from functionality, the decision-making weight of cost has greatly increased .
For consumers, to evaluate a smart driving product, it must have good functional experience, affordable price and reliable performance. For developers in the upstream of the industry chain, to evaluate a smart driving chip and its software tools, it is necessary to achieve controllable cost, reliable supply, convenient development, and timely response on the premise of meeting the functional and performance requirements .
Three firsts, leading energy-efficient computing for intelligent driving
In the future, Chaoxing will mainly provide energy-efficient computing solutions with intelligent driving computing chips as the core and software and hardware synergy. From the perspective of the product line , it includes the intelligent driving computing chip "Jingzhe R1", the self-developed AI processing architecture "Pinghu", the whole process development tool chain "Luban", and supporting the launch of intelligent computing development kits and intelligent driving reference solutions.
From the perspective of the business line , Superstar’s future is mainly aimed at the mass production of intelligent driving pre-installation, targeting customers such as OEMs, Tier1, and solution providers, and focusing on potential large markets; at the same time, the team is also actively exploring other edge-side AI scenarios. Including vehicle-road coordination, limited scene L4, industrial intelligence and other fields, mass production and shipments can be quickly produced.
From the perspective of delivery , Superstar will adopt the "self-selected supermarket" model in the future, which can provide flexible combinations of products and services from the product matrix according to the capability structure and differentiated needs of downstream customers.
In the intelligent driving business line , Chaoxing Future has launched multiple gradient reference solutions based on the Jingzhe R1 chip according to different application scenario requirements and sensor configurations , including NOVA-ADCU Ultra high-end driving and parking integrated reference solution, NOVA-ADCU Pro lightweight An integrated reference solution for traffic and parking, NOVA-ADCU Air (Bino) binocular smart driving reference solution, and NOVA-ADCU RS roadside perception reference solution.
In addition, Zhu Yuqi revealed that Superstar will be very concerned about the application prospects of AI large models in the car in the future. Currently, it has deployed large model perception based on Transformer, and will evaluate the application of cockpit interaction and data intelligence systems in the future.
Talking about the future development of Chaoxing, Zhu Yuqi believes that it can be summed up by "three firsts". "We have experienced the industry changes in the past few years, and we have created several precedents: first, we were the first to formally put forward the concept of 'intelligent driving energy-efficient computing solution provider'; second, we were the first and most persistent in the track to put forward the 'system Third, we were the first to call out the idea of "precise positioning, lightweight smart driving" among smart driving chip players. We are honored that with the development of the industry, our point of view has been consistent Once recognized and validated."
At present, Chaoxing Future has cooperated with more than 20 partners in the automotive field, and the marginal side has also begun to implement specific businesses. Zhu Yuqi said that 10% of chip design depends on money, 30% depends on people, and 60% depends on persistence. The upstream and downstream of the intelligent driving industry chain are comrades. We look forward to the establishment of an open and win-win cooperation ecology, and work together to bridge the innovation gap of intelligent driving. Believe it. Earn it.