In the era of mobile phones, the mutual achievements of Xiaomi and Qualcomm were once rumored to be a good story. The battle for the first launch of Qualcomm chips is also a reserved program when domestic flagship mobile phones are released every year.
Entering the age of automobiles, Qualcomm is also becoming an unavoidable name.
Starting with the Snapdragon 8155 chip, Qualcomm cockpit chips have become the standard configuration of almost all mid-to-high-end models. The debut of the next-generation Extreme Snapdragon cockpit platform (SA8295P) has also attracted widespread attention in the industry.
But apparently, Qualcomm doesn't intend to stop there.
During the Beijing Auto Show, Nezha and Qualcomm announced that Nezha Automobile will be the first to be equipped with the Snapdragon Ride Flex SoC (SA8775P) cockpit and driving fusion platform.
In the past two years, the popularity of cabin and pilot integration has been rising, with followers and wait-and-see. And Qualcomm's decisive bet is a heavy weight on the side of the scale.
As the leader of cockpit chips and a new player in the field of intelligent driving, Qualcomm showed off the circle of friends of the new platform at the Beijing Auto Show.
In addition to the world of car connection, Zhuoyu Technology and Hangsheng also released cross-domain integration solutions based on Qualcomm's Snapdragon Ride Flex SoC during the auto show. Thundertech, Momo, Momenta, etc. have released intelligent driving solutions based on the Snapdragon Ride platform.
Qualcomm's footprint in the automotive era is gradually unfolding.
Qualcomm fills up the puzzle of intelligent driving
The Snapdragon Ride platform is Qualcomm's solution in the field of intelligent driving, including products such as system-on-chip, safety accelerator and autonomous driving software stack.
In January 2020, Qualcomm released the first Snapdragon Ride platform (SA8540P) for ADAS and AD. The latest generation of the Snapdragon Ride platform, including SA8650P and SA8620P, is said to support intelligent driving solutions from Level 1 to Level 4.
Until this year's CES exhibition, Zhuoyu Technology demonstrated its urban NOA solution based on a single Snapdragon 8650 chip and does not rely on lidar, and Qualcomm intelligent driving chips began to surface.
At the 100-person meeting of electric vehicles, Zhuoyu Technology officially released the urban NOA plan with a cost of only 7,000 yuan, and the Snapdragon 8650 became a hot spot.
Shen Shaojie, the person in charge of Zhuoyu Technology, said in the exchange that Snapdragon 8650 is a very good chip, and Zhuoyu is happy to be the first batch of companies to use this chip and land the product as soon as possible.
It is reported that the new solution of the Chengxing platform is expected to be mass-produced in China this year, becoming the first batch of mass-produced models equipped with the Snapdragon 8650.
At this Beijing Auto Show, Zhuoyu Technology and Qualcomm also announced the expansion of their cooperation, in addition to the intelligent driving products of the Chengxing platform based on the Snapdragon 8650, and also launched a cross-domain integration solution based on the Snapdragon 8775 chip.
According to reports, the solution developed by Zhuoyu Technology on the Snapdragon 8650 chip has been relatively mature, and it is expected that the cross-domain solution based on the Snapdragon 8775 will also be implemented next year, becoming the first batch of cross-domain products based on this Snapdragon cabin and driving fusion SoC on the market.
Released in January 2023, the Snapdragon Ride Flex SoC (SA8775P) supports functions including smart cockpit and smart driving with a single SoC. At the hardware architecture level, the Snapdragon 8775 can implement isolation, interference-free, and quality-of-service control functions for specific ADAS functions, and has a built-in ASIL-D dedicated safety island. At the software level, multiple operating systems can be supported running simultaneously, with isolated virtual machines and real-time operating system support hypervisors that support automotive open system architectures to meet the needs of multiple mixed, critical workloads.
At the Beijing Auto Show, Zhuoyu Technology, Auto Link World, Huayang and Hangsheng all launched SA8775P-based cabin and driver integration solutions.
The cooperation solution between Qualcomm and Momo is based on Snapdragon 8620HP370, this product is a 5000 yuan level intelligent driving system, with 36 TOPS dense computing power, which can support high-speed, urban expressway NOA, and memory driving function in the city, and the parking level can realize functions such as teaching-free memory parking and intelligent obstacle bypassing.
Momenta will also develop intelligent driving functions based on the Snapdragon 8620 and Snapdragon 8650, supporting a variety of scenarios including high-speed pilot assist and city pilot assist, which are expected to be installed in production models later this year.
From cockpit to intelligent driving, cockpit integration is Qualcomm's inevitable choice
This is Qualcomm's first large-scale announcement of a collaboration in the field of intelligent driving.
In the cockpit field, Qualcomm has an absolute advantage.
Since its release in 2019, its third-generation flagship Snapdragon cockpit platform (SA8155P) has been favored by almost all car companies. After the release of the fourth-generation Extreme Snapdragon cockpit platform (SA8295P), it has also become a hit in the market. Not confined to the level of the model, in practical application, as a model within 200,000 yuan, Leapmotor C10 also uses the Snapdragon 8295 chip, which can be seen in its recognition in the market.
Qualcomm's advantage in cockpit chips stems from its accumulation in the mobile phone consumer market.
Ai Hezhi, senior director of product marketing at Qualcomm Technologies, said in an exchange with Leifeng.com's "New Intelligent Driving" that compared with mobile phone chips, automotive chips are not easy to do, the volume is not large, and the research and development cost is very high. However, Qualcomm has a lot of technical accumulation in the field of mobile phone chips, and at the same time, the software development experience can also be transferred and reused in the cockpit field, which can dilute the cost of some automotive chips. This is the advantage of Qualcomm.
Although today's automotive chips are "difficult to make and not profitable", smart cars have clearly become the largest growth market after mobile phones. According to Frost & Sullivan data, the global automotive chip market size will be about 310 billion yuan in 2022 and is expected to exceed 600 billion yuan by 2030.
Qualcomm's revenue forecast for the automotive chip business is to exceed $4 billion by fiscal 2026 and increase revenue to more than $9 billion by fiscal 2031.
With the improvement of the intelligent level of automobiles, the integration of the whole vehicle has increased, and the integration of cabin and driver has become the next direction of the development of electronic and electrical architecture.
At the same time, the expectation of cost reduction in cockpit and driver integration has also attracted investment from OEMs. The reuse of hardware such as sensors, domain controllers, etc., means reduced costs. The high level of integration can also reduce the amount of wiring harness, reducing the cost and reducing the weight of the whole vehicle.
Following Tesla, NIO has also announced that the electronic and electrical architecture of the 2024 second-generation platform model will be upgraded from a domain control architecture to a central computing platform, and the biggest change is that the original intelligent cabin and intelligent driving separation will be upgraded to cabin-driving integration.
In the face of the market trend of cockpit and driver integration, supply chain enterprises need to have both the mass production and delivery capabilities and large-scale effect of cabins and drivers, and at the same time, they also need to have cross-domain capabilities in basic elements such as hardware and software.
For Qualcomm, the development of intelligent driving products and cockpit integration products has become an inevitable choice for deep cultivation of automotive chips.
Ai Hezhi said that from a technical point of view, cabin-driver integration can not only reduce costs, but also increase performance under the same conditions and bring more innovative structures.
For example, now the large model of the car is more used for simple interaction, but after the integration of the cabin and the driver, if the vehicle information and camera information are combined, as well as the environmental information of the vehicle on the road, there will be richer content and provide a broader imagination space.
According to Ai Hezhi, Qualcomm began to put forward the concept of cockpit integration two or three years ago. "At the beginning of this concept, the customer's reaction was to wait and see, and as the product gradually matured, more and more customers began to do some development based on this platform and implement the product. Since last year, more and more car companies have begun to be determined to use the products of cabin and driver integration in their models. ”
The Snapdragon Ride Flex SoC (SA8775P) is Qualcomm's first cockpit fusion product, and it is believed that the next generation of products will develop in a more high-end direction and support flagship-level cockpit and driver integration. In addition, Qualcomm will also introduce more cost-effective solutions.
"But the first step is to do a good job of this product, so that car companies feel that this is a good product and a very promising direction, so that they can go on. Ai Hezhi said.
The market is Qualcomm's biggest advantage
Although it is a newcomer in the field of intelligent driving, Qualcomm quickly found its own direction.
Some industry insiders commented on Leifeng.com's "New Intelligent Driving" that compared with other manufacturers' strategies that focus on high performance, some of Qualcomm's intelligent driving products are more suitable for cost-effective routes, and the software can be quickly reused, which is more suitable for models that require a high volume.
However, Ai Hezhi said that Qualcomm's cost-effective route first guarantees performance, and on top of performance, it provides customers with more possibilities, rather than just the price itself.
On the other hand, under the pressure of the price war of car companies, the market's perception of computing power is also returning to rationality. The capability of intelligent driving chips does not only depend on the size of computing power, and the criteria for car companies to select products are shifting to the adaptability of chips.
The Snapdragon Ride platform offers an open, programmable architecture that can be customized by OEMs and Tier 1 vendors.
For visual perception, parking, and driver monitoring scenarios, Snapdragon Ride uses multiple software stacks that allow users to choose one or more combinations. In addition, the Snapdragon Ride vision system offers a modular architecture that integrates map crowdsourcing, driver monitoring systems, parking systems, cellular telematics and positioning modules to support better customization and up-integration.
The cost of the 100TOPS urban NOA program disclosed by Zhuoyu Technology is 7,000 yuan, and if the intelligent driving program accounts for 4% of the vehicle cost, it can be on the model in the price range of 15-200,000 yuan, which is the price of the most mainstream model in the current market.
Judging from the current feedback, the release of the program is obviously a mutually successful cooperation for Qualcomm and Zhuoyu Technology.
In the field of cockpit integration, Qualcomm has a distinct advantage.
Cockpit integration is not easy to achieve.
At the technical level, the requirements for functional safety in the two fields of intelligent driving and cockpit are different, and the consumption mode of chip resources is also different. In order to integrate the two domains, different aspects such as chip architecture, operating system, and underlying middleware need to be considered comprehensively.
According to incomplete statistics from the Gaogong Intelligent Vehicle Research Institute, at present, the layout and landing speed of suppliers in the field of intelligent cockpit in cross-domain capabilities are much faster than those in the field of intelligent driving.
According to industry insiders, the integration of cabin and driver must first ensure that the driver is multi-screen, so as to have a better experience. Specifically, the cockpit fusion chip must first ensure that it can at least drive the instrument screen and the main display at the cockpit level, and today's cockpit is generally required to be able to drive games.
Most smart driving chips do not have the ability to drive multiple screens.
This is related to the underlying architecture of the two chips. In general, intelligent driving emphasizes perception, planning, and decision-making, which requires high NPU computing power, while cockpit chips need to process a large amount of image content and run applications such as large-scale 3D games, which have higher requirements for GPU capabilities.
The GPU capability of intelligent driving chips is usually low, but most of today's relatively high-end cockpit chips have a certain amount of NPU computing power, so it is relatively easier and faster for cockpit chip suppliers to do cockpit integration.
After supplementing the capabilities of intelligent driving chips, Qualcomm's capabilities in the field of cross-domain integration have also been further strengthened.
The Snapdragon Ride Flex family of SoCs enables sophisticated cockpit capabilities such as immersive high-end graphics, integrated instrument clusters for infotainment and gaming displays, and rear entertainment displays. The pre-integrated Snapdragon Ride vision software stack enables a highly scalable and safe autonomous driving experience that scales up to support higher levels of autonomous driving.
Qualcomm provides the platform with the fastest development progress in cabin and driver integration. At CES 2024, Bosch has launched a cockpit and vehicle convergence in-vehicle central computing platform based on the Snapdragon 8775 chip. At the Beijing Auto Show, Zhuoyu Technology also displayed its demo based on the Snapdragon 8775, which, according to the staff, is expected to be mass-produced in 2025.
Automobiles have entered the era of software-defined vehicles, and chips have also entered the era of software-defined chips.
Compared with the third-generation Snapdragon 8155, the Snapdragon 8295 in Qualcomm's fourth-generation Snapdragon cockpit platform has increased AI functions in addition to the performance itself, which is determined by market demand.
With the development of vehicle intelligence, the needs of users are also changing rapidly. For supply chain enterprises, being able to get more feedback data will also help them develop products that are more in line with market demand.
Compared with intelligent driving products, cockpit products have a higher penetration rate in today's automotive market. As a leading supplier in the cockpit field, Qualcomm has a huge user base.
According to the data, in 2023, 3,490,100 passenger cars in the Chinese market will be delivered as standard with cockpit domain controllers, and the front-loading rate will exceed 16%, about 7 percentage points higher than the intelligent driving domain controller market. And Qualcomm's market share in the cockpit field is also considerable.
More product and market accumulation, as well as a deeper understanding of users, will help Qualcomm further deepen its development in the field of automotive chips.