Automotive Intelligent Cockpit Platform Configuration Strategy and Industry Research Report, 2024

According to the evolution trends and functions, the cockpit platform has gradually evolved into technical paths such as cockpit-only, cockpit integrated with other domains, cockpit-parking integration, cockpit-driving-parking integration, and cockpit-driving integrated central computing platform. OEMs and Tier1s are actively laying out.

Layout strategy of intelligent cockpit platform solution (by type)

(1) Cockpit-parking integration: the first choice for mid- to low-configured models, mass production is expected to accelerate in 2024

Cockpit-parking integration fuses parking functions into the cockpit, saving hardware costs for parking controllers. Therefore, OEMs consider it from the perspective of user experience and cost optimization, and mid- to low-configured models are the main market for this solution.

In 2024, with the support of high-computing power cockpit SoC products and multi-camera hardware, many mainstream intelligent cockpit solution providers such as Bosch, Visteon, Desay SV, ADAYO, BICV, ECARX, and Yuanfeng Technology have launched relevant cockpit-parking integration solutions.

Most cockpit-parking integration solutions are mainly based on a single cockpit SoC chip to realize intelligent cockpit and parking functions. Qualcomm 8255, Qualcomm 8295, etc. have become the preferred chips for cockpit-parking integration's main single SoC solutions. Products based on this chip platform can enhance system collaboration, improve system performance, and more fully explore and realize cross-domain integration, active intelligence, multimodal interaction, and other functions.

(2) Cockpit-driving-parking integration: the optimal solution for cost reduction and efficiency improvement, with L2 ADAS functions integrated in the cabin

Cockpit-driving-parking integration integrates the cockpit functional domain with driving and parking functions, while driving domain currently integrates assisted driving functions at L2 and below. Therefore, this solution is mainly suitable for mid- to low-configured models equipped with L2 ADAS functions. High- or top-configured models will still give priority to being equipped with independent driving control units, which have higher requirements for computing power and functional safety levels, and are more difficult to integrate into the cockpit.

Leapmotor C10 model, launched in March 2024, is based on the latest 'Four Leaf Clover' central integrated EEA. In standard- and mid-configured versions, it uses one SoC chip (Qualcomm 8295/8155) and one MCU chip (NXP S32G), based on QNX Hypervisor to drive multiple systems such as QNX, Android, Linux, and RTOS, achieving deep integration of cockpit and intelligent driving, and realizing L2 intelligent driving (e.g. ACC), instrument cluster, infotainment, sound effects, power amplifier, surround view, fatigue monitoring, whole vehicle, body, gateway, and other functions.

(3) Cockpit-driving integration (central computing platform): Advanced intelligent driving integrated with cockpit, realizing One Box + unified vehicle OS platform, expected to be landed in 2025

Cockpit-driving integration is a central computing platform dedicated to achieving One Box, One Board, One Chip (including multi-SoC inter-chip cascading), among which One Chip is widely regarded as the 'ultimate solution' for cockpit-driving integration, which runs both intelligent driving domain and intelligent cockpit domain on a single chip.

Currently, multiple suppliers or OEMs have announced the next-generation single-chip cockpit-driving integration layout, but there are slight differences.

Among them, mainstream cockpit Tier1 suppliers such as Bosch, Autolink, ThunderX Auto, Hangsheng Electronics, ADAYO, Zhuoyu Technology, BICV, and Megatronix have announced the creation of cockpit-driving integration solutions based on Qualcomm Snapdragon Ride™ Flex 8775. Some suppliers have even implemented a product layout that integrates software and hardware with cockpit-driving integration, aiming at OEMs' demands for cost reduction on intelligence.

ThunderX Auto, a subsidiary of ThunderSoft, created RazorDCX Tarkine, a single-SoC cockpit-driving integration solution based on Qualcomm Snapdragon Ride™ Flex SoC. Equipped with ThunderSoft's Dishui OS cockpit-driving integration system, it supports a through-type 8K long screen, displays a full-scenario, immersive, and full 3D interface, and can realize 360° surround view, driver monitoring, game audio & video entertainment, interconnection and other cockpit functions. Meanwhile, it can also support automatic parking, L2++ intelligent driving functions for highways and urban areas, planned to achieve mass production from 2024 to 2025.

In 2024, companies such as Desay SV, Xpeng Motor, Li Auto, ZEEKR, Jiyue, BYD, and Xiaomi announced that they will create cockpit-driving integration products based on NVIDIA Thor, which can provide 'cockpit-driving integration' technical solutions for high-level autonomous driving above L4, and through systematic integration, the software and hardware are fully integrated, shortening the development cycle, significantly reducing costs and improving efficiency.

At GTC 2024, Nvidia announced that BYD will adopt Nvidia's next-generation intelligent vehicle chip Thor, which is expected to be used in the central brain of BYD's newly released Xuanji architecture to achieve a cockpit-driving integration layout, while also being compatible with deploying various AI models.

In addition, cockpit-driving integration solutions based on local chips from Black Sesame Technologies and SemiDrive Technology are also being laid out. In April 2024, NESINEXT cooperated with Black Sesame Technologies to release the CoreFusion, a cockpit-driving integration software open platform based on Black Sesame Technologies' automotive cross-domain computing chip Wudang C1296, providing developers with efficient operating system-level software base, development tool chain and complete ecology.

It can be seen that major suppliers and OEMs have accelerated the layout of single-chip cockpit-driving integration solutions, and are expected to achieve mass production as soon as 2025.

(4) High-end cockpit AI solution: running AI foundation model in the cabin to achieve 4K/8K HD display and 3D immersive high-performance experiences

As generative AI, 4K/8K HD display, 3D immersive experience, multimodal interaction, multi-scenario integration, etc. continue to be introduced into cars, the demand for higher computing power and high-performance cockpit platforms continues to grow, especially in the high-end car market, where market competition is becoming more intense.

Beginning in 2023, AI foundation models have become popular. Both emerging and conventional OEMs have introduced them into cockpits. AI foundation models to create future cockpits have become the current mainstream development trend.

However, current intelligent cockpit relies more on cloud AI in terms of computing power. Although it can run foundation models with more than 100 billion parameters, the car side currently only has 1 billion parameters or even less, and the performance level is still insufficient.

Considering that the cloud-side foundation model is subject to privacy security, latency, stability, cost and other issues, part of calculation and storage during the use of foundation model needs to be completed by vehicle-side computing resources, so the vehicle-side AI foundation model needs to be added. In the future intelligent cockpit, it is necessary to support an AI large language model with more than 7 billion parameters to achieve strong end-side intelligence. In addition, it will also have a full-time online foundation model in-car assistant, full-scenario multimodal interaction and other performance experiences, all of which place higher requirements on intelligent cockpit platform.

In 2024, many OEMs begin to introduce end-side foundation models into their cars. How to match capabilities of end-side foundation models will become one of the key development directions of new intelligent cockpit platforms in the future. Suppliers such as BICV, ECARX, and Shanghai Ruicheng have launched cockpit solution products with high AI performance.

In April 2024, BICV announced the creation of MARS06 cockpit solution product based on MediaTek 3nm chip MT8678, which can support end-side running of a 13 billion parameter AI large language model. It is outstanding in performance, multimedia, security technology facilities, audio and video processing capabilities, end-side AI foundation model, etc. And can meet the needs of intelligent vehicles for infotainment and intelligent driving assistance systems, and bring safer, more comfortable and smarter mobility experiences to occupants.

And CT-Y1 chip cockpit solution product based on MediaTek's 4nm sub-flagship chip supports a 7 billion parameter foundation model and can generate AI images within 1s. It also supports up to 12 cameras, up to 32MP, as well as multi-screen display and built-in 5G.

In March 2024, ECARX has released Qogir intelligent cockpit computing platform product, which uses flagship 4nm mobile phone chip Qualcomm Snapdragon 8 Gen3 and is equipped with Flyme Auto + WiFi 7 system. Its hybrid end-side AI reaches 60 TOPS, supports up to 10 billion parameter foundation model platform operation, supports up to 8K display, and hardware supports light tracing technology, providing the ultimate phone-car connectivity experience.

In addition, in 2024, high-performance cockpit platform products based on Qualcomm 8295 from Neusoft Group, Aptiv, Desay SV, Harman and other suppliers have been installed and achieved mass production on models from OEMs such as Jiyue, Geely Galaxy, ZEEKR, Li Auto, Xiaomi, Mercedes-Benz, etc. Its AI computing power reaches 30TOPS (the high-performance version reaches 60TOPS) and adopts a dual NPU architecture, which can effectively improve the speed and performance of foundation model calculations and help the efficient operation of end-side AI foundation models.

Furthermore, products based on cockpit-only platform solutions such as Samsung V920, Intel and Qualcomm consumer chips will also be mass-produced from 2024 to 2025, making market competition more intense.

Mainstream next-generation cockpit platform supplier layout: high computing power, focus on AI, flexible and customizable

Qualcomm 8255 cockpit platform solution: booming in 2024, may be the next-generation mainstream automotive solution

From the perspective of supplier layout, Qualcomm remains the main product in the market, mainly Qualcomm 8295 and Qualcomm 8255. Especially the Qualcomm 8255 product, highly recommended by Qualcomm, many OEMs who used to make 8295 intelligent cockpits have begun to access to 8255. However, performance of SA8255 is lower than that of SA8295, especially in terms of GPU performance. CPU performance is also slightly lower than SA8295. But in AI, SA8255 can reach up to 48TOPS, and the price is significantly lower than SA8295.

Beginning in 2024, many mainstream Tier 1 companies have launched their cockpit platform products based on Qualcomm 8255 chip, positioned as an upgraded version of Qualcomm 8155, which can directly undertake Qualcomm 8155 cockpit system ecology and shorten the development cycle for vehicle installations. At the same time, based on high computing power of 8255, cockpit integration solutions such as cockpit-parking integration can also be realized. It is expected that in 2024-2025, Qualcomm 8255 cockpit platform products are expected to inherit the Qualcomm 8155 market and become the mainstream solution in the market.

Samsung V920 platform solution: mainly deployed by Harman, Visteon and other foreign Tier1s, and Hyundai is the first to mass-produce and install it

In 2024, Harman launched the latest Ready Upgrade Advanced cockpit domain controller, which provides optimized solutions for high-end models or OEMs who want to add more displays, richer safety functions, and connectivity services to existing vehicles. Provide superior in-car experience through new Samsung Exynos processor, and introduce new features and service capabilities for on-demand access through Harman Ignite Store app and third-party applications. Meanwhile, Harman also collaborates with TE Connectivity to achieve easier hardware upgradability via patented NET-AX + modular hybrid connectors.

AMD V2000A platform solution: creating the ultimate cockpit for PC-level performance

ECARX builds Makalu computing platform, based on AMD Ryzen embedded V2000 processor and Radeon RX6600 series GPU. Equipped with ECARX Cloudpeak cross-domain system, it can provide console game-level graphics and image rendering capabilities. It also supports the latest graphics processing interface of PC-level computing platform and Unreal Engine 5, which can realize real-time rendering of 3D environment and 7. X.4 panoramic spatial audio. Comprehensively empower partners to create full-scenario 3D immersive cockpit functions and provide a complete soft and hard capability base for the next-generation intelligent cockpit products. It will be mass-produced in Smart models in 2024.

Automotive Ecological Domain Research: How Will OEM Ecology and Peripheral Hardware Develop?

Ecological Domain and Automotive Hardware Expansion Research Report, 2024 recently delves in the dynamics of automotive ecological domain in peripheral hardware expansion, supplier ecological solutions and OEM ecological implementation, and comprehensively combs and forecasts the development and future trends of automotive ecological domain.

First, the ecological domain will realize large-scale interconnection between IVI and external software and hardware, and obtain efficient collaborative perception and calculation of multiple devices.

The automotive ecological domain refers to the software and hardware ecosystem centering on IVI and including various peripheral devices. The IVI can call terminal sensors (such as mobile phones, cameras, UAVs, tablet PCs, etc.) and computing power, and attain efficient collaborative perception and computing of multiple devices through the cloud platform. The ecological expansion of automotive applications has always been a hot spot for OEMs and suppliers, and the development can be divided into three stages. In the first stage, ecological expansion is mainly realized through IVI application mall. In the second stage, the mobile phone ecosystem will be transferred to IVI through phone-vehicle interconnection. In the third ecological stage, watches, cameras, tablet PCs, and even UAVs, smart homes, seats, etc. are added.

Xiaomi Automobile is a typical ecological domain constructor. Relying on its HyperOS and Mi Home ecosystem, Xiaomi Automobile can support 1,000+ Mi Home devices to get on vehicles seamlessly, and can be connected without passwords under the same account to realize automatic scenario linkage.

In order to support ecological expansion, Xiaomi has reserved a large number of interfaces in central instrument displays (CIDs), center consoles and seat back. For instance:

• PIN points are reserved around CIDs to expand interfaces, hardware devices of the third parties can realize plug-and-play, and there is a magnet at the bottom of the screen to attract a group of physical buttons for the purpose of control in vehicles.
• The storage space of center consoles reserves hardware interfaces, power interfaces and magnetic attraction, which can expand more innovative hardware of the third parties, such as Mi Home automotive fragrance diffusers.
• The back of seats features 22. 5W magnetic charging. After Xiaomi’s tablet PC is adsorbed, it will directly become the original IVI, which can control 30+ automotive systems such as navigation, air conditioning, seating, audio and video. In addition, the IVI has CarPlay and AirPlay wireless connections, and the expansion of the seat back also supports iPad, which can control air conditioning, music and seats.

Second, OEMs often start from scenarios to design the ecological domain expansion.

According to Ecological Domain and Automotive Hardware Expansion Research Report, 2024, the automotive ecology has included hundreds of kinds of hardware and thousands of accessible hardware devices. From OEMs' point of view, Xiaomi Automobile can support the first batch of 1,000+ Mi Home devices to get on vehicles, GAC's Green Smart Car Home can support interconnection with 4,000+ devices, and Huawei's house-wide intelligence is connected with 7,000+ smart items carrying HarmonyOS.

Facing massive hardware resources, OEMs often conduct hardware layout from scenarios when designing the ecological domain, especially typical scenarios like smart home, gaming, viewing, karaoke, virtual reality, shooting, office, camping, charging, smart mobility, sports and health. At present, most of automotive hardware products are optional, and some of them can be purchased conveniently through official App stores of OEMs.

1. In entertainment scenarios, OEMs introduce a series of entertainment equipment such as game consoles (SWITCH, PS, etc.), movie-watching equipment, microphones, automotive UAVs, sports cameras, etc. to link multi-terminal information with IVI, thus improving playability of cockpits.

Hardware interconnection cases in entertainment scenarios:

• Dongfeng Warrior is equipped with a UAV system: Warrior 917, which was launched in August 2023, is equipped with a UAV system. The system is composed of a S400 UAV from GDU and a control system, focusing on safe and intelligent outdoor scenarios of IVI integration. Its roof can be expanded to be an apron. In addition to regular flying functions, the UAV can monitor roads, survey road terrain, slope, angle and other environmental information, and support synchronous intelligent path planning.
• Avatr has a custom karaoke microphone: Avatr boasts a wireless karaoke microphone (dual microphone) customized for the smart cockpit, with a built-in DSP processing chip, which can beautify human voice, optimize reverberation curve, and support linkage with the screen, sound and atmosphere lights in the cockpit.

2. In smart home scenarios, OEMs have promoted 'the whole ecology covering people, cars and home'. Users can remotely control and manage smart home devices (including air conditioners, water heaters, lighting, security systems, etc.) through mobile phones, cars and other devices.

Hardware interconnection cases in smart home scenarios:

• AITO and various smart home terminals realize car-home linkage: AITO integrates the capabilities of all kinds of terminals including smart screens, smart air conditioners, sweeping robots and home surveillance cameras in smart home scenarios based on HarmonyOS. In addition, AITO also supports preset scenarios like 'go home happily', 'leave home confidently'. For example, car owners can turn on air conditioners, lights, TV sets and music at home with one click on their way off work.
• GAC's “Green Smart Car Home” solution enables car-home interaction: GAC's “Green Smart Car Home” supports the interconnection of more than 1,000 home appliance brands and 4,000+ devices. It allows car owners to remotely control smart home appliances in their cars through voice, touch screens, etc. It will also wake up and control vehicle functions at home.

Third, the way to expand automotive ecological domain

In the process of ecological domain expansion, OEMs give priority to mobile phones as the entry point for ecological domain expansion. Through the deep interconnection between mobile phones and IVI, functions like computing power sharing, data seamless connection and hardware sharing can be attained, bringing users a more convenient and intelligent mobility experience. At the same time, massive user habit data and ecosystem on mobile phones are migrated to cars. On this basis, OEMs gradually expand interconnection with terminal products such as cameras, UAVs and tablet PCs, and build a perfect intelligent ecosystem. This expansion process not only provides users with richer functions and services, but also strengthens the competitiveness of OEMs in the field of intelligent ecology.

The cooperation between OEMs and mobile phone/technology vendors can be summarized and classified into the following modes: OEMs acquire mobile phone vendors, OEMs conduct independent R&D, OEMs cooperate with technology companies in research and development, mobile phone vendors manufacture cars (through cooperation), OEMs cooperate with mobile phone/ technology vendors. From the perspective of cooperation modes, some OEMs tend to get through the underlying operating systems of mobile phones and IVI, and promote Internet of Everything based on a unified platform. Other OEMs focus on improving IVI systems and seek cooperation with other vendors to achieve partial IVI interconnection.

In addition, it is worth noting that AM automotive hardware products are also an indispensable part of building a complete automotive ecosystem. When comparing the automotive hardware products of OEM app stores, we find that Huawei and Xiaomi, as top smart phone players, have a huge user base and mature e-commerce platforms (such as Huawei Mall and Xiaomi Mall). These platforms not only sell core products such as mobile phones, but also offer products in many fields such as smart home, smart wear and accessories. Therefore, for the car owners of Xiaomi and Huawei HarmonyOS, they can completely purchase automotive hardware products which they desire through these existing e-commerce platforms instead of OEM app stores.