In 1885, Karl Benz invented the first automobile, a creation that ushered in a new era of human transportation. Little did he know that over 130 years later, the phrase "a supercomputer on four wheels" would echo through the times, a notion now well-known and widely recognized.

Every time a statement of this sort is heard, the automotive chips that underpin the transformation of cars into supercomputers bask in the reflected glory.

01

Three roadmaps for automotive chips: Who can provide the optimal solution?

Crafting selling points is a crucial aspect of automotive marketing.

Over the past two years, car manufacturers have come to understand one thing: a smart cockpit is not omnipotent, but without it, they are utterly helpless. Data reflects this in the market, with car manufacturers achieving an 80% adoption rate of smart cockpits in the segment above 500,000 units.

This raises a question: what kind of cockpit chip should be chosen?

From a basic requirement standpoint, at the very least, a high-computational-power vehicle chip is needed. Under the demand for "one core with multiple screens," the computational power required for cockpit chips has already caught up with or even surpassed that of smartphone chips. The pursuit of computational power has become an inevitable hurdle for automotive chips.

Advertisement

Looking at future development, AI large models will pervasively shift from the cloud to the edge, with the endpoints being not only smartphones and computers but also, of course, automobiles. Chips integrated with generative AI technology are naturally a rigid demand for car manufacturers.The following question is, who can provide such cockpit chips?

Currently, there is a three-way competition among automotive chip manufacturers.

The first category consists of traditional automotive chip manufacturers, such as NXP and Renesas; the second category includes PC chip manufacturers, such as Intel and AMD; the third category comprises mobile chip manufacturers, like MediaTek, Qualcomm, and Samsung, which have entered from the smartphone AP chip segment.

The market for automotive chips has already begun to undergo significant changes. An increasing number of chip manufacturers are entering the automotive race, but the ultimate winners will definitely be those who have the capability to manufacture high-computing-power chips and the experience to integrate generative AI technology and ecosystems into end products.

Who can meet both criteria?

The answer is not difficult to give: MediaTek.

Traditional manufacturers struggle with the design and manufacturing of high-computing-power chips, which we will not elaborate on further. PC manufacturers have an innate advantage in computing and processing capabilities and are currently focusing mainly on the development of autonomous driving chips.

Mobile chip manufacturers like MediaTek have both the successful experience in the mobile field, such as smartphones, and significant advantages in low power consumption, high performance, integration, and application ecosystems. These characteristics are more in line with the needs of the automotive cockpit chip market.

02

MediaTek, Building Momentum for GrowthIn the past two years, Qualcomm has been unrivaled in the field of smart cockpit chip domain. According to statistics and organization by the GaiShi Automobile Research Institute, in 2023, the installed volume of Qualcomm's cockpit domain control chips exceeded 2.26 million, accounting for nearly 60% of the market share.

In the mobile chip field, MediaTek has always been a strong competitor to Qualcomm, and this competition has also extended to the automotive industry. In fact, MediaTek has long been a "veteran player" in the automotive field. With over a decade of experience in the flagship market, MediaTek has been deeply cultivating the automotive field and last year launched an integrated automotive solution—the Dimensity Automotive Platform. This platform covers four business directions: the Dimensity Automotive Cockpit Platform, the Dimensity Automotive Connectivity Platform, the Dimensity Automotive Driving Platform, and the Dimensity Automotive Key Components, offering a combination of automotive solutions with high computing power, high intelligence, energy efficiency, and reliability as their core advantages.

From the latest business situation released by MediaTek, the Dimensity Automotive Platform is experiencing explosive growth. Among them, the Dimensity Automotive Cockpit Platform has shipped over 20 million units globally, and the 3nm flagship and 4nm sub-flagship cockpit chips have been adopted by more than six leading domestic car manufacturers.

03

Inequality Appears in Automotive Chips

The wise adapt to the times, and the knowledgeable act according to circumstances.

In the pursuit of combining AI with high computing power, MediaTek has found a powerful partner. In 2023, when MediaTek announced a partnership with NVIDIA to provide a complete AI smart cockpit solution for software-defined vehicles, it caused a significant stir.Based on the cooperation plan announced by MediaTek and NVIDIA, both parties will fully enter the fields of smart cockpit chips and automotive SoCs, which is equivalent to directly entering Qualcomm's stronghold.

One is a super giant in the mobile chip market, and the other is an absolute king in the AI chip market. The combination of the two hegemons has begun to shake the automotive chip market.

In the future smart cockpits, a single SoC can support digital cockpits, ADAS, and autonomous driving functions. This requires the chip's computing power to not only meet data computation but also to meet graphic computation and artificial intelligence computation, and to achieve automotive-grade stability and safety.

From the current official announcement of the cooperation, the collaboration between MediaTek and NVIDIA has brought three impacts to the automotive market.

First, the competition in computing power.

As mentioned earlier, smart cockpit chips require high computing power, and high-process manufacturing technology is a necessary condition to support high computing power.

Currently, mainstream smart cockpit SoC chips have basically achieved a process below 10nm, including Samsung V9 at 8nm; those at the 7nm level include Qualcomm 8155, Huawei Kirin 90A, and Xingjing Technology SE1000.

Even automotive chip companies capable of investing in automotive cockpit systems, such as Texas Instruments, NXP Semiconductors, and Renesas Electronics, can only use the most advanced process at the 5nm node at this stage.

On April 26th of this year, MediaTek officially launched the 3nm process Tianji automotive cockpit chip CT-X1, as well as the 4nm process CT-Y1 and CT-Y0. Using the 3nm process in the field of smart cockpit SoC chips, MediaTek is the first in the industry. The advanced process will greatly release the chip's computing power, allowing the performance of automotive cockpits to take a leap.

As early as March of this year, MediaTek had already launched Tianji automotive cockpit chips with AI technology: C-X1, C-Y1, C-M1, and C-V1. These four products all support NVIDIA DRIVE OS software, covering the automotive market from luxury to entry-level segments. It is worth noting that these Tianji automotive cockpit platforms also integrate NVIDIA RTX GPU, which can support ray tracing technology.The enhancement of game graphics through ray tracing technology needs no elaboration, but its application on automotive chips signifies a significant breakthrough in the graphical processing capabilities of car cabin equipment. Traditionally, due to hardware limitations, the graphical processing of cabin chips often couldn't compete with PCs or gaming consoles. This integration now allows users to enjoy a more realistic visual experience within the car cabin, which is indeed surprising.

Secondly, the AI revolution.

Both MediaTek and NVIDIA agree on the role of AI in shaping the future. For instance, You Renjie, Senior Vice President of MediaTek and General Manager of the Computing Connectivity Metaverse Business Group, stated: "Just as personalized, intuitive computing revolutions brought to the mobile communications market, generative AI is also transforming the automotive industry."

NVIDIA's Vice President of the Automotive Business Unit, Ali Kani, also mentioned: "Generative AI and accelerated computing are reshaping the automotive industry."

In fact, the unique feature of the Tianzhi automotive cabin chips (SoC): C-X1, C-Y1, C-M1, and C-V1, launched by MediaTek in conjunction with NVIDIA technology, lies in the embodiment of AI technology in all aspects. On one hand, these cabin chips integrate AI computing accelerated by NVIDIA's next-generation GPU and NVIDIA RTX graphics processing technology. On the other hand, they can support the operation of large language models (LLMs) within the vehicle, making it possible for AI large models to be onboard, supporting advanced AI safety and entertainment applications such as in-car voice assistants, multi-screen displays, and driver alertness monitoring. Running these applications on the car's edge not only improves safety but also offers the advantages of high response speed and low latency.

Thirdly, the future of integration.

With the basic realization of domain integration in both the smart driving and smart cabin segments, the concept of "cabin and driving integration" has also been put on the agenda.

Cabin and driving integration refers to integrating the cabin domain with the smart driving domain into a high-performance computing unit, supporting both intelligent driving and smart cabin functions simultaneously.

In 2023, the Chinese market (excluding imports and exports) saw 6,346,700 new passenger cars equipped with cabin and driving intelligence (L2 and above intelligent driving + smart cabin) as standard, a year-on-year increase of 66.37%, with the penetration rate breaking through the 30% mark for the first time. According to predictions, this figure is highly likely to exceed the 50% mark this year.

However, achieving "cabin and driving integration" is not a simple task.From a supply chain perspective, "cabin and driving integration" sets higher demands for the participants. On one hand, it requires the simultaneous capability of mass production and delivery for both cabin and driving, as well as the effect of scale. On the other hand, for foundational elements such as computing platforms and software, there is a need for cross-domain capabilities.

These two capabilities perfectly match MediaTek and NVIDIA.

The integration of cabin and driving is also an important consideration behind the cooperation between MediaTek and NVIDIA. In terms of smart cockpits, MediaTek has long accumulated profound experience in the field of automotive chips. In the area of intelligent driving systems, NVIDIA has already launched platforms such as the DRIVE AGX Orin.

Impressive computing power, smarter AI, and a broader future all indicate that the combined force formed by the collaboration between MediaTek and NVIDIA seems to be greater than that of Qualcomm.

Inequalities in the automotive chip field are beginning to emerge.

04

MediaTek Sets a New Trend

During the Beijing Auto Show on April 26, MediaTek unveiled new products of the Dimensity automotive cockpit platform.

Among them, the Dimensity automotive cockpit platform CT-X1 uses a 3nm process, while CT-Y1 and CT-Y0 use a 4nm process. These three chips are based on the Arm v9 architecture, equipped with an HDR ISP image processor, supporting various intelligent image optimization technologies such as AI noise reduction and AI 3A, as well as advanced on-device generative AI technology.

Specifically, the newly launched cockpit computing chips by MediaTek boast exceptional performance.In the deployment of large AI models, the CT-X1 can support AI large language models with up to 13 billion parameters, while CT-Y1 and CT-Y0 can support AI human language models with 7 billion parameters. It is important to note that the largest models currently capable of running on automotive platforms support up to 1 billion parameters. In addition, the three chips can also run a variety of mainstream large language models (LLMs) and AI drawing functions (StableDiffusion) within the vehicle.

The 4nm process sub-flagship CT-Y1 chip's AnTuTu car version score exceeds 1.07 million, already matching the Snapdragon 8295, reaching the level of top flagships. The flagship CT-X1, built with 3nm process technology, also made its debut, with performance 30% stronger than the Snapdragon 8295 (from an architectural perspective, this is a very conservative figure), and AI computing power is 4-5 times stronger.

In reality, both future cockpit scenarios and autonomous driving scenarios have extremely high demands for computing power. The ultimate path, regardless of the route taken, will lead to advanced process technology.

MediaTek's core technology is computation, that is, the computing power of CPUs, GPUs, and APUs. The support behind high computing power is the mass production capability of advanced process technology. In the medium to long term, we will find that IC design companies with advanced process capabilities are the ones that can establish themselves in the generative AI-defined smart car market of the future.

With the rivers and seas at its bosom and the long winds of thousands of miles, MediaTek leverages its successful experience in mobile chips, deeply integrates AI with computing technology, drives the development of smart cars, and reshapes the future of automobiles.

In 2024, the market pattern of the automotive market still holds uncertainties.