Intelligent connected car chips interpret "three companies are promoted"

The competition will become more intense

Whether it is computing power bases, wireless connections, smart cockpits, autonomous driving and vehicle-road collaboration, the competitive dimensions of smart connected car chips are becoming increasingly diverse. In the future, competition among Nvidia, Qualcomm and Intel for intelligent connected car chips will become more intense. If the PC era has achieved Intel, and the smartphone era has achieved Qualcomm, then everyone hopes to become the new king in the great changes in the automotive industry. NVIDIA is unparalleled in terms of single-chip computing power indicators, but this does not mean that NVIDIA has no rivals in the field of intelligent connected car chips.

In fact, the current mainstream model in the battle for computing power in intelligent connected cars is still the parallel use of multiple domain controllers, using multiple chip stacks to ensure computing power. When talking about chip computing power, Yang Yuxin, chief marketing officer of Black Sesame Intelligence, told a reporter from China Electronics News that to achieve large-scale, cost-effective mass production of autonomous driving, a single chip with high computing power is not the only way.

In other words, even if a single chip has insufficient computing power, a solution composed of two or three stacked chips can still achieve the same effect. For example, Qualcomm's Snapdragon Ride autonomous driving solution can provide computing power from 16TOPS to 2000TOPS; the EyeQ Ultra developed by Mobileye, which works closely with Intel, has a computing power of 176TOPS and is expected to be supplied by the end of 2023 and fully achieve automotive grade in 2025. mass production.

Under the trend of combining software and hardware in algorithm-driven AI chip design, Qualcomm and Intel are not showing weakness. Although Qualcomm has laid a solid foundation in the field of smart cockpits with its advantages in mobile phone chips, Qualcomm's automotive business has never been limited to smart cockpits. Unlike smart cockpit chips, which are highly dependent on smartphone chips for development, smart driving chips are exclusive products customized for cars.

Looking at the current fiercely competitive ADAS and autonomous driving chip markets, Intel and Nvidia have established ecological advantages, while Qualcomm is still on the road to catching up. Due to its late start in ecological accumulation, Qualcomm’s remedial work in ADAS and other fields will be a long and complicated process. "Qualcomm attaches great importance to the ADAS field. We are continuously expanding our product roadmap from connected cars to ADAS, and we are paying great attention to software stack development.

As you can see, this is a very long process and very complex. " Nakul Duggal, senior vice president and general manager of automotive business at Qualcomm Technologies, said. Intel's autonomous driving department is also facing considerable challenges. Since Mobileye joined Intel, although its performance has continued to grow, the growth rate has been far from the original expectations.

In the context of everyone transforming into automotive smart chips, Mobileye has not shown its dominant strength and is facing challenges from Intel and Qualcomm in the automotive chip business. For example, the computing power of the self-driving chip EyeQ5 delivered by Mobileye only reaches 24TOPS. In comparison, the computing power of NVIDIA's Orin chip has reached 254TOPS, which is almost 10 times that of Mobileye.

Mobileye revenue at a glance (fiscal years 2019-2022)

Data source: Compiled based on Mobileye financial report information

Only by seizing market growth points can we maintain performance growth. Mobileye, which is listed independently from Intel, urgently needs to improve its ability to withstand pressure. Recently, Mobileye lowered its full-year revenue forecast for 2023, and subsequently its stock price fell sharply.

According to news on June 6, Intel will sell 35 million shares of Mobileye’s Class A shares, with a total value of approximately US$1.5 billion. After the sell-off, Intel's stake in Mobileye dropped from 99.3% to 98.7%. It remains to be seen whether this sell-off will further optimize Intel's business structure. In short, in the "three companies advance" pattern, NVIDIA, Qualcomm and Intel all have their own advantages and face challenges. The "watershed" in the automotive industry has just begun.