Intel: The car gamble of "Intel Inside"

Intel, which once led the PC era, sees the automotive industry as the next hot spot after missing out on the mobile era.

Data shows that currently, 8,000 chips may be needed in a car, and as the complexity of cars increases, the demand for automotive semiconductor components is bound to grow steadily. According to the consulting firm Gartner Group, by 2020, the profit growth rate of the automotive semiconductor business segment will be twice that of the global chip market.

Therefore, the automotive industry is a new engine for the long-term development of the semiconductor industry. Faced with such a blue ocean, Intel, which missed the mobile era, does not want to miss this automotive revolution again.

Intel's way of entering the automotive industry is also through continuous mergers and acquisitions to make up for its own shortcomings. For example, by acquiring Russian computer vision company Itseez and Silicon Valley computer vision chip company Movidius, Intel has accumulated technology in deep learning and computer vision.

The acquisition of ADAS giant Mobileye not only enabled Intel to possess a powerful computing platform and communication capabilities consisting of "CPU+FPGA+EyeQ+5G", but also enabled Intel to gain nearly 70% of the ADAS market share and the low-cost advantage brought by the collaborative development of chips and algorithms.

In terms of business layout, Intel is also mainly divided into two parts:

First, in the smart cockpit, Intel has the Intel Atom A3900 processor developed specifically for automotive smart cockpits. The Apollo Lake platform based on this processor has powerful computing capabilities and can support functions such as in-vehicle infotainment systems/in-vehicle cockpits, digital instruments, rear-seat entertainment, and advanced driver assistance systems.

Secondly, in the field of autonomous driving, on the one hand, Intel relies on Mobileye's own autonomous driving solutions, and on the other hand, Intel also uses its own chips such as Xeon on the basis of Mobileye to build an autonomous driving platform, namely "Intel Go".

As is known in the industry, in March this year, Intel announced its transformation into a data-based company to tap into the potential of data. In fact, Ken Caviasca, vice president of Intel's Internet of Things Group, said earlier that self-driving cars need to process a large amount of image data, and Intel's goal is to become the brain of self-driving cars.

Of course, whether it is a smart cockpit or autonomous driving, in the highly competitive automotive chip market, Intel hopes to launch end-to-end solutions, as well as a comprehensive product line, powerful computing performance, etc. to compete with manufacturers such as Nvidia.

Transformation and layout

In October 2017, Intel's predecessor Paul S. Otellini passed away. During his eight years in office, Otellini made more profits than the previous CEOs combined. However, during the same period, Intel completely missed the mobile era.

Since Krzanich took office, he has tried to reverse Intel's position in the mobile field, but has not succeeded. At the same time, Intel's layout in areas such as smart wearable devices has also not yielded much results.

Intel, the pioneer in the computer industry, is facing unprecedented crises and doubts. Transformation has become imperative.

As mentioned above, in March this year, Intel announced its transformation into a data-based company, but in fact, it had already started its layout many years ago. Whether it is AI, the Internet of Things, or cloud and 5G technology, they are all based on data and applied.

In fact, Intel is not a new player in the automotive market, but automobiles have not become its core business.

As early as the mid-1980s, Intel and Ford had already started a joint venture to build electronic control units. It is said that Intel's 8051 microcontroller and derivatives were used in almost all Ford vehicles from 1983 to 1999.

In addition, according to Intel's official disclosure, as early as 2009, the predecessor of Google Waymo had already cooperated with Intel, using Intel Xeon central processing units, Arria FPGA for machine vision, and Gigabit network cards to provide communication for various components.

However, with the development of the four modernizations, cars are increasingly becoming mobile smart terminals. At the same time, according to data, an autonomous driving car can generate 100GB of data per second. Such a huge amount of data tests the ability of autonomous driving manufacturers in data storage, analysis and processing.

As a chip giant in the PC era, what Intel needs to do in the automotive field is to fully support the realization and application of intelligent connected vehicles and self-driving vehicles at the bottom level, from data processing to data transmission, data analysis and data security.

In fact, Intel did not have the core competitiveness to become an automotive semiconductor manufacturer in its early days, but like many giants crossing borders, it continued to make up for its shortcomings through mergers and acquisitions.

In August 2015, Intel acquired Altera, a company focused on developing FPGA chips, for $16.7 billion. It is reported that its autonomous driving FPGA chips have been mass-produced. Currently, the global FPGA market is mainly divided between Xilinx and Altera, which together account for nearly 90% of the market share.

In April 2016, Intel acquired Italian semiconductor functional safety manufacturer Yogitech. It is reported that the company mainly adds safety functions to chips, which are used in driverless cars and autonomous devices.

In the same month, Wind River, an Intel company focused on IoT software, also acquired Arynga. It is understood that the company can provide software that complies with the GENIVI standard to enable in-vehicle computers to support OTA functions.

In May, Intel acquired Itseez, a machine vision technology company. As we all know, machine vision technology is very important for self-driving cars, and the software and services developed by Itseez for driver assistance systems can warn of possible collisions, improve driver awareness and simplify driving. In response, Intel said that Itseez's technology has been used in other products such as safety systems.

In September, Intel acquired computer vision company Moviduis. It is reported that Movidius spent 9 years independently developing the low-power vision processor Myriad series VPU, which can provide powerful autonomous computing capabilities for devices and has broad application prospects.

In January 2017, Intel announced the acquisition of a 15% stake in high-precision map company HERE. At the same time, the two parties will cooperate in driverless cars and Internet of Things technologies.

In March 2017, Intel spent $15.3 billion to acquire Mobileye, allowing it to officially enter the automotive market. As is well known in the industry, Mobileye provides ADAS technology to major automakers worldwide and is also in a leading position in the field of high-precision maps and sensors, Intel has also been endorsed to quickly enter the automotive field.

Of course, in addition to continuous buying, Intel is also constantly expanding its circle of friends.

In July 2016, BMW, Intel and Mobileye formed the Autonomous Driving Alliance to jointly establish an industry standard and an open platform for autonomous driving. In May 2017, the Alliance introduced Delphi to develop a highly automated autonomous driving platform for BMW, playing the role of system integrator.

At present, it is reported that the Intel Alliance has already included automakers such as BMW, Audi, Mercedes-Benz, Ford, Great Wall, Changan, and Chrysler, automotive suppliers such as Delphi, Continental, and Magna, as well as travel companies such as Uber and Didi.

In addition, in September 2016, Ericsson, Audi AG, BMW Group, Daimler AG, Huawei, Intel, Nokia and Qualcomm announced the establishment of the "5G Automotive Alliance".

It is reported that the alliance will develop, test and promote communication solutions, support standardization, and meet society's needs for mobile Internet and road safety through applications such as autonomous driving, Internet of Everything services, smart city integration, and intelligent transportation.

Judging from the above news, Intel plays a role in the automotive industry in processing vehicle networking, big data, and recognition algorithms, and solving problems such as sensor aggregation, environmental modeling, and path planning.

At the same time, in the field of autonomous driving, in addition to Intel's own advantages, Mobileye is responsible for processing computer vision and local processing. The complementary advantages of the two allow Intel to thrive in the automotive field.

Intel's business layout

In November 2016, Intel separated its automotive team from its Internet of Things department. At the same time, the automotive team was divided into two parts, providing cockpit and autonomous driving solutions to automakers respectively.

In terms of cockpit, as mentioned above, Intel has the Intel Atom A3900 processor developed specifically for automotive smart cockpits. The Apollo Lake platform based on this processor has powerful computing capabilities and can support functions such as in-vehicle infotainment systems/in-vehicle cockpits, digital instruments, rear-seat entertainment, and advanced driver assistance systems. Specifically:

In-car entertainment system/in-car cockpit experience: It can support the next generation of navigation, radio, multi-screen entertainment and rear-seat entertainment systems, as well as connect to mobile devices, answer calls, listen to music and operate applications through voice recognition. In addition, it can use integrated virtual acceleration to perfectly support multiple operating systems to meet the integration of multiple workloads under cost-controlled conditions.

Digital instrument panel: It is understood that the Intel Atom A3900 processor makes digital instrument panel a reality, providing the required software-defined instrument panel for modern cars. At the same time, it uses powerful graphics and image processing performance to meet the development trend of 3D and high-definition large screens.