Tesla, Inc. delivered the first Model S in a California factory at the end of June 2012. its stylish shape, 4.4s 100km acceleration and 17-inch central control screen subvert the stereotype of traditional electric cars as "short and poor" and attract the attention of many tech bigwigs, including a Chinese.

This man is the founder of XIAOMI, who just announced that he will spend 10 billion US dollars to build a car within ten years: Lei Jun.

In 2013, he visited Tesla, Inc. 's headquarters twice. After a test drive of the Model S, he ordered four, two for himself, and two for Yu Yongfu, chairman of the UC browser, and Fu Sheng, chairman of Cheetah Mobile Inc. After returning home, he wrote a thousand-word essay on that Musk was a cool classmate and compared the Model S to "four wheels plus two iPad".

Not only Lei Jun, but many people think that Model S has similarities and differences with the iPhone of that year, on the one hand, because the large screen and virtual dashboard have changed the traditional way of interaction, and on the other hand, because it can achieve wireless upgrade (OTA) in the air like iPhone.

In order to understand why, IHS Markit, an American market research firm, bought a Model S out of its own pocket for $80, 000, and broke it down like a mobile phone for 800 yuan, discovering a lot of secrets.

For example, NVIDIA Corp's Tegra 3 chip is used to control the big screen, which makes its computing power comparable to that of mobile phones and tablets at that time; the big screen is made by Taiwan's TPK, which is the exclusive supplier of the first generation of iPhone. In addition, Tesla, Inc. did not completely outsource the research and development of automotive electronics to third parties like other manufacturers, but through deep participation in customization to control costs and create differences as much as possible.

All this makes people believe that the success of Model S lies not only in hardware and physical performance, but also in software and algorithm innovation, very "Silicon Valley". Since then, Tesla, Inc. has been known as the "Apple Inc of the automobile industry".

After the great success of Model S, Apple Inc announced the "iOS in the Car" plan, later the CarPlay car system, at the Global developer Conference (WWDC) in 2013, but at that time, it was not known that Apple Inc had quietly launched the "Titan" project internally, ready to challenge the authority of Detroit at any time.

However, with the progress of the project, there were huge differences within Apple Inc. Some people thought that cars should be built across borders, while others firmly opposed it. They believed that they should focus on self-driving software and take the route of light assets. The two sides reached a stalemate, and slowly, there was less and less news about Titan.

During the years when Titan was out of public view, Tesla, Inc. became the brightest star in the automobile industry, especially after crossing the production hell of Model 3 and successfully building a factory in Shanghai, his market capitalization surpassed GM, Volkswagen and Toyota to become the first in the world, and Musk even became the richest man in the world at one time.

Not only that, Tesla, Inc. 's largest "long", known as "financing for God" woman, ARK Capital CEO Katherine Wood recently even threatened that by 2025, Tesla, Inc. 's share price will reach 3000 US dollars, the market capitalization will reach an astonishing 3 trillion US dollars, Apple Inc's market capitalization is about 2.1 trillion US dollars.

Perhaps smelling the crisis (or business opportunity), Apple Inc's car-building speed increased significantly last year, and all kinds of important news gradually came out, such as asking Taiwan suppliers to stock, and Taiwan Semiconductor Manufacturing Co Ltd to jointly develop self-driving chips. the signing of a 4 trillion won contract manufacturing agreement with Hyundai Korea shows that the barbarians at the door are really coming.

If Apple Inc really wants to enter vehicle manufacturing, it means that the emerging Internet consumer electronics and century-old automobile industry will usher in an epic convergence. Will Cook "reinvent the car" like Jobs did when he launched the iPhone? What is the real driving force behind mobile phone giants dabbling in cars? Is the transformation of traditional car companies the only way to become Apple Inc?

These are the questions that this article tries to answer.

01. The same river?

Apple Inc stepped on the bones of Nokia to get where he is today.

At the MacWorld conference in 2007, Jobs unveiled the first generation of iPhone, without a dense physical keyboard or stylus, and replaced it with a 3.5-inch multi-touch screen, which some geeks called the "Jesus phone." However, iPhone was not favored at first because of its pricing of $499 and the adoption of a series of new technologies.

Mr Ballmer, former CEO of Microsoft Corp, predicted that "Apple Inc can only win 2 to 3 per cent of the market at most", mocked CEO before Palm. "Brewers definitely can't make coffee. The implication is that computer makers can't make mobile phones, and Motorola's former CTO said with disdain, "iPhone doesn't have any revolutionary technology." Fortune magazine commented: "Nokia N95 is an undisputed winner in front of iPhone."

At that time, the global mobile phone giant was Nokia, and by its standards, the first generation of iPhone was not competitive at all because of its unstable call quality, inability to copy and paste, no MMS, and no 3G support. In addition, compared with Nokia's "unbreakable body", iPhone failed the "drop test" (throwing the phone from 1.5m high to the concrete floor from five different angles).

But even so, iPhone sold 1.8 million units in the first quarter of 2008, compared with 115 million for Nokia, but more people believe that Apple Inc represents the future. In May of that year, the market capitalization of Apple Inc surpassed that of Nokia for the first time. Since then, the gap between the two sides has widened. Nokia felt an unprecedented crisis and hastily launched the 5800 based on Symbian in 2009, which was called the "iPhone killer" by some media at that time.

But the reality gave Nokia a slap in the face.

In early 2009, Risto Siilasmaa, now Nokia CEO, made a trip to New York and found that the Nokia flagship store on Fifth Avenue was so empty that you could quietly hear a pin drop, but the nearby Apple Inc store was packed with people when he thought the Imperial Building was starting to tilt.

Li Situo later concluded that it was not Silicon Valley engineers who defeated Nokia, but that Nokia's traditional leadership style followed the inherent thinking of the hardware era, that is, hardware is king, operating system and software capabilities are only supporting roles. That's why Nokia plummeted and was eventually acquired by Microsoft Corp, leaving behind chicken feathers.

Once upon a time, wrestling resistance, battery life, call quality, appearance color, variety and cost control were the main themes of competition in the mobile phone industry, but in the new era of mobile Internet ushered in by iPhone, everything had to be reversed and started all over again, with users pursuing not only better hardware, but also constantly updated software and a more silky and consistent operating experience.

Nokia, which is huge and bloated, obviously failed to cope with this. After giving up the backward, complex and bulky Symbian, Nokia turned to work with Intel Corp to develop MeGoo system, but it was not long before it was dismissed as "worthless" by the new CEO Elop, and then gradually lost, giving up Android and embracing Microsoft Corp, becoming the biggest failure.

How similar the auto industry is.

Although Model S was rated 99 by Consumer report when it was launched, there were many imperfections like the first iPhone, because before long, people found that Model S had a lot of minor problems, including door handles, seats, bumpers and screen crashes, and so on. Even now, people are still complaining about Tesla, Inc. 's workmanship, but it is undeniable that the emergence of Model S does make people rethink the definition of a car.

Over the past hundred years, a car has been nothing more than an oil-driven vehicle responsible for transporting people from point A to point B, that is, the so-called "Feature Vehicle". However, with the development of the Internet and the application of more intelligent technology, cars are changing to "Smart Vehicle" (smart cars).

Mary Bora, chairman of General Motors Co, said in 2015 that "more changes will take place in the automotive industry in the next 5-10 years than in the past 50 years". One of the key inherent logic is that the automotive industry has moved from "mechanical definition" and "electronic and electrical definition" to a new era of "software definition".

The mechanical definition means that the hardware is dominant, that is to say, the whole car factory buys parts from different suppliers and assembles them together, just as "buy horses in the east, saddles in the west, head in the south, and bullwhip in the north". Finally, under the condition of ensuring a certain profit, it will be sold with one hammer.

In the electronic and electrical architecture, it is more dependent on stacking. At present, the number of ECU of an ordinary car is about 70-80, luxury cars may even exceed 100, the wire harness can be up to 5 kilometers long, they are distributed in different parts, corresponding to different functions.

There seems to be more and more features, but the problem is that these software are not developed by car manufacturers, but rely on Tier 1 suppliers such as Bosch, mainland and Ambofo. Different ECU have completely different underlying code, vehicle manufacturers do not have the authority to maintain and update, and collaboration is expensive. To use Jobs' analogy, these ECU are like "baby software", far from powerful enough.

What's more, the distributed architecture of the automobile is just like the Symbian system at that time, because of the deep coupling between software and hardware, a button corresponds to a function, which can not achieve OTA, let alone create the so-called open ecology.

So, what is a software-defined car?

First of all, it does not mean that the role of code is above all else, nor can it be simply understood that the car of the future is "four wheels + a computer / iPhone". What is more reflected behind it is that the car has gradually transformed from a highly mechatronic vehicle into an intelligent, sustainable upgrading iterative mobile terminal, which is a new high-tech species. What are the forces driving this change?

02. New "Moore's Law" of vehicle Chip

The development of the mobile phone industry is inseparable from one of the founders of Intel Corp: Gordon Moore, who proposed the classic "Moore's Law" (the number of transistors that can hold on an integrated circuit doubles every 18 months). Although it has been more than 50 years since this experience was proposed, it is still working.

For example, the A14 chip used in iPhone 12 can perform 11 trillion AI operations per second, 10 times more than the previous generation.

Unfortunately, the automobile industry does not conform to Moore's Law. Intel Corp's former CEO Ke Zaiqi once said that if the automobile industry obeys Moore's Law, then the Volkswagen Beetle launched in 1971 can drive 480000 kilometers per hour, and a gallon of gasoline can run 3.22 million kilometers, and it only costs 4 cents.

The fundamental reason for this difference is that what drives the progress of mobile phones is the computing power of chips, while in the era of internal combustion engines, it is the engine, chassis and gearbox that determine whether a car is good or bad. because they determine fuel consumption, push back and comfort, but in the past 60 years, the thermal efficiency of the internal combustion engine has increased by only about 10%, and the speed is extremely slow.

In fact, as a typical manufacturing industry, the automobile industry obeys the "Wright Law", that is, every doubling of output reduces prices by about 15%. The more classic examples are the Ford Model T, the best-selling fuel car in the 20th century, and Tesla, Inc. Model 3, the best-selling electric car so far in the 21st century.

The popularity of electric vehicles is inseparable from the progress of the battery industry. As the most expensive component, the cost of the power battery system was less than 1 yuan / wh last year, and the lithium iron phosphate battery has dropped to 0.6 yuan / wh, down more than 85% from a decade ago, but the decline is not unlimited. The next generation of revolutionary solid-state batteries will take at least three years to be truly commercialized, and it is not realistic to completely eliminate "mileage anxiety" in the short term.

In addition to improving battery life and increasing charging stations, another magic weapon that can bring electric cars into millions of households is the improvement of intelligence, which is backed by data, chips and algorithms, as one industry expert said, "in the new energy era, data is transformed into decision-making and control of vehicles through computing, and the chip is undoubtedly the digital engine of the car. "

The trend reflected by this is that competition in the auto industry has evolved from "horsepower" in the last century to "computing power" today (1 TOPS represents a processor that can perform a trillion operations per second). Judging from the evolutionary history of the past decade, car computing is advancing along the new "Moore's Law".

Take the Mobileye under Intel Corp as an example, the EyeQ2 chip launched in 2010 has only 0.026TOPS computing power, the EyeQ3 chip released in 2014 has the computing power of 0.256TOPS, and it has reached 2.5TOPS in 2018. The highest computing power of the EyeQ5 chip that is planned to get on the car this year can reach 24TOPS, and almost every generation of chips has as much as ten times the computing power of the previous generation.

It is precisely because it is aware of the value of chips that the automotive industry has set off an arms race in computing power. For example, Great Wall Motor declared that the highest computing power of the "coffee intelligence" platform can reach 700+TOPS, the car ET7 released by NIO Inc. at the beginning of the year plans to carry four NVIDIA Corp Orin chips, with a total computing power of 1016TOPS, and the Weima W6 is known as "cloud computing power" that can reach millions of TOPS.

But this does not mean that the greater the computing power, the better, and the real moat and competitiveness cannot be formed only by stacking hardware. how much computing power can really be used depends on the cooperation between different sensors, software and algorithms, that is to say, we need to take the road of deep and full-stack self-research.

The so-called full-stack self-research is to master a series of core links, including visual perception, sensor fusion, decision-making, planning and control, so as to establish a complete data channel, instead of directly copying and pasting the supply chain scheme, which can form a data closed loop through OTA, control the rhythm of research and development, and form a real moat.

Because of the need for strong capital and human resources, looking at the world, there are only a handful of manufacturers who are really doing self-research, represented by Tesla, Inc. abroad and XPeng Inc. and SAIC at home.

SAIC has also made a careful layout. On the one hand, Zero Beam Branch was set up last year to focus on core areas such as the SOA software platform, a new generation of centralized electronic and electrical architecture, cloud services, and algorithm chips, and the number of staff will be expanded to 2000 next year. In addition, SAIC passenger cars also set up a "smart driving center" last year, and in the next five years, it will provide more than 10 billion yuan of financial support to the intelligent driving center, and set up a team of more than 1000 people to fully cover the software and hardware development and algorithm research needed for intelligent driving.

In the track of smart electric vehicles, Tesla, Inc. is undoubtedly the forerunner, but it is not perfect. In the past two years, traditional vehicle manufacturers are waking up and their supply chains are maturing. They are struggling to catch up with their accumulated experience and differentiated technical routes over the past few decades, just as Samsung, XIAOMI, Huawei, OV and other manufacturers began to take turns after the launch of iPhone.

03. How do double excellent students come into being

Dismantling a car is the best and most direct way to judge the determination and competitiveness of a company's transformation to electrification.

So someone really did it. Some time ago, the UBS team that dismantled Tesla, Inc. Model 3 released a report on the dismantling of Volkswagen ID.3. The car was chosen because the ID.3 was the first product based on Volkswagen's pure electric modular platform MEB, and soon after it hit the market, it beat Model 3 to become the top seller of electric cars in Europe.

The UBS team believes that ID.3 is as good as Tesla, Inc. in terms of battery life and cost control, but there is still a gap with Tesla, Inc. in software architecture, autopilot and AI learning ability. They also believe that traditional vehicle factories and suppliers have the ability to fully participate in the electric vehicle industry while ensuring profits, and make a comparison: "if Tesla, Inc. is Apple Inc, then Volkswagen is Samsung, with large sales and strong profits."

With this remark, Volkswagen's share price rose five points and its market capitalization exceeded the 100 billion euro mark for the first time, and Volkswagen Group CEO Herbert Dis was delighted to forward the good news on social media for the first time. Volkswagen's market capitalization now stands at 150 billion euros, 2.5 times that of the same period last year.

Samsung seems to be a good model for success when it comes to transformation. In the functional phone era, Samsung was once the second largest in the world after Nokia, while in the smartphone era, Samsung became the world's number one, shipping 266 million units last year, with a market share of 20.6%.

Looking back at the ups and downs experienced by the mobile phone industry over the past decade, we will find that the brands that eventually stand out are at least double excellent students, and some even three good students. In contrast, those declining companies may be more prominent in some aspects, such as Meizu design, hammer OS, Sony Group Corp waterproof, LG screen, but in the end did not go far.

The inspiration that this brings to the automobile industry is: in the stage of the transformation of the old and new potential energy of the industry, we should promote our strengths and make up for our weaknesses, and strive to become an all-round player.

For example, Volkswagen, as the world's largest automobile industry group, is well aware that the company's weakness lies in software and algorithms, but he dared to propose to transform into a software-driven company in 2019, and set up a car.software department to this end, and often beat employees with the example of Nokia in internal speeches, hoping that they can speed up the pace of transformation.

But at the same time, Deiss knows that traditional car manufacturers like Volkswagen have advantages over new forces in terms of capital, technology, talent, supply chain and production management, which are the foundation of the company.

At a time of great changes in the automobile industry, it is easy to imagine the traditional car factory as an image of being old-fashioned and unenterprising. It seems that all new concepts are worth touting and all old experiences need to be knocked down. But this idea is obviously one-sided and extreme. Any industry has underlying logic and rules, which are difficult to shake and subvert.

Li Shufu, chairman of Geely Holdings, wrote a ten-thousand-word letter this year, which mentioned: "Today, the electric car industry in the world is competing with each other, and everyone is living a good life driven by the capital carnival dance." However, I think the basic law of the automobile industry will not change, that is, scale, in the end, the number of global automobile industry enterprises will not be too many, only large-scale enterprises can survive. "

China has been the world's largest auto producer since 2009, and SAIC is the leader in five companies that produce more than one million vehicles a year.

Last year, SAIC sold more than 5.6 million vehicles, ranking first in the country for 15 years in a row, with a revenue of more than 740 billion yuan and a net profit of about 20 billion yuan. moreover, under the tremendous pressure of the epidemic, it achieved counter-trend growth overseas, and the export scale of complete vehicles reached 390000 vehicles. First in the country for five years in a row.

In other words, SAIC should be the best car builder in the country, but in the new era, this is not enough.

In order to meet the new challenges, Chen Hong, chairman of SAIC, put forward the four strategies of "electrification, intelligent networking, sharing and internationalization" as early as 2014, and laid out in advance in many key areas, forming a strong interest community.

For example, in the field of battery and IGBT, it successively established joint ventures with Ningde era and Infineon industry leaders to achieve strategic positions; it also walked in the forefront of the times in the integration with the Internet industry, established a joint venture with BABA to build a zebra system, and took the lead in launching the world's first Internet car, Roewe RX5, and introduced Internet ecology into the car for the first time.

In February this year, R Automobile, a subsidiary of SAIC passenger cars, launched the world's first 5G car, the MARVEL R, which took a key step in the 5G-V2X field. However, as the vanguard of promoting the upward innovation and upgrading of the group's own brand, it is clear that getting on the car brings more surprises to the industry than that.

04. Wait for the flowers to bloom

In 1928, two American researchers conducted an experiment on corn hybrid seeds in Iowa. They handed out improved seeds to more than 200 farmers and observed and recorded their acceptance. It was found that only 25 people were using new seeds in 1933, and the number of users continued to increase, reaching 159 in 1937.

In 1962, communication expert Everett M.Rogers proposed the famous "innovation diffusion curve" based on the experimental results. This theory has a strong guiding significance for industry and corporate research, and can perfectly adapt to the process of public acceptance of a new idea or technology, such as electric cars.

The first generation of electric car products like Model S attract more "technology control" and "enthusiasts". They focus on new features and experiences, so they can put up with some obvious shortcomings and defects, but as the market matures, people need not only to taste fresh, but also to become more pragmatic, focusing on a series of factors, including battery life, brand, workmanship, after-sales and price.

Although China has been the world's largest market for new energy vehicles for the past five years, until 2020, electric vehicles accounted for only about 4% of total new car sales, which is still a minority, and there is still a huge gap between China and the mass market. So, how to bridge this gap?

Only better close to the needs of users.

Take the Wuling Hongguang MINIEV, which has a "sinking dividend" as an example. The price of this model starts from less than 30, 000, and the sales of its unattractive A00 class electric car have been rising since it went on sale in July last year, reaching 36000 in January this year and has sold more than 200,000 in nine months.

For a long time, the automotive industry has been dominated by "engineer thinking", more from the professional perspective of automotive engineering to design and produce, pursuing the advanced product in design, performance, function and efficiency, ignoring the actual needs of users, so user thinking has gradually become the consensus of the industry.

However, user thinking can not be reduced to a slogan, but also needs to be constantly verified and improved in practice.

Last month, R Automotive, a subsidiary of SAIC, launched its first technology brand, R-TECH High Energy Smart, which took five years to develop and invested 20 billion in research and development, which is the product of the user's mind.

On the one hand, it is equipped with SAIC's new zero-beam Galaxy full-stack solution. Based on the full stack software platform of SOA cloud management, plug and play is realized through AIoT hardware, and with the help of open and rich ecology, it can bring users a real intelligent and personalized experience of "thousands of people, thousands of cars and thousands of faces".

On the other hand, in the intelligent driving that has attracted much attention, taking full account of the complexity of Chinese road scenes and the ever-changing needs of users, the Shanghai Automobile Intelligent driving Center has developed the world's first high-level intelligent driving scheme, PP-CEM ™, constructed a "six-fold integrated perception system" and carried out full-stack self-research.

Perception is one of the three core functions of autopilot, which is equivalent to human facial features, which is responsible for perceiving the surrounding environment, collecting data and transmitting it to the decision-making layer, which is equivalent to the brain, responsible for processing the data, and then outputting the corresponding instructions. the executive layer is equivalent to the limbs and is responsible for executing the instructions of the brain.

At the perception level, there are two different technical routes in the market at present. Tesla, Inc. adheres to the pure vision + algorithm scheme, and most other vehicle manufacturers take lidar as a necessary item of perception to deal with more corner cases (extreme emergencies).

The reason why PP-CEM ™can have six senses lies in a series of software and hardware enhancements, including the top lidar in the industry, the Premium 4D imaging radar that can see the can of cola 140m away, the NVIDIA Corp Orin chip that can provide 500-1000TOPS, and so on.

If we compare it with the mobile phone industry, the auto industry is in the first stage of switching from fuel cars to smart electric vehicles, which means that in addition to constantly solving some technical problems, such as battery life and safety, and launching popular products, the whole car factory also needs to create a corresponding ecology, just like the rise of Apple Inc and Samsung can not be separated from iOS and Android.

With the increasing expansion of the wave of the new four modernizations of the automobile, the connotation and extension of the automobile industry are constantly expanding. in the past two years, the relationship between automobile companies and technology enterprises is getting closer and closer, and there are more and more examples of marriage, such as Baidu, Inc. and Geely. Huawei and Chang'an, BYD and DiDi Global Inc. and so on.

According to SAIC passenger cars, the cooperation between car companies and technology companies should not be limited to the purchasing relationship between OEM and Tier 1, but need to climb another tall building to create new intelligent species based on user needs.

To this end, SAIC passenger cars gathered Luminar, NVIDIA Corp, Zaifu, Huawei, Zebra Zhixing, Microsoft Corp, Tencent Wisdom Travel, State Grid, Southern Power Grid, Star charging, Special call, BOE, Olympic New Energy and other leading enterprises to form the most powerful world-class co-creator strategic alliance in history, sounding the strongest assembly number in the intelligent electric era.

05. The end.

The entry of Apple Inc and XIAOMI has pushed the car-building craze in recent years to a new and most exciting part, and has also added more variables to this restless industry. How to deal with it?

Chen Hong, chairman of SAIC, said at an internal work conference this year, "in a world full of changes and uncertainties, if we follow the old rules and follow the rules, we will never be able to find a 'new world' with an 'old map'. "

The so-called old map means that in the era of internal combustion engines, the rules of the game of the global automobile industry are in the hands of foreigners, and China has been catching up step by step, although it has not achieved a real breakthrough. however, it has accumulated very rich experience in the fields of policy formulation, manufacturing, supply chain management, personnel training, brand marketing and so on.

In the new energy era, engines and gearboxes are abandoned, and the technical barriers erected by foreign giants no longer seem to be an insurmountable natural danger, but it is by no means easy for China to make a historic leap and enter the "new world". We need to make continuous breakthroughs in the "four modernizations".

In terms of electrification, at present, China's power batteries are in the first echelon in the world in terms of installed capacity and technological leadership, and in the degree of intelligence, although Tesla, Inc. is temporarily in the lead by virtue of revolutionary electronic and electrical architecture and super algorithms, but Chinese local manufacturers are also struggling to catch up through technology self-research and various forms of alliances.

For Chinese Autobots, if there is not an energy revolution, the possibility of overtaking is basically zero, but the times have given us such a chance, we must not waste it. A small step for smart cars is a big step for China's automobile industry. This is not in line with the historical process, this is the historical process itself.