The new forces are leading the change and are most favored by the secondary market; the pan-Internet faction occupies the high ground of traffic and is good at technology migration; the traditional car companies have a solid foundation and their brand reputation is well-known.

Even a certain real estate group with deep pockets has made a bold statement - striving to become the world's largest and most powerful new energy vehicle group in 3-5 years.

Just like the martial arts competition at Mount Huashan, all the heroes were ready for battle, and all forces held their yellow turbans high and shot arrows left and right.

Look, there is the industry iron curtain in front, a desperate mentality in the middle, and a treacherous world behind. Where can a newbie like you stay?

Even so, after the Shanghai Auto Show a month ago, which was still hot, the rarely mentioned AI vision companies still squeezed in.

After getting used to the giants’ aggressive moves, we find that the entry of AI vision companies requires a gradual and orderly process with a beginning, a middle, a middle and an end.

Their quiet entry has also added several new stories to the field of intelligent driving.

Hikvision: self-development with the left hand and investment with the right hand

Hikvision, the big brother of AI security, has been deeply involved in the intelligent driving market while maintaining its usual low-key style.

Its preparation for intelligent driving began in 2015, when Hikvision planned to develop new businesses internally. Initially, three businesses were identified: Hikvision Automotive Electronics, Hikvision Robotics, and Hikvision EZVIZ.

In July 2016, Hikvision Automotive Technology, with an investment of 150 million yuan, was officially established.

Before and after this, Hikvision also invested in Weishi Automotive Technology in June 2016 and established Hikvision Automotive Software in July 2017.

2018 was a year of growth for Hikvision’s intelligent driving, with breakthroughs in market channels and technological research and development.

In February 2018, they launched the advanced driver assistance system and automatic parking APA+, and in the same year they successfully integrated it into the configuration of the 2019 Porsche Cayenne.

The automobile industry is based on stability, with a long supply chain and complex interests. It is not easy for new entrants to break into the market, but Hikvision has unexpectedly entered the high-end market.

Data shows that as of the end of 2018, Hikvision Auto has passed the audits of 20 OEMs and become their qualified supplier. The company's major customers include FAW Group, Beijing Auto, SAIC Roewe, SAIC MG, Honda Motor, etc.

Among them, there are more than 200 designated projects, more than 100 projects have been put into mass production, covering 500 channel partners.

In addition to establishing subsidiaries for independent research and development, investment is also a major path that Hikvision is more interested in.

Before establishing the automotive electronics company, Hikvision acquired a stake in millimeter-wave radar company Sensortech in 2016 and became the latter's second largest shareholder.

Founded in 2013, Sensortech is not only one of the first pioneers in millimeter-wave radar, but also one of the leading companies with outstanding performance.

Qin Yi, the founder of Sensortech, is a radar expert who returned from the UK and has been engaged in radar research and development and manufacturing in the UK for more than ten years.

It is reported that 80% of the team members gathered by Senstech have military backgrounds and master almost all core technologies such as radar hardware, software and mass production processes.

It is reported that Sensortech's millimeter-wave radar has established R&D centers in Beijing and Shijiazhuang, a general factory in Wuhu, and an automotive business unit in Hangzhou.

Shijiazhuang is known as the base camp of military radars. It has a large concentration of military and civilian millimeter-wave radar R&D talents. The 54th and 13th radar R&D institutes of China Electronics Technology Group Corporation are both in Shijiazhuang.

Sensata also performed quite well.

In 2019, Si-Tech achieved the localization of large-scale 77GHz automotive millimeter-wave radar for the first time, breaking the monopoly of international giants.

Sensortech's 77GHz millimeter-wave radar has become the first ADAS millimeter-wave radar sensor in China to be truly "on the road".

At present, Sensortech has become a supplier within the systems of domestic and foreign automobile companies such as Hongqi, FAW, Hyundai Motor, Dongfeng Nissan, Great Wall, and Changan.

The high-resolution millimeter-wave imaging radar + visual fusion technology jointly developed by Hikvision and Sensing may compete with low-beam lidar.

Dahua Technology: Focusing on complete vehicles,

Three-electric, Internet-connected, and autonomous driving are launched simultaneously

Leapmotor was born out of the automotive department of Dahua Technology, and after becoming independent, it received technical and financial support from Dahua Technology.

In 2015, Zhu Jiangming, Vice Chairman of Dahua Technology and CTO of Dahua Technology, personally founded Leapmotor.

After experiencing a sharp decline in new energy subsidies in 2019, many new car-making companies have encountered serious funding problems and their monetization is in doubt.

Leapmotor is no exception.

In 2018, after Leapmotor suffered a loss of 307 million yuan, it continued to lose about 200 million yuan in the first half of 2019.

On January 4, 2019, Leapmotor's first model, the S01, was launched on the market. Approximately 1,000 units of the car were delivered in 2019.

Regarding Leapmotor, which has been suffering from continuous losses, negative comments have been brewing online.

Zhu Jiangming said, "Even without financing, Leapmotor can survive for another three years." He revealed that Dahua Technology will continue to provide funds for Leapmotor, "Of course, we hope to get more financing and develop faster."

After experiencing financing difficulties, at the beginning of 2021, Leapmotor officially announced a financing of 4.3 billion yuan, including the Hefei government investment platform.

At the beginning of this year, the Hefei Municipal Government, which had previously invested in NIO, signed a strategic cooperation agreement with Leapmotor. In the future, Hefei will invest approximately 2 billion yuan in Leapmotor's Series B financing and launch more cooperation.

In terms of cash flow, the company was never favored by the industry, but with the arrival of huge amounts of financing, the market seemed to see possibilities again.

On the technical level, Leapmotor said it has independently developed three core technologies: three-electric system, intelligent network system, and autonomous driving system, and has fully mastered the core hardware platform and algorithm technology of autonomous driving, achieving full autonomous coverage of key technologies in the perception, decision-making, and execution layers of autonomous driving.

At the product level, Leapmotor currently has three mass-produced models, namely: Leapmotor T03, Leapmotor S01 and Leapmotor C11.

The three products have different styles and sales volumes.

In 2020, according to official news from Leapmotor, the cumulative sales in 2020 reached 11,391 units, of which T03 was the main force, contributing 10,266 units.

Founder Zhu Jiangming is also very confident: "In 2023, Leapmotor will enter the top 3 of new car-making forces, and its market share in the domestic new energy vehicle market will reach 10% in 2025."

SenseTime: Pursuing advanced sensing technology, both inside and outside the cabin

Compared with other AI unicorns, SenseTime started its layout in autonomous driving earlier and more comprehensively.

SenseTime entered the field of autonomous driving in 2017. Its automotive industry layout can be divided into two major levels: in-cabin (smart cabin) and out-of-cabin (smart driving).

The smart cabin layer is based on the pre-installed mass production solution and uses visual perception technology as an anchor point, covering multiple scenarios from getting in the car to using the car from point to surface.

SenseTime’s SenseAuto Cabin smart cabin solution includes driver perception system, cabin perception system, smart entry and other functions.

It is reported that in the past two years, SenseTime has won designated mass production projects for smart cabins from more than 30 leading domestic and foreign partners, covering a total of more than 13 million vehicles, of which more than 10 projects have achieved mass production and delivery.

In terms of intelligent driving, SenseTime chooses to cooperate with OEMs and become a solution provider for automobile manufacturers (OEM) and Tier 1 suppliers.

Among the three major elements of autonomous driving perception, decision-making and execution, automobile manufacturers and Tier 1 play an important role.

In 2017, SenseTime signed a five-year long-term cooperation agreement with OEM manufacturer Honda to develop L4 autonomous driving solutions suitable for passenger car scenarios.

In 2018, SenseTime completed autonomous driving without takeover in semi-open venues in Hangzhou and Shanghai. In 2019, the "AI Autonomous Driving Park" was established in Japan, which will be used for the research and development and testing of autonomous vehicles and will be open to the public.

SenseTime’s autonomous driving business is positioned as vision-based, with other elements as supplementary.

In addition to vision, SenseTime has technical reserves in high-precision maps, lidar, millimeter-wave radar, etc.

By combining a variety of different sensors, technical functions such as perception, analysis and prediction, decision-making planning and control, city-level three-dimensional map reconstruction and high-precision positioning capabilities for unmanned vehicles can be achieved.

At present, SenseTime has iterated its autonomous driving technology many times and formed a relatively mature intelligent driving solution: SenseAuto Pilot intelligent driving solution, focusing on innovations from L2+ advanced assisted driving to L4 autonomous driving, and released the SenseAuto Pilot-P driving navigation solution for the first time at the Shanghai Auto Show.

In addition to software, in March 2019, SenseTime also launched its first original robot, the SenseRover X autonomous driving car, which is a teaching product for autonomous driving.

Obbec: Strategic investment + self-development, walking on two legs

Orbbec is one of the AI startups that has invested the most in smart cars.

As an AI 3D sensing technology solution provider, Orbbec, founded in 2013, has been deeply engaged in the 3D sensing field for nearly 8 years.

As the core visual perception technology in the field of artificial intelligence, 3D sensing integrates multiple interdisciplinary technologies such as chips, algorithms, optics, and software. It is the most core technical carrier for perception and recognition, new human-computer interaction, and other fields in the era of artificial intelligence.

In addition to 3D structured light, Orbbec also has long-term plans in mainstream 3D vision perception technology fields such as binocular, iTOF, dTOF, and lidar.

As early as 2018, Orbbec invested in Feixin Electronics, a lidar chip-level solution provider.

Founded in 2016, Feixin Electronics is a high-tech enterprise focusing on optoelectronic equipment, lidar research and development, and integrated circuit design.

Only two years after its establishment, Feixin Electronics received investment from Bosch and others.

It is reported that Feixin Electronics' main business is the research and development, and production of lidar systems and core chips, and its customer base is mainly domestic and foreign automobile, robot, drone and other production and development manufacturers.

Feixin Electronics said that to address the industry pain points, it has adopted continuous wave carrier modulation or coherent heterodyne detection solutions, and used focal plane point cloud ranging technology to meet higher spatial resolution and a larger field of view. The detection distance can exceed 200m, and there is no need for complex and expensive mechanical scanning devices, which continuously improves system reliability and makes the obtained images clearer.

In April 2019, Orbbec established Orida, a provider of automotive 3D vision sensing solutions.

Oruida's business focus is on smart cockpits, and its products include ToF camera modules, lidar and other hardware, as well as 3D ToF smart cockpit solutions.

Inheriting Orbbec's 3D visual perception technology, Orida can bring a variety of 3D intelligent functions such as DMS, OMS, gesture recognition, face recognition, and identity authentication to smart cars.

Its financial-grade secure 3D face recognition solution protects the information security of drivers and passengers; multi-zone gesture control is achieved through the 3D-ToF camera; at the same time, smart cars can also use 3D information to determine the body shape of drivers and passengers, their position in the cabin, etc.

Recently, Oruida also released a 3D ToF smart cockpit solution tailored for smart cars.

ArcSoft: Focusing on the interior of the cabin, taking the path of software and hardware integration

In 2018, in response to the peak and saturation of the mobile phone market, ArcSoft officially expanded its business from the smartphone field to smart cars, IoT and other fields, making a horizontal breakthrough into the autonomous driving market in one fell swoop.

Deng Hui, founder and CEO of ArcSoft, once said that in the future every car will have more than 10 cameras, and the smart cockpit will become a key application scenario for intelligent driving vision AI.

Similar to its mobile phone positioning, ArcSoft’s smart car adopts an integrated hardware and software solution, striving to become a one-stop solution provider for in-vehicle vision.

According to the prospectus, as of the end of 2018, ArcSoft's business revenue from "automobile and other IoT products" was only 3.6795 million yuan, accounting for less than 1%.

Unlike most vision companies that add technologies such as LiDAR, ArcSoft's autonomous driving solution is completely based on vision, with the core focus on in-vehicle intelligence.

ArcSoft's intelligent driving vision solutions include in-vehicle safety driving warning, driver identity recognition, in-vehicle safety assistance, assisted driving warning, automatic parking and many other solutions.

In March 2019, ArcSoft acquired a stake in Kaiyi (Beijing) Technology, whose main businesses include active safety intelligent terminals (ADAS+DMS+face recognition), SDK software services and overall hardware solutions.

In 2019, ArcSoft was listed on the Science and Technology Innovation Board.

ArcSoft said that it has profound experience in the field of computer vision and has integrated its low-light high-contrast shooting, anti-shake and other image and video enhancement algorithm technologies. Even in special circumstances such as poor lighting inside the car, changing face angles, and vehicle shaking, it can also well perform functions such as vehicle surrounding environment monitoring and vehicle occupant monitoring.

After going public, ArcSoft has made great efforts to develop smart cars and other IoT smart devices, and the results are beginning to show.

According to ArcSoft, the smart car segment will begin mass production in 2019.

Data shows that in 2020, the intelligent driving vision solution business grew rapidly, with operating revenue of 65.9299 million yuan, a year-on-year increase of 310.61%.

It is reported that ArcSoft's intelligent driving related products include DMS (driver identification system), ADAS (advanced driver assistance system), BSD (blind spot detection system), OMS (passenger identification system), Interact (visual interaction system), Authenticate (biometric authentication), AVM (3D surround monitoring system), AR HUD (AR head-up display) and smart trunk, as well as other related software solutions based on core algorithms.

Data from Gaogong Intelligent Automobile Research Institute shows that the DMS (driver recognition system) algorithm business is the main source of revenue for its intelligent automobile business.

This year, ArcSoft revealed that its intelligent driving business has achieved the designated development of 37+7 pre-installed vehicle models (37 mass-production models and 7 pre-research models), mainly providing pure algorithms. The company directly signs contracts with Tier 1 or vehicle manufacturers, involving many domestic mainstream car companies (including new car-making forces) and some joint venture car companies.

DeepGlint: The earliest player to enter the market and grow together

Founded in 2013, GreenVision is one of the earliest AI vision companies and also one of the earliest AI vision companies to invest in autonomous driving.

That year, DeepGlint, together with Wu Gansha, director of Intel Research Institute, and Jiang Yan, leader of the National Intelligent Vehicle Future Challenge championship team, founded a company focusing on the field of autonomous driving - UISEE Technology.

In 2016, UISEE Technology was established in Beijing, and DeepGlint invested in UISEE Technology as an investor.

Over the past five years, UISEE Technology has been forging ahead amidst the turbulent tide.

At CES in January 2017, UISEE Technology introduced its driverless concept car “Urban Mobile Box” to the world. This model became the third driverless car in the world to win the Red Dot Design Award.

In the same year, the company took the lead in launching commercial driverless operations for the general public at Baiyun Airport in April and Raffles City Hangzhou in June respectively.

On January 21 this year, Hong Kong International Airport announced that the driverless logistics vehicles jointly developed by UISEE Technology and the Hong Kong International Airport Authority will replace human-driven trailers to undertake the task of transporting baggage between the airport and the Haitian Ferry Terminal, which means that their use at the airport has gradually increased.

In the past year, UISEE has established business cooperation with dozens of companies including Changan Minsheng Logistics, FAW Logistics, and BASF.

It is revealed that the software algorithm provided by UISEE has also been pre-installed in a certain domestic luxury brand car model and has helped the independent brand take the lead in launching the L3 level autonomous driving function. Last year, UISEE delivered hundreds of "AI drivers" and achieved a year-on-year growth of 150%.

Not long ago, UISEE Technology announced that it had completed a new round of financing totaling more than 1 billion yuan, in which UISEE Technology obtained participation from state capital.

UISEE Technology has been working hard and delving deeply into unmanned logistics, and its achievements include business layout in the field of unmanned logistics accounting for almost 70% of the domestic market.

Since its establishment in 2016, UISEE Technology has gone through numerous hardships and explored an optimal solution among the chess pieces as dense as stars. With the airport formula, it has been running desperately on the road of self-improvement.

GreenVision's road to autonomous driving is also moving further and further with UISEE Technology.

Megvii: Based on AI vision, providing a full set of in-vehicle solutions

In November 2018, Megvii publicly demonstrated its in-vehicle AI vision solution.

At that time, Megvii's solutions were mainly based on functions such as face unlocking, account switching, driver identification, and multimodal interaction in the vehicle system and during driving, and it charged corresponding software usage fees and service fees.

"Face unlock" can capture the driver's face information through the camera outside the car and identify and confirm the identity, realizing face unlocking and temporary authorized face unlocking;

The camera inside the car can be used to start the engine and safe by scanning the face. The "account switching" function can accurately identify the driver's identity through facial recognition without perception, and cooperate with the in-vehicle intelligent system to quickly adjust the user's preset personalized vehicle configurations (seat position, mirror angle, air conditioning temperature, music, lighting, navigation, etc.).

The "Driver Identification System" can use the in-car camera to check the driver's driving status and behavior in real time, and trigger an early warning when the driver shows signs of fatigue or distracted driving, thereby ensuring driving safety.

Megvii once stated that it has achieved in-depth cooperation with NIO on future smart car applications. True commercial use of driverless cars is still a long way off, so Megvii focuses on understanding and assisting human drivers.

DeLu Deep Vision: Based on 3D vision camera,

Empowering the industry

DeLu Deep Vision's role in the field of smart cars is more about cooperating with third parties.

As a leader in the field of 3D vision, DeLu Deep Vision has been deeply engaged in technical fields such as high-precision depth perception imaging, 3D real-time high-precision reconstruction, 3D tracking recognition and perception for many years.

Last month, DeLu Deep Vision attended the 2021 Global Autonomous Driving Summit and demonstrated its latest 3D CV camera and its applications.

The error of Delu Deep Vision's two self-developed 3D CV cameras is less than 1mm within a range of 5 meters, which exceeds the indicators of international 3D camera giants, and the mass production yield rate is over 99%.

Based on the front-end low-power embedded platform, both cameras can achieve non-contact accurate recognition. Based on the principle of structured light, they can restore high-precision 3D detailed information of the face and accurately identify the identity of the person through the three-dimensional size information of the face. At the same time, the accuracy rate of two-dimensional and three-dimensional attack recognition is as high as 99.99%.

One more thing to mention: in terms of security, it can reach financial level.

It is reported that in addition to the field of smart cars, the two cameras are also being deployed in the fields of smart homes, financial payments, smart transportation, etc.

Intelligent driving: The second spring of AI vision

The entry of many AI vision companies into the smart driving track is not off-topic.

First, the layout of intelligent driving is a result of the outward-looking strategy.

Ever since computer vision escaped from the laboratory cage, AI security and autonomous driving have received "S cards" from a large number of investors.

When AI was first put into use, security provided an excellent breeding ground for AI companies to integrate technology and industry.

During this period, AI and security achieved mutual success:

HaiDaYu and other proud players in the security industry that have been exported to the world have almost dominated the global security market. The industry has expanded rapidly and spread to all areas of the city.

AI unicorns also started out in security and have gradually expanded into thousands of industries and into the entire domain.

On the left, AI security has become the main source of revenue, and on the right, AI security has gradually occupied a place. The question facing the entrants is how to maintain continuous vertical growth.

It is imperative to break away from path dependence and find markets beyond AI security.

If the path of AI vision companies in the past five years was "general AI SDK → re-customized integration project implementation" , then in the next five years, they can try the path of "standard market in non-standard field → forming standardized products → low-cost large-scale replication" .

Where are the standard markets in non-standard fields? Autonomous driving, medical treatment, and chips are among them.

Looking at the AI market, almost everyone in the sector is losing money. Jujinzhi believes that this has nothing to do with high labor costs, as losses are increasing; it also has nothing to do with hardware reserves, as OEM is possible.

The core issue is: AI security has not been standardized and project requirements are endless.

Then go to the standardized market? Someone asked.

The standardized market can reduce prices to infinitely low levels overnight, but the high operating expenses are beyond the ability of AI companies to afford.

Where can AI companies break through if they cannot enter the standardized market or the customized market? The answer is: find a standardized path in the non-standardized market.

In the field of autonomous driving, it is an obvious standard market in the non-standard field. Similar to AI security, smart driving start-ups also rely on capital input.

However, the former has fragmented scenarios and customized projects, and the road to product standardization is long; the latter uses smart cars as the carrier, and once the software-defined and human-machine collaboration technologies are formed, it will be a big hit in the world.

At present, many new forces in smart driving have achieved mass production of products and obtained a certain amount of cash flow.

Compared with all the competitors rushing to enter the market, the entry of AI vision seems a little late.

However, the smart car market is hot and the pattern is uncertain. The smart car industry chain is long and the subdivided fields are complex. You can say that AI vision is entering the market a little late at this time, but you cannot say that it does not have opportunities.

Second, autonomous driving may be the pinnacle of computer vision technology application.

In recent years, machine learning has continued to deepen, and computer vision applications have also made rapid progress.

Facial recognition, which has crossed thousands of mountains and rivers, is the most successful and fundamental part of AI.

Real AI is a long chain that runs through perception-decision-execution, which is particularly evident in autonomous driving.

The perception layer captures the vehicle's location information and external environment information through various hardware sensors;

The "brain" of the decision-making layer models the environment based on the information input by the perception layer, thereby forming an understanding of the overall situation and making decisions, and then sending execution signal instructions to the vehicle;

The final execution layer converts the signals from the decision layer into the car's action behaviors.

Autonomous driving technology is a combination of technologies from multiple fields, including artificial intelligence, high-performance chips, communication technology, sensor technology, vehicle control technology, and big data technology. The difficulty of its implementation has moved all kinds of AI.

Computer vision has numerous application scenarios, and autonomous driving is undoubtedly one of the most challenging and imaginative ones.

The closer the flowers grow to the top of the cliff, the more fascinating they are.

There has always been a dispute over the technical paths of AI vision and lidar in the field of environmental perception.

Regardless of which path is better, it has been brutally verified in the field of video IoT, and AI vision companies have also accumulated a lot of experience in AI technology reserves.

There are too many wolves and too little meat. How much can one eat to be full?

“Is autonomous driving a low-level industry? Everyone wants to get a piece of the pie.”

Those who have joined the game would probably feel particularly aggrieved after hearing this ridicule.

Most are stuck at the first hurdle: money.

The voice of “Don’t build cars without 20 billion yuan” was deafening, and the financial situation of the car-making star NIO once hit rock bottom.

Although AI vision companies, with the exception of Dahua's Leapmotor, are currently focused on smart driving hardware and systems, this is also an expensive business.

Many companies themselves rely on capital transfusions. Whether they have more funds and energy to participate in the battle of autonomous driving is a question they need to think about.

Industry barriers should not be underestimated.

The automobile industry has developed over a hundred years to form a rigorous and complete set of production processes and systems, and even spawned a set of industrial civilization based on safety, which cannot be subverted by latecomers in just a few years.

As the core embodiment of smart cars, autonomous driving technology is far from mature; there is also plenty of room for imagination in the intelligent experience within the cabin.

In other words, if cross-border players want to find their place in the world of smart cars, they must not only attach great importance to the topic of safety, but also have strong software capabilities.

But after the previous round of talent arms race between traditional leading OEMs and new car-making forces such as NIO, Xpeng, and Ideal, how can new players attract more professional talents? And how can they weigh the opinions and suggestions of talents from all over the world to make the final decision?

At the same time, the research and development of smart cars is not something that can be done successfully just by understanding software.

With the advent of electrification and intelligence, the threshold for car manufacturing seems to have been lowered a lot, but the internal and external problems encountered in this process may be far more than imagined.

Industry resources still need to be accumulated.

Compared with AI security, smart cities and other fields, AI vision cross-border players lack brand influence and channel resources in the field of smart cars, and their profitability is low in the short term.

Moreover, AI vision companies started their smart driving deployment at different times, and although their technologies have similarities, they are ultimately different. Compared with most vertical companies, they still have many shortcomings.

Therefore, it can be seen that in the past few years, even AI vision giants have been relatively cautious in their steps, mostly focusing on the in-cabin intelligence and ADAS markets.

If the giants’ cross-border business brings with it the ability to become hot searches, the brilliance of AI vision companies’ cross-border business has been somewhat dimmed.

The former has a wealthy background and enters the market with a top-tier experience card, looking down from a high vantage point, while the latter is more like a small boat, wading through the waves.

Of course, as technology advances day by day, resources accumulate, and revenue grows year by year, they will invest in areas including but not limited to R&D, marketing, capital, etc. There is no guarantee that this small boat will unexpectedly become a heavy cruise ship capable of long-distance voyages one day.

Don't say it's too late

The influx of numerous cross-border players into the smart car market has stimulated new vitality.

No matter from which perspective you look at it, the smart car market holds unlimited opportunities.

This market needs the existence of catfish.

In the trend of the new era, we certainly look forward to seeing more and more new players with strong strength entering the market and leaving a significant mark in the history of China's smart cars.

We also earnestly hope that this is a fertile land that can support the blooming of a hundred flowers and is full of new vitality and vigor, rather than a harbor for speculators who force growth.

With the first-mover advantage, many entrants may have temporarily ranked at the top of the industry, but with the continuous increase in the strength of all parties, it is not impossible for them to catch up.

Stay alert and keep growing.

Don’t say it’s too late. What you are not afraid of today may become your opponent tomorrow.

This article is originally written by Leifeng.com, and the author is Yu Kuai. Please reply "reprint" to apply for authorization. Reprinting without authorization is prohibited.

END