current position:Home>AI new century problem: the more you earn, the more you lose

AI new century problem: the more you earn, the more you lose

2022-05-07 13:46:1836kr

For the high-tech industry , The most important thing is the tuyere and the first year ,AI No exception .

Carrying people's fantasy of superhuman , The AI industry includes but is not limited to machine translation , Intelligent control , expert system , Robotics , Complex technical fields such as language and image understanding , It also makes AI From birth , Almost standing on the commanding height of Technology , Accept looking up . In this fantasy , From the initial flaunting of scientific and technological belief , Up to now, the emphasis on commercial landing , Ebb and flow , There are many participants in the industrial chain .

With BATH Led by traditional technology giants 、 The four little dragons are typical AI Algorithm service providers and Comview Dahua Co., Ltd Hkust xunfei And other comprehensive solution providers , Are the core participants in the industry , They compete in the upstream and downstream of the industrial chain , More layout . among , Once occupied a market share of up to 60% Of AI Four Little Dragons , Undertake more market attention , With AI Industry is cold , In the pursuit of listing , Compared with the technical “ The arms race ”, The ability of scene landing has become the core focus .

At the crossroads of fate , Their stories have also become more intriguing .

01 Jackie Chan in the wind

AI It is not the product of the new century , You may not know this .

The concept of artificial intelligence was formed in 20 century 50 years , It has experienced three waves in history .

20 century 50 years , The era of machine translation that focuses on logical reasoning , Robots and intelligent software are beginning to appear , But due to technical limitations , There was no bigger wave , People for AI The discussion is still shrouded in science fiction ; here we are 70 years , Artificial intelligence has entered the era of expert system , A large number of academic studies have sprung up , Although knowledge is constantly accumulating , But it has not been put into practice , The industry is also rapidly cooling down .

until 2006 year , Introduction of deep learning algorithm , People have just begun to enter the era of cognitive intelligence that attaches importance to data and autonomous learning . In the data 、 When the algorithm and computing power are mature , In the outbreak of artificial intelligence, technology began to land , Go deep into the application level , The era of entrepreneurship belonging to artificial intelligence is also quietly coming .

Now we know AI Four dragons , Founded on 2011 Years of absentmindedness , Is the big brother of the industry , The founder is from Tsinghua University “ Yao ban ” Young people of .“ At first we didn't think of starting a business , We want to apply computer vision technology to games .” In Tang Wenbin's memory , The participating game 《 Here comes the crow 》 That's where they start , It was inky who brought them to the research of face recognition technology .

2012 year , Zhu long, who is aware of the maturity of the application of computer vision in the industrial field, decided to start a business , I found it in Alibaba cloud Lin Chenxi, a high school classmate in charge of Technology , Yitu technology was officially established . Almost at the same time , Thomson technology Tang Xiaoou also met his partner Xu Li , It also starts from face recognition ,2014 The technology of Nianshang soup , Successively took Xiaomi 、 Huawei 、 Meitu Xiu Xiu And picture chat software FaceU、Snow Waiting for customers , Mainly engaged in B2B Business . as for Cloud comes from technology , It was the last of the four dragons to get on the runway .

2014 To 2016 Year , The four dragons led by Shangtang technology mainly invest in two places : One is “ Talent monopoly ”, The second is to build a hardware computing platform . Although at that time AI The field has not received attention yet , But for developing technology companies , Who has the technology , Who can lead , In this competition , Clouds seem to have inherent advantages , The founder was born in China University of science and technology , Most of the team comes from the Chinese Academy of Sciences , Clouds have always been seen as AI The national team in the industry .

The time came to 2018 year , The wave of artificial intelligence has really swept across China , Stand on the air outlet , Beardless AI The four little dragons also meet Jackie Chan , The influx of hot money in the capital market , Less than 3 Months , The four dragons have already got 100 The financing amount of more than 100 million yuan .

02 The foam subsided

Gartner Once proposed a technology maturity curve , Divide the mature evolution speed of new technology and the time required to reach maturity into 5 Stages , That is, the technology trigger period 、 Expected inflation period 、 The bottom stage of disillusionment 、 Enlightenment climbing period and plateau period . Now it seems ,AI The development of the industry , It also fits this curve .

It has gone through the entrepreneurial stage triggered by technology , The pronoun of artificial intelligence market has become “ Burn money ”, High-income 、 High gross profit margin and high loss , Has always been a AI The most typical characteristics of the industry , Although Tang Xiaoou, the founder of Shangtang technology, has responded to this , say “ We are not a money burning company , It's a profitable company , Be self-financing .”、“ Financing is not for burning , But doing great things .”

But what can't be ignored by the outside world is : The four dragons face almost similar problems in their continued losses . First, we need to maintain high investment in R & D and talent recruitment ; Second, maintain high investment in market expansion and regional expansion , And these two problems are also “ The higher the revenue , The more losses ” The immediate cause of , Accompany each other .

Labor costs are too high , On the one hand, scientific and technological research and development has high requirements for talents , Keep the labor cost high . For example, Shangtang technology was once known as Bell Labs in China , Because it has a very luxurious technical team , According to the prospectus , Shang Tang has 40 Famous professor ,5000 More than employees , About two-thirds of them are scientists and engineers , Include 250 More than doctors and doctoral candidates .

A similar situation , And more or less happened to the rest of them . When technology cannot be realized quickly , In a sense, the number of scientists is the guarantee of the company's product quality . In order to gain more trust from capital , In addition to the technical volume , The four dragons inevitably fell into the inner volume of talents and R & D expenses at that time .

Put aside the dependence on talents , In market expansion , because AI Industry characteristics of , In particular, highly customized fragment scenes , It often requires more human investment , Resulting in low per capita efficiency , Calculated by relevant institutions ,AI The per capita cost of the industry is about 50 ten thousand , Equivalent to per capita income .

According to the incomplete statistics of xinmou , stay “AI Four Little Dragons ” among , Cloud from technology in 2017 Year to 2020 The net loss in the first half of the year was nearly 23 One hundred million yuan , According to figure technology 2017 - 2020 The total loss in the first half of the year was nearly 73 One hundred million yuan , Kuangshi technology is 2017 - 2020 The total loss in the first three quarters has exceeded 130 One hundred million yuan .

chart :AI Comparison of net profit of algorithmic companies ( source :Wind, Guojin Securities Research Institute )

besides ,AI The business development of the four dragons is highly coincident ,2020 Year , Security and finance are AI The first two areas with the largest market share in the real economy , The proportions are as high as 53.8%、15.8%, The total is close to 70%, On one side, the commercial imagination space is compressed , On the one hand, the number of losses is rising ,AI The foam of industry valuation has been punctured , The cold winter also comes quietly .

review 2018 Peak year ,AI The industry has completed 523 Start financing , Total amount about 667.1 RMB 100 million , And by the 2020 year , This group of figures is reduced to 305 rise 、243.3 RMB 100 million , Cut it in half . The carnival is behind the scenes , The four little dragons actually have no winners . however , although AI The industry is experiencing a bottoming out period of disillusionment , But it can also be understood as a new climbing period .

03 According to one side

2021 year , Is seen as a AI The first year of industrialization .

AI The industry began to bid farewell to the crazy burning of money in the past , Everyone tends to consider building a reasonable business model , help AI Industrialization landing , Also find a way out for yourself . This makes the past crowded track , Also ushered in differentiation . When I first started , Almost all four are related to security 、 It has something to do with finance , As the business evolves , The myth of technology is no longer beautiful , The landing scene has become the focus of attention , This time, , The four dragons are on each side .

Shangtang technology positions itself as an artificial intelligence platform company , Mainly rely on the sales of software platform to make money . Based on the company's self-developed SenseCore, Centralized mass production artificial intelligence model . At present, it has intelligent business oriented SenseFoundry-Enterprise( Shangtang ark enterprise open platform )、 For smart cities SenseFoundry( Shangtang ark city open platform )、 Face wisdom Smart Life Of SenseME、SenseMARS、SenseCare Platform and intelligent vehicle oriented SenseAuto( Shangtang jueying intelligent vehicle platform ), The four plates covered are also regarded as the four-wheel drive mode of Shangtang .

Smart business and smart city are the main businesses of Shangtang , The proportion of the two has increased year by year , The total income ratio of the two once exceeded 85%. But these two tracks are also the mainstream landing fields of artificial intelligence , Relatively mature , The competition is also very fierce , In addition to the other three dragons : Open vision 、 Cloud from 、 According to map ; And Hikvision 、 Dahua Co., Ltd .

It is worth mentioning that , In fact, the main roads chosen by the other three companies are different from Shangtang .

And soup selling software 、 Different platforms , Yinqi, who has done face recognition for six years , In the later stage, Kuang is more likely to be positioned as a product company , Emphasize the integration of software and hardware . What the world offers AIoT Hardware and software integration solution , The main thing is to treat yourself as Brain++ Core AI Algorithmic system 、AIoT Software composed of operating system and industrial applications , The combination consists of sensor terminals and edge devices 、 And automation equipment .

On the commercialization landing , Mainly for individuals 、 Three scenarios of city and supply chain , make AIoT Platform and ecology . Take the Internet of things as the carrier of artificial intelligence technology , By building AIoT Product system , Consumer oriented Internet of things 、 Internet of things in cities 、 The three core scenarios of supply chain Internet of things provide proven industry solutions .

Also emphasize the combination of , In fact, there is Yitu . stay AI In the early development of the industry , Almost all of them are focusing on computer vision technology , The moat constructed by visual recognition technology is no longer reliable , The difference turns into the direction of betting according to the plan at that time , Combine chip technology with algorithm technology , According to the picture, I chose to AI Force in the field of computing technology and products , Introduction AI chip . In addition to traditional areas such as security , The biggest feature of Yitu is the deep cultivation of medical treatment , Focus on medical imaging .

But betting on chips is bound to be difficult , After being included in the list of entities , Cloud chip research and development once stopped , Differentiation does not bring good luck to Yitu , Early capital market , Compare the first of the four dragons A Shares of the cloud from technology , Yitu has not come out of its own darkest moment .

As for the youngest cloud from , Revenue comes from providing man-machine cooperative operating system and artificial intelligence solutions , Main push operating system , The latter is the main revenue of cloud , In the past, the proportion was as high as 93.59%, until 2020 Only the first half of the year saw a decline .

On the one hand, the cloud from the use of man-machine cooperative operating system , Through the analysis of business data 、 Connection of hardware equipment and software application , Got the core entrance of artificial intelligence Ecology , Provide customers with information services 、 Digital and intelligent artificial intelligence services ; On the other hand , Cloud has been trying to expand application scenarios around man-machine cooperative operating system , Including smart finance 、 Smart governance 、 Smart travel 、 Smart business, etc AI Solution .

Each family goes its own way ,AI The tide of the industry has faded , Since the second half , Players also become more calm , The period of walking with your head down is over , The madness of burning money has also experienced , This calm may not be a bad thing . The scenery is long, so it's better to look at it , It's like Zhulong 2017 As I said in ,“ Even if AI Overdo , We should not underestimate its leapfrog progress .”

This article is from WeChat official account. “ New eyes ”(ID:xinmouls), author : Ruan Xue , edit : Sang Mingqiang ,36 Krypton authorized release .

copyright notice
author[36kr],Please bring the original link to reprint, thank you.
https://en.fheadline.com/2022/127/202205071345023567.html

Random recommended