Shulou(Shulou.com)06/02 Report--

2019-12-31 21:25 introduction: entrepreneurship is not a helpless rush, but a persevering journey against the countercurrent.

Looking back at the AI security road in the past year, there are ups and downs, but the ups and downs are less than the ups and downs.

Internal economic downturn, external economic and trade frictions. Some people joked that the winter of 2019 was really long, and the casual cold wind made people shiver.

Fortunately, in the cold wind, some people have been holding on, waiting for the next spring, they believe:

Entrepreneurship is not a helpless rush that does not advance or fall behind, but a persevering journey that impacts against the current.

Today, China's security has reached the top of the pyramid of world security, and behind the miracle is the innumerable Chinese security entrepreneurs who do not accept or believe in fate.

For this group of wave riders:

No matter how bad the cards have been, every hit is bound to give the maximum potential energy a rebound.

In English, the root of the entrepreneur empresa has the meaning of "take risks to get". In the coming 2019, Lei Feng.com AI Nuggets has sorted out eight AI security topics from nearly 100 "adventurers" to present readers with important memories in the field of AI security in the past year.

Hu Yangzhong: top Ten Summary to resist internal and external troubles

In the security market, Hu Yangzhong, president of Haekang Wei Vision, is like a "big brother." once the internal competition is too fierce or the external market is too deteriorating, he can always make a "stable military morale" voice at the right time.

Prior to this, Hu Yangzhong rarely attended the latest investor research conducted by Haikangwei, and answered most of the confusion and questions of the participants, including Haikang strategy, competitors, industrial changes, industry prospects, and so on.

From his ten thousand-word answer, Lei Feng chose ten major viewpoints:

1. Today, we can still rely on hardware to make money, and software will have a heavier say in a few years.

2. The development of AI edge-aware devices is not as expected, and personal judgment is still greatly influenced by macroeconomic; of course, maybe Haikang's expectations are too high.

3. Frequent mergers and acquisitions can not solve the practical problems of security. In the past, few mergers and acquisitions were made by Haikang or its competitors. Once there was too much capital integration in China's security industry, but it eventually became a game and farce of capital.

4. TO P, TO B and TO SMB markets are changing too fast, and the future growth prospects are affected by many factors, so it is difficult to judge in the short term. In the long run, I am more optimistic about the enterprise market.

5. There are more projects this year and less next year, which is not particularly important and need not be calculated too clearly. We don't have to be too obsessed with numbers, revolve around numbers, and take a long-term view. This is what we should do.

6. Data is aggregated on demand, and some enterprises have proposed to put all the data on the cloud. I don't think all the data of the Internet of things should be on the cloud, and it is not necessary to put all the data on the cloud.

7. The road to intelligence is not yet fully open. In some directions, we thought there would be good market space, but there may not be; there are also many directions, and there are some special gains inadvertently.

8. In the past, you could make money by making 10 million yuan in revenue, but now you may lose money if you do 100 million a year. The security market is very fragmented, it is easy to survive, it is difficult to grow, anyone can take a bite, but it is difficult to live a comfortable life.

9. Some large enterprises have proposed not to touch applications and data. Haikang is not the case. Haikang will do intelligent applications in some areas. Partners can do it for partners, and partners cannot do it ourselves.

10. We have seen too many projects stop work and end badly due to various benefit distribution problems, leading party ability problems, and leading party business adjustment problems. Enterprises should learn to cover the bottom and dare to do so.

Li Ke: dreamer, make Dahua more imaginative

Li Ke, 47, has worked for Huawei for 21 years, and the top official is the president of South America and vice president of the group.

In April 2017, Li Ke parachuted into Dahua and became the director and president of the company.

Li Ke believes that the technology of the industry continues to evolve, the needs of users are constantly updated, and whether the technology, products or solutions must be transformed into measurable value actually generated, customers will be interested in buying.

Based on this, Dahua, led by Li Ke, put forward a brand-new "AIoT+SDTV" twin-engine strategy.

AIoT emphasizes the ability of perception and intelligence with video as the core to achieve a comprehensive perception of the real scene; SDTV strengthens customers' business understanding, builds software and algorithms around the value calculation of data, and builds a closed loop of customer-oriented value chain through AIOT+SDTV.

AloT and SDTV are not only the technical route and architecture that support the landing of Dahua HOC, but also the business planning and direction of Dahua.

Dahua's future business strategy is to walk on two legs: data up and business down.

Based on the success of customer-oriented value, a closed-loop management from the perceptual layer to the business layer is realized. At the same time, they use ecological logic to build future development, that is to say, no manufacturer can monopolize the market in the future.

Dahua wants to be an ecological company that can form partnerships in different scenarios through architectural design flexibility and understanding of the business.

On the other hand, Dahua has accumulated in the business of AIOT in the perception layer, including video and voice capabilities based on cloud architecture, big data's modeling algorithm, and some capabilities of the customer application layer of the software platform, to form Dahua's development advantages in the industry.

Zhang Pengguo: work together vertically and horizontally, just want to win again

According to IHS research data, in the video surveillance market over the past eight years, Yushi has gone from being unknown to the fourth in the world and the third in China.

Behind a few simple figures is persistence and forge ahead that can not be touched by ordinary people's naked eyes.

The times gave Yushi the heaviest gift, but also secretly marked the price.

Behind one good report card after another, negative voices around Yushi follow: suspension of listing in Hong Kong, merger into A-share listed companies, holding hands with Internet companies, and so on.

Such as the recent measures to join hands with Ali Yun, some industry insiders commented: Yushi has no sense of existence without security, and the layout of Alibaba's urban brain is far from what Yushi can eat.

For Alibaba, Yushi is just one of them, but for Yushi, Alibaba seems to be their straw.

What on earth is Yushi thinking behind all the market movements?

In the view of Yushi president Zhang Pengguo, they constantly meditate on their original thoughts and missions, constantly open up their cognition and boundaries, and constantly talk about themselves like monks to warn themselves not to sit back and watch the sky.

He said that Yushi has always been very determined, but has been looking for the possibility of 12 points of growth.

For the industry, starting with a camera and a DVR, the purpose of collecting data and establishing a knowledge graph is to serve security, but if these data are left idle, after a hundred years, people are still unable to play a role in the development of information and knowledge, and still cannot provide nourishment for promoting the development of technology.

In the impression of Zhang Pengguo, he did too many such things in the first half of his workplace, leading a team to charge, to open up a new field and to win a war.

Today, standing in front of the watershed of industry and technology, he and Yushi want to win again!

He said that people always tend to overestimate their abilities, in fact, they are not as strong as they think, and it is very important for a person to be able to look at their own level objectively beyond themselves.

He firmly believes that when you are strong, you may just pander to the trend, not that you are particularly good. There are no excellent companies, only those that catch up with the trend.

And at present, the biggest potential energy may be Ali.

Up to now, Zhang Pengguo is still unable to define himself and Yushi, or to make a perfect summary of past achievements and efforts.

"We want to do better and go further. Yushi's starry sea is not security, but AIoT. For this starry sea, we are willing to accept all the changes and challenges."

Liu Guang: empire builder, lost control

March this year is a good time for security manufacturers to discuss cooperation and strategy.

In that case, Haikang Yushi has just finished lecturing on the company's strategy and partnership plan on the bank of Qiantang Lake; here, Liu Guangguang, chairman of Oriental Netpower, has talked with Chuantou Xinsheng about a strategic stake, which will replace Liu Guang and become the actual controller of Oriental Netpower.

On March 26 and April 4, Dongfang Netpower issued announcements successively to explain the introduction of strategic shareholders.

Dongfang Netpower announced that on April 4, 2019, Liu Guang, the company's controlling shareholder, actual controller and chairman, and Jiang Zongwen, the shareholder and director of more than 5% of the shares, respectively signed a "share transfer agreement" with Sichuan Investment Xinsheng. Liu Guang and Sichuan Investment Xinsheng signed the "Voting entrustment Agreement".

According to the share transfer Agreement, Liu Guang transferred his 54385175 unrestricted conditional shares (about 6.3642% of the company's total share capital) to Chuan Tou Xinye; Jiang Zongwen transferred his 9500000 unlimited Sale condition shares (about 1.111% of the company's total share capital) to Chuan Tou Xin. The transfer price is 12.53 yuan per share.

According to the Voting entrustment Agreement, Liu Guang entrusted the remaining 163155526 shares of the company (about 19.0927% of the company's total share capital) corresponding to the laws and regulations or the articles of association other shareholders' rights except the right of income, proposal and nomination to Sichuan Investment Xinye to exercise.

If the aforementioned equity changes are finally completed, Sichuan Investment Xinjian will directly hold 63885175 shares of the company, accounting for 7.4759% of the company's total share capital, and the total number of shares with voting rights in the company will be 227040701 shares, accounting for 26.5686% of the company's total share capital. Sichuan Investment Xinjian will become the controlling shareholder of the company, and Sichuan SASAC will be the actual controller of the company.

As for the introduction of strategic shareholders in state-owned assets, Dongfang Netpower's explanation is that the introduction of strategic shareholders in this transaction will help the company to optimize its shareholder structure and improve its ability to undertake large-scale projects. further enhance the company's credit ability and anti-risk ability, enhance the company's brand, capital and market development ability.

"after the completion of this transaction, CIC will make use of its own government resources, industry resources and the advantages of communication and coordination to provide more support and convenience for the company in terms of project resources, financing costs, etc. Further promote the strategic development of the company from video to data, from AI to DI, from software to services, and provide momentum for the long-term development in the next five to ten years, and enhance the company's comprehensive competitiveness."

As soon as the news came out, many people in the industry thought that Dongfang Power "liked to mention" state-owned capital, which deepened the "AI+ security", but there were also voices that Dongfang Power's move meant that the management, especially Chairman Liu Guang, had lost control.

And why the actual controller has changed, Oriental Network Power has also explained on the official website, one of the important items is to "help the company to improve its ability to undertake large-scale projects and further enhance the company's credit and risk-resistant ability."

The item "improving the ability to undertake large-scale projects" can be said to be a pain point for many security enterprises that often undertake large-scale government projects.

At present, many large-scale government projects are carried out in the form of PPP projects, with regional cities and counties as a whole for construction and development, covering a wide range, involving many departments, high investment and long construction cycle, which is difficult for non-large security enterprises to bear. Take the Ping an City PPP project as an example, the first departments involved include finance, the National Development and Reform Commission, the political and legal Commission, the municipal government and so on.

For the security suppliers, the model is mixed, because it means that the security enterprises need advance capital, the project income can only be recovered during the operation service period, and the fund is returned slowly.

In particular, Oriental Network Power has been focused on the "big scene" of urban public safety and participated in many large-scale projects and activities in Ping an City. However, due to the significant changes in the general environmental and economic situation since the beginning of this year, the great downward pressure on the economy, and the long construction cycle and payback period of large-scale projects, it will have an impact on the operating efficiency of the company.

Some investors believe that although the government accounts will be recovered, due to the economic downturn, the local government's finances are tight, and the cycle of getting the money back will be further lengthened. Most of the time, the income is determined, but the money can not come back. It has been reported before that there may be more than 2 billion yuan of uncollected accounts receivable in 2018, which is related to the business of Dongfang Power, which is mainly concentrated in Ping an city projects and government projects.

Prior to this, although Oriental Network Power also came forward to refute the rumors, the massive introduction of state-owned capital and the change of actual controllers inevitably brought worries such as "high bad debts and insufficient liquidity" to the surface again.

Duan Aiguo: domineering words, strong attack on the city

In the past year, Huawei Security has undoubtedly been the fiercest in terms of firepower in the AI security market.

Duan Aiguo, president of Huawei's intelligent security product line, opened fire many times. "listen to competitors say that Huawei is going to quit security? I would like to stress here that Huawei will not quit until it is the first!"

Not long ago, at the launch of Huawei's intelligent security business strategy in 2019, they launched a new brand "HoloSens" and launched what is known as the industry's most powerful camera HoloSens SDC X2382-HL and HoloSens IVS3800 intelligent video cloud platform.

This camera with Penton 310 AI chip, is the industry's most powerful AI SoC (system-on-chip) for edge computing scenarios. It can achieve on-site computing power as high as 16Tops, can achieve intelligent drag 8, and will support simultaneous recognition of 200 different objects, including cars, people, obstacles and traffic signs.

In addition, another product, "Intelligent Video Cloud platform", carries the new Kunpeng 920 chip.

According to Duan Aiguo, the platform, from chip to architecture, is the industry's first fully independent innovation edge video cloud platform, with three times higher performance than the mainstream products in the industry, and the price is relatively low.

The release of these two new products undoubtedly allows Huawei Security to move from a small jog to a strategic rush.

With Teng 310 blessing at the front and Kunpeng 920 at the back, a complete set of software and hardware standardized solutions for Huawei security have been clearly demonstrated in front of the industry.

For a long time, in the security edge side, especially in video processing, Hayth, Horizon, Yuntian Lifei, Cambrian and so on have been laid out in the field of security chips.

In the security cloud chip market, mainly by Intel, Nvidia, Google and other foreign companies to control, domestic enterprises have not yet made a major breakthrough.

And Huawei this "small step" leap, may make China's security forward "a big step".

Interestingly, in the face of Huawei's strong attack on the city, domineering and domineering, the always low-key security giant Haikangwei also frequently spoke out and fought back: Huawei will definitely withdraw from the security market in the future.

Haikang's logic is very simple: "you can't keep big fish in shallow water. It won't be long before you shoot flies with a cannon."

At the same time, at a recent investor research meeting, Haikangwei rarely commented on some of Huawei's security achievements in the past year.

1. "460%".

At the launch of Huawei's intelligent security business strategy in 2019, Duan Aiguo mentioned that global shipments of Huawei security cameras have increased by 460% in the past year, shipments of some platforms have increased by 350%, and Anping transportation and the park have achieved explosive growth.

According to Haikangwei, the figure of 460% is of little significance. At present, the per capita output of Haikang is about 140w to 160w. Huawei achieved 700 billion sales last year and produced 350w per capita. At present, Huawei Anping has about 1000 people, and being able to achieve 2.5 billion a year is the limit, but it is almost impossible to double profits.

In other words, Huawei cannot do high-profit security, while Huawei cannot do low-profit security. In the end, the security business will be in an awkward position within Huawei.

2 、 HoloSens .

Duan Aiguo previously announced Huawei's new security brand-HoloSens. He explained that Holo means holography, and HoloSens products aim to meet the security needs of various industries and urban development for holographic perception and data intelligence in the smart era.

Haekangwei mentioned that holosens is really cool, with chips 16T and SDC integrated, similar to the Porsche version of mate30. But the more flamboyant this kind of product is, the worse it is, and there is no scene in the security industry that needs to use such a cool camera.

3. Hayes chip.

In the security market, Hayes chip is dominant.

Haikangwei believes that as an independent entity of Huawei, Hayth's business goal is Anba, and the horizon and Cambrian are also in hot pursuit, and the only way for Haisi to maintain its competitive edge is to ship a large number of chips and then iterate quickly.

In the security market, Haikang buys tens of millions of Hayes chips every year, while Huawei Security, on the other hand, sold only 100000 cameras last year, and Huawei executives will consider the benefits.

This problem can be compared to Ali investing in Qianfang, and Qianfang acquired Yushi.

From Ali's point of view, the only way to compete for the market with Huawei Anping Division and Tencent 2B Division is to quickly expand the system, which needs the support of excellent hardware suppliers in the market. Yushi, as the third largest security manufacturer in the country, is Dahua 1pact 6 and Haikang 1gam12, Ali naturally will not particularly support Yushi to compete with Haikang and Dahua.

Inky: security standing up, the listing is blocked.

When it comes to neglect and security, people in the industry should have heard of it.

At present, the overall valuation of more than $4 billion is the first batch of unicorn enterprises in China to apply AI technology to the traditional security industry.

Seven years ago, AI was not yet popular, and there were few start-ups, and the VC circle did not recognize it. In those years, it was necessary to have a rare feeling of putting all your eggs in one basket as an AI. Not long after entering the venue, Xianyang aimed at the industry's recognized hard bone: the security market.

People are confused, confused and even ridiculed. After all, there were more views in the industry at that time that AI was just a new pot of old wine, and it was a clich é that it would take a long time to apply and even commercialize it, and the technology application bubble was huge. There are also some industry customers who think that since they were fooled by face recognition ten years ago, they will no longer believe this stuff.

Ignorance ignored this and took root in this vertical field. When the technology is proved to be mature with one result after another, the industry really realizes that the era of AI has really come.

The experience reflects the most real appearance of a startups from infancy to growing up, and when he is born, the business picture gradually unfolds.

In the early days of the landing of technology, his crying, laughing, every move, quarrelling and quarrelling were all waited for by people, watched by the times and recorded by historical pictures, and the voice of the times outlined their outline, more and more.

Finally, when he matured, the line draft of the painting was clearly displayed in front of people.

Today, the product matrix developed by discerning security involves many fields, such as human, vehicle and object feature recognition, and the business line involves cloud, edge and end.

Ignoring the fact that the technology CEO Yinqi had predicted, "2018 will be a year of industry consolidation, and we can see clearly whether we win or lose by the end of this year."

Indeed, in the AI security products and technology side, as an AI start-up, neglect has done well enough, but in front of the capital market, neglect has encountered some problems.

December 3 news, according to foreign media reports, neglect technology postponed up to 1 billion US dollars of Hong Kong IPO to 2020.

The progress of the listing has attracted a lot of attention since the IPO prospectus was submitted to the Hong Kong Stock Exchange on August 25, 2019.

It was previously reported that neglect did not go through the listing hearing on the Hong Kong Stock Exchange because Hong Kong regulators asked the company to provide more information. Ignoring the official response: the report is not true.

Currently, the three investment banks that ignore cooperation are Citigroup (Citigroup), Goldman Sachs (GS) and JPMorgan (JPM).

In fact, it is normal for enterprises to be asked for supplementary materials by the Hong Kong Stock Exchange in the process of listing. Many enterprises have experienced additional inquiries. In fact, there is no such thing as "failed". Enterprises just need to answer these additional inquiries before they can get formal permission to continue stock sales. For example, the two companies listed on the HKEx in October, Zhongshou and Master Lu, also resubmitted their listing materials to the HKEx.

It is reported that the disregard shareholder camp is very strong, Alibaba, Ant Financial Services Group, Lenovo, Innovation works and so on are all neglect technology shareholders.

According to the prospectus, disregarded income increased from 67.8 million yuan in 2016 to 313 million yuan in 2017, and then to 1.427 billion yuan in 2018, with a compound annual growth rate of 358.8% from 2016 to 2018.

Ignoring the fact that the adjusted net profit in 2018 was 32.2 million yuan, the adjusted net losses in 2016 and 2017 were 92 million yuan and 1.42 yuan respectively.

The adjusted net loss and adjusted net profit for the six months ended 30 June 2018 and 2019 were $137 million and $32.7 million respectively.

Yan Shuicheng: bid farewell to 360 and join Yitu

In August this year, Professor Yan Shuicheng, former vice president of the Group and president of the Institute of artificial Intelligence, officially bid farewell to Yitu and joined Yitu as CTO.

Professor Yan Shuicheng is a specially appointed expert of the National Thousand talents Plan and an outstanding scientist of IEEE Fellow,IARR Fellow and ACM. His main research fields are computer vision, machine learning and multimedia analysis. he has published 500 high-quality academic papers and was selected as Thomson Reuters High Citation Scholar (TR Highly-cited researchers) three times in 2014, 2015 and 2016.

"this time I integrate into Yitu in the role of CTO, looking forward to further strengthening the basic research of AI, while going deep into the industry, so that AI technology can be really landed in more scenarios." Yan Shuicheng told Lei Feng net.

Faced with so many choices, why did Professor Yan Shuicheng scoop Yitu out of the three thousand weak water in AI?

"in fact, before joining Yitu, I also considered it for a long time and specifically considered three factors."

First, they are familiar with Zhu long and Lin Chenxi and take a fancy to their low-key pragmatism and complementary positions.

"most of the early CV startups intersect with Microsoft or Prof. Thomas Huang. I spent time in both places when I was a student, so we all know each other very well."

He said that he spent a long time with Leo (Zhu long), the founder of Yitu, in the same office of Microsoft Research Asia; he also intersected with co-founder Lin Chenxi at Microsoft Research Asia, and he also had a good understanding of Yitu's development process, so when Leo was invited, the decision to join Yitu was very simple and quick.

Second, identify with the three major ideas of development according to the map.

In Professor Yan Shuicheng's view, now the layout of CV startups has its own characteristics. From this side of the picture, the software-hardware solution, multimodal technology precipitation, and Tech for Good product concept attract him.

In May this year, the release of the chip formed a prototype of the hardware and software solution layout that was planned several years ago. The release of voice recognition capability last year and the publication of NLP work on Nature Medicine this year have made the multimodal technology layout begin to take shape. Security, medical, pharmaceutical and other layouts all adhere to the core concept of Tech for Good.

Third, curious about the mystery behind the list.

Professor Yan Shuicheng told Lei Feng that he knows a lot about Yitu's development, but Yitu's technical team has always maintained a certain mystery and has dominated the National Bureau of Standards and Technology (NIST) face recognition competition for several years, which makes him curious and drives him to learn more about the team and is willing to do something different with such a team.

Ma Yuan: it is difficult to start a business, the biggest dark horse.

The past 2019 has been a difficult year for AI Security Entrepreneurship.

This year, a list of the demise of startups has stimulated the sensitivities of one entrepreneur after another; most startups spend the money they raised in previous years. Cold, 365 days and nights throughout the past year for most entrepreneurs.

In the same year, there were also "lucky ones" who successfully broke through. Before the opposition between the economic trend and the survival of enterprises, Ma Yuan and his Pengsi technology won a phased victory.

As soon as they were established, they completed the first 10 million yuan financing, and at the beginning of 2019, they received the second investment led by IDG Capital, followed by Gaojie Capital, Ancient Capital and Hongtai Fund. In April, they announced that they had completed the 150 million yuan A round of financing.

Ma Yuan, who graduated from Xi'an University of Electronic Science and Technology, is the first batch of graduates majoring in artificial intelligence "Intelligent Science and Technology" in China. Pengsi Technology, which he founded, is a start-up company specializing in computer vision and the Internet of things. They provide AI+IoT technology and software and hardware solutions for vertical industries such as security. At present, they have formed a team of more than 100 people, the vast majority of whom are R & D.

"there is never a solidified industry, only slow enterprises." Ma Yuan believes that "under the AI tide, a BAT-level company will certainly be born."

Excavators replace hoes, cars and airplanes replace feet. When people feel that they have liberated their hands and feet, the emergence of AI liberates our eyes and minds.

Just as cars and airplanes have rewritten the way humans walk, we are not afraid that they can easily exceed the limits of physical exercise. People can also look forward to AI optimistically, believing that it will redefine security, health care, education and many other industries, and will eventually create a new type of intelligent society.

In the process of constructing this intelligent society, Ma Yuan thought that some earlier AI companies did not take root deep enough, spread a little wide, and did not make up for the shortcomings fast enough.

Too many technology-driven AI enterprises, not only to maintain the technology leadership, controllable thinking, but also to meet the needs of capital profit, often lose themselves in the balance between gain and loss.

"I have seen too many mismatches between actual needs and product functions, which is one of the main reasons why I left to start a business."

He said that the first wave of AI startups swarmed into the security market, inclined to the Internet-style business style of fast rush, quick work, and fast land grab, and did not hesitate to use subsidies and other ways to get customers and occupy high ground as quickly as possible.

But the To B market is completely different from the ToC market. The ToC industry has mastered users' thinking and habits, it is easy to appear popular styles, but the change cycle is not long; the demand of the ToB market is relatively vertical, and it pays more attention to a stable, reliable and predictable service model.

"To B and ToC are completely two sets of logic. unfortunately, in the past few years, both capital and enterprises want to use the tactics of ToC to quickly erode the market of To B."

Ma Yuan believes that the word "diversification" should not be mentioned frequently by startups, and "full empowerment" should not be mentioned frequently by startups. Start-ups, including Pennsylvania, have limited resources, limited capacity and limited personnel, so they need to do their best to work together.

In his words, slow is fast, less is more, and some things seem slow in order to rise abruptly based on accumulated strength. People often overestimate the results that can be achieved in the short term, while underestimate the goals that can be achieved in the long run. This is what start-ups should do. Don't be in a hurry and take a down-to-earth approach.

Http://blog.sina.com.cn/s/blog_cfa68e330102zs77.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.