domeet webmaster

This article is from WeChat official account:Touzhong.com (ID: China-Venture) span> , author: Zhang Lijuan, the original title: “” mown investors chives, “the AI ​​unicorn who can usher IPO of spring? 》

How popular it was back then, how much it is questioned now.

As Yitu, JD Digital, Hesai, Yunzhisheng and other AI star companies have been blocked from listing on the Science and Technology Innovation Board, some pessimistic practitioners even put forward the argument that AI has collapsed. AI unicorns I thought it might be our own spring, but some people might never see spring again.

The AI ​​unicorns who came from the 2016 “AI First Year” are all vying to submit for listing at the end of 2020. However, with the disclosure of various prospectuses or other various reasons, some of them did not It is impossible to obtain high valuations that match the financing in the secondary market, and some may not even go up. Even as Li Kaifu publicly admitted in 2020, “Many AI companies have cut investors’ leek.”

Of course, with the optimization and upgrading of AI technology, it is bound to achieve a large-scale landing, but how good the stories told before, how cruel the reality is now. The key question is, how can AI technology be truly integrated with products and implemented on the ground?

For AI companies, what is the top priority? Listed? profit? What about the industry consensus? How to develop for a long time?

The listing persuasion is only a partial phenomenon, and the essence of business must be pursued

Many investors have reached a consensus: AI companies still have to return to the value of the company itself. As for Yitu, Hesai, Yunzhisheng, etc., the temporary suspension of listing declarations is not a serious problem. The reasons for suspension are diverse. From the perspective of the enterprise itself, it is still necessary to pay more attention to the creation of its own competitiveness.

The views mentioned by Xiangfeng Capital’s executive partner Xia Zhijin were also recognized by more guests. “A company’s competitiveness does not only come from technology and products, but also from operations, customer relations, manufacturing and other environments. It has an impact on enterprises, and many technology companies often ignore the competitiveness of the entire business environment.”

In Xia Zhijin’s view, the country is indeed controlling the pace of listing. On the one hand, too many companies are queuing to go public, and there may be mixed good and bad. Not all companies in every industry have the opportunity to go public. This is the reality. On the other hand, it is a big misunderstanding if you think that those companies that could not be listed before can be listed and have good valuations because of the science and technology innovation board; in particular, technology companies have no profit and no problem, even some Want to go public without income, this may be a misunderstanding.

Li Jianing, executive director of Mingshi Capital, added that the entire listing channel has been more open in the past two years, much better than it was five or six years ago. The overall trend is more beneficial to enterprises and entrepreneurs. Generally, there is a bubble in an industry. After the bubble bursts, many good companies and good talents can be left, which is still conducive to the overall development of the industry.

In the development of the AI ​​industry, scenarios are more important than technology. Take the 100 2B companies invested by Mingshi as an example. Few companies have AI as their first label, but 80% of companies have AI attributes. “We invested in a company that uses AI as a boiler energy-saving company. What is more important here is that there are people who understand the business scenario and apply a relatively advanced technology to the scenario, opening a significant gap with other players in the industry.”

Li Jianing also pointed out that the AI ​​we are talking about now has many bottlenecks, because AI is still limited by its upstream data collection capabilities, computing power, and acceptance of downstream scenarios. The so-called bottleneck may be broken through changes in upstream and downstream.

Zou Yanshu, executive director of Huaxing New Economic Fund, made the bottleneck encountered by AI companies more concrete:

1. The scene is too concentrated. Security does have a very good foundation and a very huge market. As long as AI companies have a certain market share, they can create a very good income volume.

2. Return to the essence of business. For AI companies, there must be a model for making money in the end. “From another perspective, where is the technological content of an AI company? What is the essential difference between it and a project-based company with a revenue of 1 billion yuan and a gross profit of about 20%? This requires the company to answer the question and tell the market. Tell investors.”

Chen Siyu, executive director of Haisong Capital, also pointed out sharply, “Haisong pays more attention to whether the innovation and application of the underlying technology have solved the real pain points(must have)(must have)(must have)

span>, instead of pseudo-demand (nice to have). The combination of AI and traditional industries is good, and algorithms replace manpower to reduce costs and improve efficiency Fortunately, it is necessary to grasp the true core needs. Not only must the product end be uniquely competitive, and be able to solve the differentiated demands of customers, but also the income end must be healthy and sustainable.”

This undoubtedly expresses investors’ views on the overall AI industry: Whether it is an investor in the primary market or from a regulatory perspective, what everyone cares most is the authenticity and availability of the business. Persistence.

For this, Guo Kejia and partner Ding Runqiang stated, “Whether any new technology can really solve real problems, the business is real, the customer is clear, someone pays for it and has certain pricing power, And under the premise of real business, it can be copied to different industries and customers.”

In addition, what AI companies must solve is the ceiling problem, and the solution to the problem is to really solve some practical and valuable problems.

“This problem must be very concrete and specific in the short to medium term. It cannot be large. Do we use a higher cost to solve a simple problem, or can I solve the same problem at the same cost? The problem? This is actually the most critical point for us to distinguish between true technological innovation and a pseudo-science.” Ding Runqiang said.

As for the final success in listing, Melissa Yang, partner of CCV, the founding partner, said that China’s stock market has moved from the earliest approval system to the approval system, and then to the registration system of the Science and Technology Innovation Board. The most important thing is that the future is more important than the past. , Because the approval system looks at past data. “We are investing in cutting-edge technologies. AI is one of them. On the whole, the Sci-tech Innovation Board is still good for AI. This general trend has not changed. But now the Sci-Tech Innovation Board is a new thing and is constantly being adjusted. “

In this way, for companies, Aolin Technology Investment and Financing VP Guo Nan mentioned that for 2B companies, everything they do needs to create value for customers before customers are willing to pay. And only product delivery can truly show the company’s ability to productize.

Lai Ye Technology Finance VP Guo Yu also pointed out, “First create unshakable value at the customer level, and after successfully verifying the customer value and market value, then in the capital market through the role of capital amplification to achieve more Many contribute to the value of the whole society. Therefore, focusing on the essence of business may be more important to the company than creating value in the capital market.”

Qu Han, Vice President of Pensee Technology, provides another dimension of approach: “Now it should be in a period of large bubbles. Therefore, technology companies like ours, especially those engaged in 2B business, need to pursue cost-effectiveness. , Is to greatly reduce the cost of acquiring AI value for customers in business. For example, in the 2B market, the cost of our products may be only a fraction of that of the 2G market.”

Yuan Peijiang, founder & CEO of Deep Wake Technology, concluded that the AI ​​industry will eventually return to its value and business model. Improving efficiency and reducing the cost of the entire transaction are the ultimate mission of the company.

If a company can’t face this problem objectively, can’t solve these problems, and find a suitable business model, the road will not go far. Whether it is Baidu, Ali, Tencent, or Huawei, the development process has never been smooth sailing. This process may be a kind of experience for these companies, and then they will do more solidly.

Fortunately, Hu Hetong, a partner of SenseTime’s Artificial Intelligence Industry Fund, also mentioned that strict supervision and public opinion is conducive to the growth of leading companies. Head companies, especially the absolute leading companies, have less impact on listing, but it is more difficult for long-tail small start-up companies to meet the listing requirements than before. Therefore, from the perspective of the overall capital market trend, the head effect becomes more obvious, and resources will be further tilted to the head enterprises in the future, and the “2-8” law becomes “1”.-9″ law.

For the autonomous driving track, Jiang An, the founder and general manager of Zhongke Huiyan, said that autonomous driving is a bit like a high-pitched singer. Since the previous years, the tune was too high, which made it impossible to keep up. Recently, with the further implementation of autonomous driving application scenarios, this track has begun to be sought after by the capital market again.

“In recent years, the autopilot industry has experienced ups and downs. Products, technologies, and business models need to be real and sustainable. These are very real issues that everyone is concerned about. So I feel that after experiencing this wave, right The entire autonomous driving industry is also a kind of washing. Many companies that have survived this stage will be able to embark on a more healthy and stable development process. After all, the market capacity and landing scene are too large.”

In addition, in March 2020, 11 national ministries and commissions including the National Development and Reform Commission jointly issued the “Smart Vehicle Innovation Development Strategy”, which clearly stated that by 2025, intelligent vehicles with conditional autonomous driving (L3 level) will reach mass production and achieve a high degree of automation. Driving (L4 level) smart cars are marketed and applied in specific environments. In Jiang An’s view, this is also a huge benefit to the entire industry.

As a gold digger in the tide of autonomous driving, Yang Yuxin, CMO of Black Sesame Intelligence, also said that the current understanding of autonomous driving in the entire automotive industry is more biased towards L2 to L3 level autonomous driving. The market is beginning to return to rationality, the technical direction is relatively clear, and the time point is relatively clear. The automakers are basically preparing for mainstream models in 2023, striving to catch up with Tesla’s current level, which is two years.

How can AI companies build their own ecosystems steadily?

So, after the status quo of the AI ​​industry is reached, many AI practitioners are more concerned about how to be a good service provider of the scene and truly become an AI gold nugget under the current trend of continuous decentralization of the AI ​​scene ?

Hu Hetong said, The maturity and integration of new technologies have given birth to a variety of new business models and application scenarios. The era of AI empowering all industries has arrived. On the one hand, there are more and more opportunities for AI technology in various sub-industries to be implemented and realized. On the other hand, the strategic deployment of “full digitalization” at the national level is also beneficial to the development of AI companies.

But in these vertical fields, it is impossible for a platform company to do products and services in every field; platform companies need to cultivate ecological partners in different fields. Therefore, for each AI startup company in each vertical field, the future There is still a great opportunity for development.

Taking SenseTime as an example, a platform company like SenseTime is a provider of infrastructure in the digital age.

On the basis of the supercomputing platform or open technology platform of platform-based enterprises, each AI company can develop more subdivided application scenarios in the industrial ecology. This ecology is actually an open and tolerant ecology, and There is still a big difference between the so-called disputes between the princes just mentioned on the 2C side.

Xia Zhijin also pointed out the issues that more entrepreneurs are concerned about. Is AI becoming more and more scattered?

From the application side, it is becoming more and more dispersed. We see that AI will be used in various industries, agriculture will also use AI, security, automobiles, and all walks of life will use AI. This is completely decentralized. This dispersion may be different not only in technology, but also in business models.

Going back to the issue of AI startups doing the ecology, who leads the ecology is not necessarily the developers of downstream AI applications, and may go to the upstream core competence providers to empower various applications, forming a relative A more relaxed ecology, rather than a particularly tightly integrated ecology, is more of empowerment.

Yang Yuxin said that her views on AI may be relatively simple and rude:

AI is not an industry in the first place, but the AI ​​business model can be simply divided into a platform business model and a service business model. The AI-oriented scene itself is decentralized. It is not that AI is decentralized. Because AI is facing a decentralized market, the more weapons you have, the higher the amount of value that can be generated when empowered. Therefore, the best business model for AI that sinks into the scene is the service-oriented business model.

“Large companies have the ability to do the above things. All platform resources are in your hands. Whoever has more resources will have greater commercial value. And the small companies below The industry is deeply rooted and the quality of customer service is high, and whoever has a premium right will become bigger. But it cannot be separated from the business, and the business model is to do service.” Yang Yuxin said.

Ding Runqiang emphasized that whether it is technological maturity or enterprise scale, there must be practical problems. SenseTime can of course set up an industrial fund, because it was done earlier, but for more entrepreneurs, it is necessary to find micro-customers in the vertical field first, first as industry customers, and then do industry migration. Not to mention that each ecology has a large scale. The more upstream it is, the more intensive it is. For example, the most upstream chip is actually whoever is it, but the downstream industry goes down.It may become more divergent and open.

Guo Nan’s Aolin Technology’s entrepreneurship is precisely because several founders took a fancy to the timing of the trend from automation to informatization to intelligence: many companies’ production lines have been able to achieve Industry 4.0, and even have “black “Light Factory”, but to the corporate management level, each department basically uses some very primitive methods to align the company’s strategic business plan. This is one of the opportunities that Aolin Technology discovered.

“Originally, the business may be digitized, so that the people above can be displayed and seen by the sub-department or sub-modules. Now, I hope to return the global data of these companies to guide some small businesses. Decisions on business scenarios and points to guide some major strategic directions. This is another opportunity.”

Qu Han also pointed out that the AI ​​industry chain is gradually dividing labor, and some things are gradually standardized and productized. Pence is now experimenting with 2B-end AIoT standardized products, including AI terminals and Pence Cloud’s intelligent services. “We made functions such as AI services, equipment management, video processing, and data analysis into mid-office capabilities, and then opened this part to channel partners for cooperation. In this closed loop, partners are business experts, and Penz is an AI expert.”

“Second, we may also need to pay more attention to the activity of small ecology. For example, our partners have various industry informatization experts. In various fields such as pan-security, education, transportation, industry, and retail, AIoT products And the service is embedded in the partner information system. The richer the partner, the stronger the integrity of the solution when facing different customers, and we can be channels for each other and help each other.” Qu Han said.

Yang Yuxin pointed out the essence of the problem, From the perspective of doing business alone, doing business requires volume and profit. A relatively better way is to do business. Because the platform is doing well, the development can be sent to the supplier, and the company can directly face the customer.

How do you make profits?

Yang Yuxin said that profits will return to AI technology itself. “Everyone is going to do AI and use AI to help empower the industry. The premise is that AI can help you and save you the cost of investment. This is to increase your volume and use your technology to iterate continuously to increase your investment. Profit. Only in this way can the company expand. Of course, there must also be a consulting team that can turn customers’ pain points into needs. This is a very important point.”

Ding Runqiang also pointed out that first, if the product or the core of the business is to be used as a tool to empower other applications, the premise is that there is sufficient scarcity; second, the development tool itself, whether it is software or hardware, There is a relatively complete system, with a large audience, and easy to get started; third, there is a high enough margin, this margin is so high that in addition to raising its own core R&D team, it will be a loss for any downstream project development. This is the prerequisite, and these three are indispensable.

But in Li Jianing’s view, it is possible that the company is not hitting the pain points of the customer’s life or death, but his itchy points. For example, using AI to help customers save 5 manpower, the value of cost reduction is clear. “The problem is that the customer is facing too many things today. He wants to engage in sales and internal quality control. There are many things that are exhausted. Of course, saving these 5 people can produce value, but it is not necessarily a problem that he has to solve immediately. “

Zou Yanshu also said that when Huaxing New Economic Fund was talking to companies, also often asked whether the company must have or nice to have. Everyone can desperately verify that this is a nice to have. But this is actually a more contradictory issue.

“I shared a case. We invested in Shell. Shell is now a company with a market value of 100 billion U.S. dollars. Now if everyone goes to the real estate forum again, they will regard Shell as a benchmark company in the industry, capable of disrupting the market. New economy company. But if you go back to 10 years ago, this investment is not necessarily a must have. At that time, traditional real estate companies did not feel the threat of shells. Some things may not be what you can see today , And the leading company will definitely take the lead in the entire market and become a vane.”

Therefore, Yuan Peijiang of Deep Wake Technology said that many companies are now building ecosystems and promoting the implementation of technologies and scenarios. Similarly, these ecosystems will also form princely divisions, and data security and data sharing have become bottlenecks. The data “chimney”, so technological innovation and new rules are needed.

Ma Dongjun, the managing partner of Haisong Capital, believes that on the one hand, the Chinese are very smart and start a business immediately to focus on a scenario that is particularly easy to achieve.

But more importantly, when can a company or a group of people appear, when the cash flow problem is not so immediate, they can use the first principles to ask some questions like Musk..

“In the entire industry, we now have enough smart people and excellent companies, but there need to be some companies that can do the long, medium and short things well. I especially look forward to our entrepreneurs, maybe the time may come, and they can appear Leading subversive things. In this way, it is good for all people and participants.”