In the field of artificial intelligence (AI), there is a bubble in the world. Now that the bubble has blown up and it has already broken, the real winners are beginning to show signs. Behind the 50 times PS value, AI Vision's swarming enters, and the imbalance between talent and commercialization occurs. AI will subvert the entire industry.
The AI ​​industry has moved from the stage of playing technology concept to the stage of landing. The winners will be those that can quickly solve real economic problems, improve industry efficiency, and even subvert traditional business models.
Alibaba Cloud is the hand of Alibaba Group's penetration from individual consumers into traditional industries. It is the artificial intelligence brain of the entire Ali Group. Hu Xiaoming, president of the company, said in a speech on December 28, 2017 that there is a bubble in the world of artificial intelligence (AI). Although the whole industry will brag about the “high and deep†of artificial intelligence, AI should be something that people can see and hear.
Hu Xiaoming is not the only voicer who is worried about the "AI bubble". Four days earlier than him, Liu Qingfeng, chairman of Keda Xunfei, predicted in a forum that there will be a large number of artificial intelligence startups going bankrupt this year, but the overall AI industry Will be developed on a large scale.
"AI is a 2B business. You can't count on the speed of 2C. Be careful that the bubble is coming." Li Kaifu, founder of the Innovation Workshop, said earlier.
People usually use two indicators to measure whether there is a bubble in a field: Is there a large amount of hot money entering the field, which makes the valuation of companies in this field far higher than the actual value? Is technology widely applied in the real world and has changed the previous business model to a large extent?
Hu Xiaoming believes that artificial intelligence needs to have three conditions at the same time to be considered “valueâ€: first, there must be scene-driven, solve problems, how much cost is reduced for this society, and how much efficiency is improved; second, behind artificial intelligence Is there enough data to drive the improvement of AI capabilities; and third, whether there is enough computing power to support it.
But more AI traders interviewed by reporters believe that the first thing Hu Xiaoming mentioned is the only factor that determines their life and death - those companies that can quickly solve real economic problems, improve industry efficiency, and even subvert traditional business models. It is possible to win in this competition.
In China's VC circle, AI is very expensive, and it is not a secret that money is not yet available.
A late night in Beijing winter, in an inconspicuous restaurant, several financial consultants in the AI ​​field are discussing the financing progress of the project. They found that in 2017, the image recognition company's PS (valuation divided by the multiple of revenue) is about 50 times, much higher than the normal company's 10 times.
Compared with the Internet industry, the artificial intelligence industry is very new and has frequent financing activities. Therefore, most investment institutions use a relatively comparable and comparable transaction method when they value.
The so-called comparable trading method is to refer to the financing situation of similar companies in the past to formulate the PS value. In 2016, the global AI field has a minimum PS value range of 43.8 times and a maximum of 241.9 times.
Therefore, in the past year, most investment institutions took the median value, financing valuations generally maintained at 50 times, while the normal company's PS value is generally about 10 times.
A venture capitalist who has long been active in the AI ​​investment and financing community said: "Even with such a high multiple, you don't want to vote."
In artificial intelligence companies, only image recognition companies have taken billions of dollars in investment. Many investment companies are vying for the head company represented by the "image four little dragons". These companies have the same qualities: starting with AI technology, usually with a strong technical team; already generating revenue, future income can be expected.
Compared with the star companies that rely on machine learning technology for content recommendation, the valuation of Shangtang Technology is considered overvalued in the investment community. According to a headline today, the headline is expected to generate revenues of 17 billion yuan in 2017, with a valuation of 20 billion to 22 billion US dollars and a PS of no more than 10 times. But this contrast is also controversial in the industry.
One of the main reasons for the image recognition company's fire is that companies have found the most suitable area for AI image technology to quickly land.
The main product of security is the camera. This simple image intelligent landing field is uniquely unique. The scenes of visual recognition in the security field include deployment tasks, human-to-person comparison and intelligent alarms. These scenarios can quickly and easily deploy AI products, such as algorithms. SDK, integrated control system, face image retrieval comparison system and smart front-end camera (embedded SDK).
Almost all of the AI ​​visual identity startups, including the "Image Four Little Dragons", have been involved in this industry. Investors and their joint development of expected sales, these sales, will become these companies continue to move forward (listed independently or The cornerstone of selling to large platforms is also an important basis for investors to realize.
For many years, security was a non-sexy and heavy-mode industry. Resources were almost in the hands of industry giants Hikvision and Dahua, and few startups were involved. Haikang’s profit is equivalent to five Dahua, while Dahua is equivalent to five third-ranked Eastern Netpower. Other companies’ profits are much less than those of Eastern Net. Hikvision's total operating income for the first half of 2017 was 16.448 billion yuan, an increase of 31.02% over the same period of the previous year. The net profit was 3.292 billion yuan, an increase of 26.22% over the same period of the previous year.
The enthusiasm of AI Vision Inc. has also enabled the two traditional giants to accelerate the pace of +AI.
But the security camera loaded with AI capability is only a small part of Hikvision's business products. Some people analyzed that the ratio is about 5%.
An entrepreneur who once served as a technical expert to the public security department told reporters that the state's technical strength in the security field has been quite high. There is no technical obstacle for the public security department to arrest fugitives. He said: "Why does Hikvision not go to AI products? Because it is not the right time." The company founded by the entrepreneur is a partner of Hikvision.
Another founder of AI, which also focuses on security, told reporters that security is currently the most adequate industry for AI, and the current technology penetration rate is only 1%. In other words, only 1% of security devices use AI technology.
“In 2017, the output value of the security industry was 450 billion yuan, and all current AI security products have an output value of less than 2 billion yuan,†the founder said. Video surveillance cameras usually have 4 channels, 8 channels, 16 channels and other specifications. One camera can be installed all the way. Today, the largest order in the field of intelligent security is in Shanghai Baoshan, which is 3000 roads. Compared with the huge security field, this number is almost Can be ignored.
Some people are optimistic that AI will subvert the security market in the next few years. By then, AI products and solutions will occupy at least 50% of this market, and will also completely improve the efficiency of this field.
One visible change is that, in less than a year, product and technology costs have been falling at a rate of 10 times, and development prospects are expected.
Low penetration and high growth make investors want to squeeze into this gamble. A number of financial advisers told reporters that the four head companies of image recognition no longer need financial consultants to assist in financing, and investors will find their own homes.
Security is also a strong regulatory industry, with major needs concentrated in public institutions such as public security, finance, and operators. In the second half of 2017, the national team investment institutions began to enter the market, their role is crucial, in addition to funds, but also for these companies to enter this relatively closed market to level the barriers.
But frequent and high-funded financing has made these early-stage startups, and the number of shareholders far exceeds normal levels. Shangtang Technology has nearly 20 investors, and there are more than 10 shareholders who despise technology, including many strategic investors who need to combine business.
When a field accepts resources far greater than the value that can be created in this field, it is seen as a bubble. However, if AI is put into the entire technology field, its enthusiasm seems not surprising.
The hundreds of millions of dollars in financing is a huge sum for the early AI technology companies, but it is only a small amount if it is placed in the Internet. Since the establishment of Didi Travel, the total amount of financing has exceeded 130 billion yuan, and the financing of multiple rounds has reached 3 billion US dollars. In the shared bicycle field, the new round of financing of Moby and ofo has exceeded 600 million US dollars.
In 2017, the total amount of financing in China's AI sector was 345.8 billion yuan, less than three times the total amount of financing for a company in three years.
"Today, the money spent by China's Internet on subsidizing users is far higher than technology. I think this bubble is acceptable." Voice Intelligence went out to ask founder Li Zhifei to comment on the reporter.
At the beginning of November 2017, after the Vision Technology in the “Image Four Little Dragons†completed the largest financing in the AI ​​field in the same year, the company’s CTO Tang Wenbin was hesitant when asked by reporters whether the visual AI company was overvalued. After thinking for a while, he replied: "The industry does have some false fires. How to spend this money to maximize business development is indeed a problem."
72V Battery Pack ,Lithium Ion Battery Pack,Lithium Battery Pack,Battery Power Pack
Zhejiang Casnovo Materials Co., Ltd. , https://www.casnovonewenergy.com