27 June 2020
In the summer of 1956, scientists in a small two-month conference at Dartmouth College in Hanover, New Hampshire, USA, discussed for the first time "the use of machines to imitate human intelligence." It will come to mind that until March 2016, 60 years later, Google's AlphaGo defeated Go World Champion and professional ninth player Li Shishi, this round of artificial intelligence (AI) revolution has only begun to pick up.
After 2016, the global artificial intelligence technology upgrade and industrialization momentum is surging. Thanks to the rapid progress of artificial intelligence technology, by 2019, artificial intelligence technology has entered a full-scale commercialization stage.
In financial, medical, security and other fields The implementation technology has landed, and the application scenarios are becoming more and more abundant. The commercialization of artificial intelligence has brought positive effects such as the acceleration of enterprise digitalization, improvement of the industrial chain structure, and improvement of information utilization efficiency. It is foreseeable that the development of the artificial intelligence industry will further accelerate in 2020.
But the process will be long. For startups in this field, the prosperity brings opportunities, but the risks are always accompanied, even death.
Starting in 2016, startups in the field of artificial intelligence have been given the best treatment in this era. Especially in China, a large amount of capital has flowed into the field of artificial intelligence. Such a grand situation has continued for three years, until the big cooling in early 2019.
In 2019, the overall financing enthusiasm of the AI industry has declined significantly. So far, Shangtang Technology, which has the highest valuation in China's AI field, completed four financings in 2017 and three financings in 2018. In 2019, this company has not disclosed Any financing information.
Financing becomes difficult and we need to find new ways. In March 2019, the introduction of the science and technology innovation board provided new opportunities for many AI companies that have not yet been profitable.
However, the road to listing of AI companies is also slightly difficult. On August 25, 2019, AI unicorn companies defied technology to submit IPO applications for Hong Kong stocks. So far, they have not successfully listed. "Finance" reporter was informed that a number of well-known AI startups are preparing for the IPO of science and technology board, but as of now there is no exact schedule.
One is the narrowing of the financing channel, and the other is the step by step of the giants. The technology giants represented by BATJ Huawei, as well as the industry leading companies that are actively transforming into intelligence, are seizing the right to speak of AI. But it also gave AI companies a new outlet they might start thinking about selling to giants.
In the past year, a reporter from Caijing interviewed a large number of people in the AI industry and investors concerned about the AI field. The consensus was that the cold winter of capital had arrived, and it would be colder in 2020. The next most important thing is to improve self-hematopoiesis and find new ways.
In the past few years, the abnormally enthusiastic venture capital field has encountered the "coldest winter". Data service business card data shows that in the field of artificial intelligence in China, a total of 523 financing transactions were completed in 2018. In both years, these two figures declined significantly, which were 371 transactions and a total of about 27.19 billion yuan in financing.
The reason is not complicated. The enthusiasm of capital has caused many AI companies to rapidly expand and obtain high valuations that do not match the current business volume. Once the market cools down and capital starts to be rational, the more highly valued the company, The harder it is to introduce new financing.
The viewpoint of Hu Bin, founding partner of Strategic Capital, is that capital-driven entrepreneurship is established in some industries, such as those that need to burn money in the early stage to seize users, but it is uncertain in the current AI field. The status quo is that the services provided by startups have not been significantly differentiated, which has discouraged investors. Hu Bin acknowledges that the technology of head AI companies is very good and has potential, but it is difficult to judge which step will be taken in the future.
The venture capital environment has matured rapidly in the past few years, and once unicorn companies that were very rare are all over the world today. "So many AI unicorns with billions of dollars in valuation, what will happen after going public? For investors, The risks are getting bigger and bigger. "Hu Bin said.
AI companies are generally talented and have high R & D investment costs. At the same time, they are mainly engaged in To B businesses. It is difficult for the To B business to grow rapidly in a short period of time. The rate of earning money cannot keep up with the rate of spending money.
AI entrepreneurship is divided into hardware and software. Hardware includes AI chips and intelligent hardware, and software is mainly solutions. The representative companies of hardware are the Cambrian of AI chip company, and the contempt technology that transforms from soft to hard. Although many AI startups have released their own AI chips, most of their revenue is still based on software.
In the field of AI chips, Nvidia is currently at the forefront. In China, Huawei has the strongest chip capabilities. In addition, Alibaba is also pressing harder by virtue of its semiconductor company, the flat-headed brother. Not much.
Smart hardware has a relatively low threshold, and at the same time can bring more revenue growth than software. At present, the main product forms include smart speakers, robots, smart cameras, etc. However, hardware naturally requires higher cost investment. In the background, big companies are more competitive. Taking smart speakers as an example, the only players left are Alibaba, Baidu and Xiaomi.
The difficulty of the software business is that it is difficult to do large-scale. An AI startup founder told Caijing reporter that AI companies are mainly engaged in government and bank business, but both types of customers are very strong.
"Many AI companies claim to have reached cooperation with certain government departments and some large banks, but these cooperations cannot be directly converted into orders and revenue." The founder said, "One is the problem of repayment, and many other cooperations are actually a very Small orders, you may only be a small part of the integration. "
The two largest software companies in China, Kingdee and UFIDA, had revenues of 1.485 billion yuan and 3.313 billion yuan in the first half of 2019, and their latest market capitalizations were 30 billion yuan and 80 billion yuan, respectively. year.
This is the status quo of China's software industry and the reality that AI startups need to face.
Where is the way out?
In 2020, the focus of AI startups will shift from financing to business.
During the financing period, the company needs to "tell a story" to tell investors the core technological strength and the great value that technology can generate in the future. When such a "story" actually landed, it was not very convincing. The customer only values one thing, and that is whether technology can make him reduce costs and increase efficiency.
Investors are also changing their minds. A number of investors who are concerned about the AI field recently told the Caijing reporter that they will no longer ask technical details, but will need the company to come up with real business data. "How valuable is the technology in real-world scenarios, not as high as 99%," said one of the institutional partners.
The founders of AI companies are mostly technically born, and technically born CEOs care about cutting-edge technologies. In order to achieve small breakthroughs in technology, they are willing to pay a lot of energy and costs. However, a company is not a research institute or laboratory. At present, the entire AI industry is still in the early stages of development. On the other hand, advanced technologies will be unconvinced.
The CEOs of AI companies with multiple technical backgrounds told Caijing reporters that when they are truly facing the market and customers, they find that understanding the industry and solving problems is the core competitiveness, rather than letting AI break from 95% accuracy to 96% .
When AI startups go deeper into the industry, they will see that demand is almost everywhere.
This is the best era of AI entrepreneurship. Almost all industries are developing in the direction of intelligence. The former "Internet +" has been transformed into "AI +". AI retail, AI medical, AI finance, AI education, smart cities , Smart factories, smart homes ... opportunities are endless.
National policies are also strongly supported, and capital is still flowing in. The Chinese government has announced that it is striving to become the world's leading AI innovation center by 2030. It has formulated a national AI strategy and plans to invest tens of billions of dollars in AI research and development. Local governments are also responding positively. Beijing announced the investment of 2.1 billion US dollars to build an AI science and technology park; Tianjin plans to set up a US $ 16 billion AI industry fund; Hangzhou Future Science and Technology City plans to invest 3 billion yuan to build China (Hangzhou) during 2018-2020 Artificial intelligence town.
This is also the worst era for AI entrepreneurship. As the next epoch-making new technology, competition has become fiercer than ever. No one wants to fall behind the trend of the times, especially the giants who have occupied the highlands in this era.
In 2020, it is foreseeable that we will see more AI startups start the IPO process, and at the same time, a large number of companies will stop there and withdraw from the competition.
Perhaps you can expect new ways to be acquired by the giants. In the United States, the ultimate dream of a large number of To B technology companies is to be acquired by the giants. The latest case is that Google is considering the acquisition of Salesforce. For the giants, such mergers and acquisitions can increase the technical strength faster and gain more market share; For the acquiree, they can focus on technology without worries and better integrate with the actual business.
But in China, there are few such cases, but some clues have already appeared, and more and more industry leaders have begun to participate in investment in the AI industry. Once the M & A channel is opened, it will generate more stimulation for the entire AI industry.
In 2020, AI startups began to part ways, some will step down from the altar and leave sadly; some will join the sea of the secondary market to accept greater tests, but given the important position of AI in the future of humanity, these are all It's not the end. The era of AI has just begun. The question is only who can stay and who will withdraw.