The “AI+” era in China has just begun.
Editor’s note: This article is from the WeChat public account “Light Source Capital” (ID: lighthousecap), the author of the light source capital. Authorized to reprint.
Recently, the “AI+ Venture Capital” forum jointly held by the light source capital and Inspur held the “2019 Artificial Intelligence Computing Conference”. Many well-known investors and artificial intelligence industry entrepreneurs gathered at the scene to explain the investment and the chain. The significance of value and continuous empowerment for the long-term development of the AI field also shares the views on the future development of the AI+ scenario application.
The founder and CEO of Light Source Capital Zheng Xiaole started the theme of “Investing in AI Industry: Above the Glacier, the Sound of Life”, and the industry situation, investment preferences, development pain points and future trends of the artificial intelligence industry investment. Anatomy. The following is the speech:
Founder of the Capital Capital, CEO Zheng Yule
Light source capital has been established for five years, mainly focusing on the new economic field, providing financial advisory services based on financing and mergers and acquisitions for enterprises in the growth and middle and late stages. Up to now, we have helped more than 60 companies to complete 100 rounds of financing, with a total financing of more than 8 billion US dollars; 18 of the companies we serve have become unicorns in this field, and all market capitalizations add up to about 650. About US$100 million, they are the best companies in China’s new economy and the most outstanding entrepreneurs.
After five years of accumulation, the head enterprises that the light source serves have covered many industries such as e-commerce, consumption, sea, logistics supply chain, automobile travel, corporate services, IOT, and so on. Among them, AI is the field of continuous attention and deep cultivation of light source capital. We have provided services for many excellent AI companies, including the fourth paradigm, Tess, DataVisor, Zhujian Intelligent, echeng Technology, Yicaitong Technology, Roobo. , drop technology, non-code technology, Weimar car, cloud account, onion math. As the capital market is in full swing, the fundraising situation is unresolved, and the AI market seems to be above the glacier. However, the deep integration of AI and industry has just begun. The era of “AI+” is coming. The name of the song is the name of a song I like very much. I hope to use it to express hope for the future.
above the glaciers
China’s Internet wave has undergone a transformation from an incremental market to a stock market, relying on demographic dividends to achieve growth in the first epochIt has passed, and now, with the long-term bonus of capturing the fragmentation time of mobile users, it has spawned the birth of a new batch of giants such as today’s headlines. However, we have to think about when the demographic dividend and the long-term dividends are gradually peaking, what can continue to trigger a new round of Internet convergence, and the efficiency dividend will undoubtedly become the keyword of the Internet third century, in the era of stock and stock optimization. Under the premise, the most powerful weapon to improve unit economic efficiency is AI, whether it is To C or To B.
The capital winter has not passed, the AI industry concept dividend is fading
Let’s first look at some of the capital markets. In terms of the amount and amount of funds raised by the fund, in the case of a significant decline in the amount of funds raised last year, this year’s year-on-year decline is about 16%, compared with the US dollar fund, the decline of the RMB fund is even greater. . From the perspective of investment, the year-on-year decline in investment cases this year is greater than the decline in fundraising last year. Despite the temporary blowout of US dollar investment last year, the US dollar’s fundraising situation was also worrying last year, which led to a rapid contraction of US dollar investment this year. Therefore, the stock funds of the entire capital market are being seriously consumed, and the capital winter will continue.
Data source: Pedata, Wind, light source finishing
Focusing on the AI field, the number of startups has dropped sharply, investment frequency has slowed down, and the trend of capital concentration to the head is very obvious. Statistics show that from 2015 to 2018, the amount of corporate financing of more than 1 billion US dollars in all industries accounted for 50%, 60%, 60%+ and 70% of the private equity financing in the primary market, respectively. It may be further improved this year, which essentially reflects that all emerging areas and head office financing do not seem to be difficult. However, the financing of the long tail company has become obviously difficult, which has led to a significant increase in the threshold for entrepreneurship. The AI field is also the same.
Data source: business card, billion euros
It can be seen that the entrepreneurial window centered on technological innovation and platform innovation is gradually closing, and the “first half” of relying on the AI concept to start a business is coming to an end.
The window of entrepreneurship narrows, and AI investment tends to be rational
At present, the entire capital market has a general sentiment towards the AI industry – it is difficult to make a profit and its valuation is high. In addition to today’s headlines, the University of Science and Technology, and the fourth paradigm, there are still many companies that are not profitable. There is a gap between the volume of revenue and the scale of valuation. The fund investment AI company cannot guarantee the compound rate of return, which directly affects LP. Willingness to invest.
We recently made a statistic, interviewed about 20 major investment institutions in AI or key promoters in investment institutions, and got the first-hand market feedback data:
1 The number of investments is reduced. The investors surveyed saw 14 AI projects per capita this year, with an average of 2 in a month, and in the early and middle stages. The number of AI projects was less than 2 per capita.
2 Focus on early investment. Investors are most concerned about finding good teams and investing in early projects with reasonable valuations, rather than chasing high in the middle and late stages. Even if the investment in the middle and late projects is more concerned about the progress of commercialization.
3 The three most well-received areas. The results of the questionnaire show that the most optimistic areas are industrial manufacturing, followed by financial technology, and third is speech semantics.
4 The three major segments of the most uncertain business prospects. The first one that everyone is concerned about is autopilot, the second is the chip, and the third is medical imaging.
5 The three major segments of the commercialization are most clear. Investors believe that the prospects for commercialization are more clearly defined by human resources, customer service and industrial manufacturing. These are the prospects for commercialization.
Data Source: Light Source Capital
In summary, the investment in AI-based products tends to be rational, and how AI helps companies reduce costs and increase efficiency has become the core entry point for the AI industry. In addition, how to ensure that the commercial landing situation is good, and avoiding the valuation bubble has become an element of the second half of the AI field.
Sound to the
There must be fertile soil above the glacier
The test of the ice age is actually a double-edged swordUnder the impact of internal and external environments, 70% or more of AI companies will die out, and AI companies with real scene capabilities, liquidity, and technology leadership will stand out as super unicorns, which is full of expectations. As in the harsh environment of the Quaternary glacial period, 70% of the mammals were extinct, including giant animals such as mammoths and saber-toothed tigers, and humans survived the test and survived until now.
In essence, AI is still an enabling enterprise service. The next opportunity lies in the innovation of AI plus industry or scenes. This trend is gradually emerging. The trend of digitization and intelligence of the fourth industrial revolution has changed to this day, the infrastructure of deep learning has been relatively mature, and objectively provides a powerful boost for the development of the AI field.
The first is the development of the Internet, cloud computing and big data. At present, the big environment, infrastructure and social cognition are still relatively mature. The second is the increase in the chip computing power of the GPU, and there are also a lot of computing chips that can significantly improve the performance of each scene for AI. The third is the improvement of IOT’s ability to collect data, including the depth, breadth, and continuity of data. The fourth is that the accuracy of the deep learning algorithm has begun to break through. The fifth is the promotion of capital. The sixth is the middle-seat that everyone is talking about recently. The middle-class will make some advanced technologies and services input from outside the enterprise, and gradually become the internalization capability of the enterprise. The evolution of the medium-sized or enterprise organization form makes the AI available in the enterprise. An important driving force for the rapid landing of all links.
Data Source: Gartner
Along with the deep integration of AI and various fields, a number of head artificial intelligence companies have emerged in many industries including chip, vision and voice. From the perspective of revolutionary and maturity, the application of speech semantics is particularly prominent in the development maturity and commercialization ability, chat robot, natural language generation, AI+C&SI service, intelligent application, AI acceleration chip, Process automation software, etc., will also begin large-scale commercial applications within 2-5 years.
Six-point expectation under the three-dimensional evaluation model
After long-term business deepening, the light source capital sums up the three-dimensional evaluation model of AI+ commercial application value to determine whether an industry can combine with AI to generate 1+1>2 market space.
The first is the basic level of data informatization in this industry.Industry data networking is not separated in every device, it is the foundation of AI development; secondly, the industry application space, the bigger the industry will breed the bigger the company, which is related to the ultimate ceiling of AI development; finally the degree of marketization of the industry Industry regulation, decision-making cycle, commercialization progress, etc. all determine how deep AI penetration in this industry can be.
Based on the above three dimensions, it can be seen that the areas including content information, finance, security, etc. have achieved extremely obvious AI. For example, face recognition has played a big role in the airport security process; financial anti-fraud has completely replaced the original risk control duplication audit in financial institutions; whether it is headlines or fast hands, AI has solved hundreds of millions or even billions of users. The issue of content distribution. Retail, education, automotive, medical and other industries will also be highly anticipated.
Specific to see the future development of each field, we believe that the following six areas can achieve rapid growth with AI.
AI+ chip. Whether global or domestic data is exploding, the demand for algorithm complexity and precision is rising rapidly. The general-purpose chip can no longer meet the algorithm’s demand for computing power. At this time, not only the development of GPU is rapid, but also There are a large number of edge chips that can greatly improve efficiency, and these companies are worthy of attention.
AI+NLP. The way of human-computer interaction is shifting from 3.0 to 4.0. Natural language processing should become the mainstream form of human-computer interaction in the next era. This core technology will also have a lot of floor space in various scenarios—finance, customer service, and retail.
AI+ vehicle manufacturing. With OEM as the center, in the next 5-10 years, the first self-driving car with L3 will appear. In the future, L4 may gradually begin to land. But the threshold for autonomous driving is very high, requires a strong technical ability and a very good platform. There are only a handful of such companies, so the giants or startups that can keep abreast of the giant ecology have more opportunities in this field.
AI+ medical imaging. Most of the companies in this field were still in the trial stage of various hospitals. It is worth looking forward to the large-scale commercialization of the people who started the mainstream hospitals.
AI+ Industrial Vision. The efficiency of China’s manufacturing industry has reached its peak in the era of manufacturing 1.0. How to break the game? It still depends on how we can integrate with AI in the era of the 4.0 industrial revolution, including the scale application of various sensor products, industrial Internet of Things and 5G commercial, cloud computing and edge computing, industrial big data, measurement and specific industrial robots.Applications have a lot of market space.
AI+ data in the middle. In essence, AI Zhongtai is an infrastructure that every enterprise can have in the future. Through this infrastructure, whether it is big data or other artificial intelligence, including image recognition, etc., all have the opportunity to touch more quickly. There are a large number of applications in all aspects of business operations, whether it is the client or the operator, the production terminal, and the inventory management delivery terminal. The customer of Light Source Service is the practitioner of AI+ China Taiwan. Its self-developed DEEPEXI series products and platform solutions help enterprises to quickly land on the Internet in the middle of the platform, data center and intelligence based on the open source system. Taiwan, quickly realize the value of digital construction, and achieve the business objectives of data creation value.
The era of “AI+” in China has just begun
The era of AI as a single product has passed, and the era of combining AI and industry has just arrived. At present, more than 90% of AI enterprise customers on the market are B-end customers. The development of AI enterprises is inseparable from the development of various To-B companies with AI as their core competence. The AI industry entered the second half. The four barriers of AI enterprises were first of all the ability to productize; the second was the refined scene; then the business ecology; and finally the industry Know How.
The competition in the field of AI is a long-distance race. It is not about the heroes of the company’s market value in one or two or two years. It depends on whether the company can become an economic model and have long-term hematopoietic capacity. A profitable company. This requires companies to identify their most powerful genes and AI integration, emphasizing the ability of AI to organize and deepen the ability of all aspects of business management, emphasizing the ability of AI to deliver products under the scene, and finally embracing capital.
The average time for most US corporate service companies from establishment to listing is 8.3 years. The vast majority of AI companies in China are no more than 5 years old. From 2014, a large number of AI companies have not been in business for more than 5 years, so China The long-distance running of AI companies has just begun.
Finally, I think that AI entrepreneurship is actually a friend who makes time, and still has a lot to do.
Thank you all.
Light source AI company