Tracing back the development history of artificial intelligence and laying the theoretical knowledge of current commercialization can be traced back to 70 years ago. Up to now, we think it can be roughly divided into three stages:
The first stage (1943-2006): This is a period of theoretical knowledge accumulation for the development of artificial intelligence. During this period, some phased achievements have been produced, but generally speaking, due to the data scale and computing power, the relevant research progress is slow;
The second stage (2006-2016): the algorithm, computing power and data have made a qualitative leap over the previous stage. The triple joint force has promoted the application of artificial intelligence, achieved remarkable results, made major technological breakthroughs in computer vision, speech recognition and other fields, and artificial intelligence began to have commercial value;
Phase III (2016 to present): taking alphago's defeat of Li Shishi as a milestone, artificial intelligence began to be fully promoted to commercial applications. Countries all over the world pay high attention to the development of artificial intelligence, especially in China. Under multiple factors such as policy support and strong digital infrastructure support, China's advantages in artificial intelligence application scenarios are particularly prominent.
The three types of enterprises have their own strengths, or will eventually compete on the same stage. We summarized the participants in the AI industry into three categories, and summarized their advantages and relative disadvantages in the competition:
Internet technology giants: represented by Google, Baidu, Alibaba, Tencent and byte beat. Have strong algorithm ability, master data resources (especially C-end data), have significant brand effect, have the most comprehensive ability, and have a great advantage in the competition; However, it is easy to face the problem of internal resource allocation;
AI startups: represented by Shangtang technology, Kuangshi technology, Yuncong technology and Yitu technology. The R & D capability is outstanding, and most of them are based on self-developed framework, and the advantages of specific scene algorithm capability are prominent; But lack of industry and customer accumulation;
Traditional hardware manufacturers: represented by Hangzhou Hikvision Digital Technology Co.Ltd(002415) and Dahua software. There are many years of business accumulation in specific scenarios, upgrading to the direction of intelligence and adapting to the AI era. Its advantage lies in its strong sales channel and customer expansion ability, but it is relatively weak in algorithm ability.
The three types of enterprises will share the industry growth dividend for some time in the future. However, with the slowdown of industry growth, or will eventually compete on the same stage, resulting in customer resource competition or redistribution of value chain, the competition pattern is expected to be further clarified.
Common issues faced by AI start-ups. We believe that in the next stage, AI start-ups will face the following two topics:
Pressure on Profitability: at present, the four AI dragons are in a state of loss. The path to achieve profitability is reflected in the business level. We can focus on the revenue growth brought by customer expansion and model innovation, and the scale effect gradually appears, driving the expense rate to decline significantly;
Path selection: the artificial intelligence scenario landing is complex and the business chain is long. We judge that the above manufacturers must make certain choices. There are two specific paths: one is to make choices along the scenario, resulting in the verticality of the business scenario; Second, trade-offs are made along the links of the industrial chain, resulting in the division of labor of the industrial chain.
Risk tip: the industry growth is less than expected and the industry competition pattern is deteriorating.
