Investment in the field of science and technology should be driven in both directions-both from the bottom to the realization of core technology, and from the application side, its adaptability to technology. That is “see what type of screw first, and then look for a matching screwdriver.”
The “New Capital” series aims to dig out the real and valuable content of investment institutions, so that a wider range of entrepreneurs and readers can understand the true look behind each institution. These capitals are one of the most active investment groups in the market today and are also future catchers.
Gorge Capital is a venture capital fund that focuses on investing in early and high growth periods. It was founded by the earliest returnees in China. Follow the track in the field of “smart technology” including artificial intelligence, smart production, semiconductors, and data services. The company’s accumulated capital management scale exceeds 3 billion yuan, achieving an overall return of more than 6 times and an average annualized rate of return of more than 80%. Its investment projects cover China and the United States.
The company’s management team has more than 50 years of accumulated investment experience and has successfully exited more than 10 projects. For example, the world ’s leading supplier of network security equipment Feta Technology (listed on NASDAQ in 2012), the world ’s leading shared computing and blockchain technology innovation company Xunlei Technology strong> (listed on NASDAQ in the United States in 2014), one of the world ’s leading providers of video decompression technology service DivXNetwor (listed on NASDAQ in the US in 2006 and acquired in 2008), Semiconductor software design EPA company Celestry (sold to Cadence in 2003).
At present, the company has launched the Phase III smart currency base, which mainly invests in smart industrial applications and nuclear technology, focusing on semiconductors, high-end manufacturing, and AI applications in the field of enterprise services. . Invested companies cover the entire artificial intelligence industry chain-from upstream nuclear component companies, smart infrastructure companies, and downstream smart application companies to achieve synergy. Includes AI foundation (semiconductor, module, communication, new materials), intelligent platform layer (perception, data, algorithms) and AI applications (hardware, systems, integration, services).
Interviewed Li Man, the founder of Gaojie Capital, and conducted in-depth exchanges on the development of the artificial intelligence industry and the company’s investment philosophy.
Li Man, founder of Gaojie Capital
Mr. Li Man is one of the earliest venture capitalists in China. He graduated from the Computer Department of the University of Science and Technology of China and obtained an MBA from New York University. He has nearly 20 years of investment and investment banking experience across China and the United States. Prior to Gao Jie Capital, Mr. Li worked for Fidelity Funds, Goldman Sachs, and WI Harper. In addition to his wealth of fund management experience, Mr. Li has served as the president and CFO of SouFun. He has deep company operating experience and capabilities.
In the era of “weak AI”, “man-machine combination” is still the dominant
As an important feature of the Fourth Industrial Revolution, since the propositions appeared in the 1960s, artificial intelligence technology has shifted from early simple machine learning to deep learning. At present, artificial intelligence is still in its early stages, and Li Man defines it as the “weak AI” era .
In Li Man’s view, this is not only reflected in the fever of the artificial intelligence investment in the market in the previous two years. At the technical level, the development and mass production of sensors and new-generation chips, the core technologies of artificial intelligence, remain a challenge. , Such as lidar, camera, optics, etc. In addition, the research and development of artificial intelligence chips need to start from scratch. Although there are some excellent systems in the current artificial intelligence framework and tools, there is still a long way to go to implement large-scale deep learning and inference systems.
Seeing the advantages and limitations of the current development of artificial intelligence technology, Li Man believes that the implementation of the current artificial intelligence technology also depends on the “man-machine combination” model, because the current AI It can’t completely replace human’s work, it’s more simple and repetitive things. On the other hand, in order to achieve further development of artificial intelligence, in addition to relying on deep learning technology, it also needs to be highly integrated with other disciplines, because artificial intelligence is related to the fields of big data, programming, robotics, natural language understanding, hearing, and visual processing. Technology and landing also need to combine the corresponding industry knowledge in the application scenario.
Compared to “hard technology” and “black technology”, Li Man is more willing to define Gao Jie Capital’s focus area as “smart technology”. “In our opinion, the core of smart technology is AI (artificial intelligence), but it is not limited to AI. Smart technology emphasizes the application and system construction of AI, that is, based on demand, in specific scenarios, based on AI Technology to solve problems, “Li Man said.
Investment logic: two-way driving of technological progress and industry demand
Although we are still in the “weak AI era”, Li Man believes that AI that can truly integrate related technologies and systematically solve practical problems still needs to emerge. In the form of smart technology, by constructing newBusiness model, complete the restructuring of certain industries, there is a greater systemic opportunity.
“What we care about is how to systematically deconstruct AI from the AI infrastructure, to the core components of the middle tier, and then to the top-level applications in order to find investment opportunities from the segmented track,” Say.
Taking the investment case-Dark Object Intelligence (DMAI) as an example, DMAI is committed to building a new generation of AI platform based on strong cognition. By integrating the six major areas of computer vision, language dialogue, cognitive science, robotics, machine learning, and game ethics, a high-level cognitive architecture for artificial intelligence is established. Unlike most “data-driven” companies, based on deep research accumulation and engineering experience, the team’s founder, Professor Zhu Songchun, has won the Marl Award, the highest award in the field of computer vision three times. DMAI uses a small amount of data and a large number of tasks to create a strong cognitive system, which has been rapidly implemented in the education industry. It aims to solve the teachers ‘real-time grasp of students’ learning status and personalized teaching needs.
For the segmented track of smart agriculture, from the perspective of Li Man, economically, the agricultural sector has an expandable business model and has extremely stable repeat purchase opportunities, “not a one-shot deal.” China has 1.8 billion mu of arable land, and the market space is huge. In addition, agriculture basically does not involve personalization issues. There is not much difference in the pest control technology between rice fields and wheat, and technology and business model have high reproducibility . For example, McFee Technology, a project led by Gage Capital, uses drone remote sensing monitoring and AI big data technology to achieve precise farmland pest control. After the data is transmitted back to the cloud, AI algorithms are used to identify pests and diseases and apply pesticides to farmland to solve farmland abuse At the same time, it reduces the cost of planting.
See more Chinese technology investment
Golden Capital ’s investment team and core consultant team have a strong technical background and deep accumulation in the fields of electronic engineering, computers, software, etc., which makes the company’s investment in the technology industry very objective. Li Man said that the technology industry is an area that requires long-term technology and experience accumulation. In the past, due to the market’s demand for returns, many bubbles appeared in the industry. “In the process of squeezing the bubble, the team has the opportunity to jump out only if it turns technology into a product and has a scalable business model,” Li Man said.
Why invest in smart technology at this time? In the view of Li Man, the blowout of data in recent years has provided sufficient fuel for the explosion of intelligent technology. In addition, the 70-year history of AI has a certain foundation. At present, it is in data storage, algorithms, and vision. There has been a breakthrough. At the same time, semiconductor Moore’s Law is nearly ineffective, and the bottleneck of deep learning algorithms requires new breakthroughs, and new opportunities are being bred. So he judged that nowEntering the market is the more correct node.
In terms of semiconductors, first of all, the opportunity comes from the global semiconductor industry transfer: from the United States to Japan in the 1970s, to Korea and Taiwan in the 1980s, and it is gradually turning to mainland China.
From Li Man’s point of view, As the world’s largest electronics producer, exporter and consumer, China has the world’s most complete industrial system, the largest end market, and government top-down support. The return of overseas talents has significant advantages. Gaojie Capital currently has layouts in the areas of chips such as computing, storage, sensing, IoT, and AI, as well as the upstream and downstream semiconductor industrial chain.
Second, China currently has great window opportunities. On the one hand, China currently accounts for about US $ 350 billion of the world’s approximately US $ 500 billion semiconductor market each year, of which the domestic share is only one-tenth, and most of it depends on imports, especially high-end chips. “We can make a simple estimation. Although it is difficult to replace all imports in the short term, it will replace 50% in 10 years. The market of nearly 150 billion US dollars per year is worth looking forward to,” said Li Man. In addition, he believes that the incremental semiconductor markets such as IoT and edge computing cannot be underestimated.
When asked about the promising segmentation scenarios in the field of artificial intelligence, Li Man told that service robots are a good development direction. At present, the common ones are care robots, smart speakers, community service robots, etc., and they can effectively replace manpower. To complete some single, foundational work. With the development of artificial intelligence technology, Li Man believes that in the future, service robots will develop from a single task to a variety of tasks, although it will take quite some time to reach human intelligence.
In general, Li Man believes that the next 10-20 years will be a golden time to invest in smart technology. Although it is still in the very initial stage, the increase in computing power and the widespread use of data in recent years will lead to the explosion of intelligent technology.