Q&A: North Summit Capital's Wanli Min

Q: Has technological innovation replaced business model innovation as the prevailing investment trend in China?

A: No. You may have a brand-new technology that makes candles very well, but most people won’t buy candles instead of light bulbs for their homes. All technological innovations must solve current problems and future problems, otherwise, they don’t make any sense. Every investor should think clearly on this point. I prefer to use the term value innovation which often combines model innovation and technological innovation.

Q: How do you define deep technology?

A: People talk about quantum, blockchain and a variety of new term combinations, as if the newer the term, the deeper the technology. Entrepreneurs often use complicated technical terms, assuming it is deep technology – but is it deep because of the unfathomable semantics? For true deep technology, one can use the most obvious language to make its value clear. For example, AlphaGo recently unleashed its deep learning power to predict how proteins fold. This is typical deep technology, and the value path is clear – it will significantly reduce novel drug development time. Everyone can understand this.

Q: What traps should investors be mindful of with deep technology?

A: No matter whether it’s deep or shallow technology, the logic of value creation must be clear. What kind of problems are you solving? If you can’t answer that, if your technology doesn’t create real value, there is no point investing in it. I know the usual patterns of tech investment failure. An investor is attracted by some new technical terms and by performance indicators, and he pays little attention to the logical link between technology and product. In some cases, the investor is drawn to a superstar scientist in a founding team who has published dozens of papers. If this is the correct model to follow, you could simply rank professors based on who has published the most articles globally and invest in the top one. The best approach is to ask the entrepreneur about the industry bottleneck he is addressing and how he solves it. Get an answer to that and the probability of success is higher.

Q: Everyone talks about artificial intelligence (AI) capabilities, but how do you distinguish the real thing from gimmicks?

A: Let’s take cameras as an example. There are many cameras in China used for security purposes, for traffic monitoring, for smart city projects. With all this real-time monitoring from so many cameras, who is going to check it? The workload is enormous. But then AI came in and automatically checked every image, attaching alerts to specific events and causing them to pop up on the big screen. The technology being used is not necessarily the best in the world, but it has solved the problem. Hikvision is one such example – data generated from its smart city cameras is invaluable. Industrial AI vision has attracted a considerable amount of investment this year. The technology is used for scene monitoring, process management and finished product inspection. It is a very useful application, and it is technically feasible. So this kind of investment is less risky.

Q: Why is Hikvision, a traditional manufacturer, more successful as an AI solutions provider than SenseTime?

A: Hikvision first sold its cameras, and then iteratively updated its solutions. As a pure tech company, SenseTime started late. Its technology must be implemented in specific industry scenarios to be valuable. Nowadays, it is hard to find value given the space is crowded. Only an absolutely crushing technical advantage, like a car replacing a carriage, makes a difference. Scenarios are king. Companies that occupy those scenarios have the final say. They can choose a technology provider or just use self-developed technology. 

Q: What do you think about Tesla versus Waymo and their different approaches to autonomous driving? Does shadow mode (installing autonomous driving systems in cars purely to collect data) mean Tesla can move faster?

A: Tesla has a huge first-mover advantage – there might be one million Tesla vehicles on the road today. They have encountered various unexpected situations and they can learn from all these. It is a process of quantitative change leading to qualitative change. Human drivers often start out very slowly, but as they spend more time on the road, responding to different scenarios, they become better drivers. With so many cars on the road accumulating data, which is solidified into the next version of auxiliary driving software, it’s a terrific iterating power. Although Waymo’s cars are equipped with LiDAR and directly targets L4 autonomous driving [the car is fully autonomous in certain environments, but it still needs a driver in the seat], the company’s user base is relatively small. Of course, if Waymo starts mass production one day, it would be a different story.

Q: Do you see intelligent manufacturing in China?

A: China’s manufacturing industry basically does the part with the thinnest profit margin in the global value chain. We all know that most of the production and assembly for Apple products happens in China, but most of the profit goes overseas. If you replace those Apple production lines with the world’s most advanced intelligent manufacturing, you would still only get a small profit. Is intelligent manufacturing any different to normal manufacturing in this context? We should first think about getting a larger share of the value. If that is possible in certain areas, the scope for intelligent manufacturing to make a difference might be a bit higher.

Q: Start-ups such as Baibu and Smart Fabric claim to use cloud factories to deliver flexible manufacturing. Are they pioneers in intelligent manufacturing?

A: The so-called cloud factory is really a network of small production lines. What kind of value is created? Has your output per unit of energy consumption increased? Has the average production rate per ton of raw materials increased? Simply aggregating small production lines to increase capacity is a factory version of Pinduoduo. You don’t empower anyone, you’re just doing linear addition – one plus one plus one. Intelligence should turn each one into a 10.

Q: But haven’t cloud factories solved problems, namely bringing orders to small manufacturers and meeting the needs of retailers that want smaller batches of inventory with more variety?

A: Why can’t some factories get orders? China’s supply-side structural reforms should eliminate excess and backward production capacity. If you are continuing to feed orders to capacity that should have been eliminated, you are just deferring the inevitable and prolonging the painful struggle. As for allocating small orders to individual factories, that seems like a kind of household contract. What are we aiming for in the textile industry – industry 4.0 or agriculture 4.0?

Q: Which parts of China’s tech sector are worthy of attention?

A: Everyone is talking about chips. Besides domestic substitution, I think there might be an opportunity in edge chips with low power consumption. Today, you see a lot of cameras and sensors in shopping malls and offices. If you add some intelligence to it – analyzing consumer behavior data and responding to it by changing in-store product positioning – edge computing can be done in a very cheap way. The technology threshold for edge chips is lower than for AI chips used in cloud computing. At the same time, China has the richest edge computing scenarios. It’s possible to take the lead in this area. Another field I pay attention to is manufacturing industry operating systems and industrial control software. There is a large gap between local and foreign companies, which makes for a good investment opportunity. We must first replace the boss – the production line brain – and make the control system smarter. Whether it is domestic substitution or the next round of intelligent manufacturing upgrades, whoever controls the nerves of manufacturing creates great value.

• Tweet  
• Facebook  
• LinkedIn  
• Google plus  
• Save this article  
• Send to  

• Topics
• Greater China
• Technology
• GPs
• China
• North Summit Capital