After a couple of years of red-hot explosion, the Chinese artificial intelligence sector is returning to rationality and reason in 2019. For Chinese early-stage venture capital firm MSA Capital, it means steering away from mainstream AI companies and try to find startup teams that are focused on previously overlooked niche markets.
"We stay outside of the mainstream types of businesses — AI is one of them," said Ben Harburg, managing partner of MSA Capital, in a phone interview with China Money Network earlier this month. Harburg said he is seeing "more humility" among entrepreneurs in terms of valuation, and find there are still pockets of new areas in AI that provide compelling prospects.
MSA Capital, founded in 2014 and has over US$1 billion under management, stayed away from the crypto bloom and bike-sharing craze in China. In AI, the firm invested in autonomous driving solutions provider Fabu Technology and healthcare big data analytics firm Yidu Cloud. In 2018, it had four initial public offerings (IPOs) among its portfolio companies, including Chinese group-buying and review site Meituan-Dianping, e-commerce firm Mogu Inc., Chinese electric vehicle maker NIO, and Pintech.
Ben Harburg is a managing partner of MSA Capital. Prior to MSA Capital, Ben Harburg worked for American management consulting firm Boston Consulting Group and founded several companies. He was a Fulbright Scholar at Freie University Berlin and Neubauer Scholar at Tufts University.
Below is an edited version of the interview.
Q: MSA Capital co-led a US$31 million series B1 financing round in Beijing-based community group buying e-commerce platform Songshupinpin.com. What are the major reasons behind this investment?
A: The Chinese new e-commerce model remains as an incredibly vibrant and rapidly evolving industry. We’re in a market where you would have thought that companies like Alibaba or JD.com would have essentially had some degree of monopoly in the Chinese e-commerce market a few years ago. But out of nowhere comes the rapidly-growing models: companies like Pinduoduo being valued higher even than JD.com. I think that’s because these e-commerce players including Songshupinpin.com target markets that are outside of the mainstream, using innovative social, collaborative, or gamified approaches to retail that are not being used by traditional e-commerce players. These new models can be very successful in what it already appears to be a very competitive e-commerce market.
Q: What is the key business focus of Songshupinpin.com in 2019?
A: We’re already in 10,000 communities across 40 cities, up from 3,000 a couple of months ago. We already built great partnerships on the supply side with fresh food markets and other producers. But now the critical thing is to think about how to continue to expand and find more folks within communities that can serve as selling agents for us.
Q: Based on your observation, what are the major differences between the Chinese community group buying industry and that of America?
A: I think the critical thing is that everyone in China is on the same social networking platform, WeChat, which dominates their daily lives. People are already very much accustomed to buying and selling things through WeChat mini-programs or WeChat groups. The existence of WeChat helps all these businesses scale up. Pinduoduo simply would not have been possible in America because we don’t have a single kind of unifying social platform that brings together day-to-day conversations with the retail element, payment platforms, and all those components that facilitate really easy sales.
Meanwhile, it’s much easier for Chinese people to buy and sell some of the critical-need goods within their communities, as most people don’t have cars and the transportation is more spread out to get to their local market.
Q: Looking forward, what are the other type of attractive opportunities in the new retail space?
A: One thing we’re quite interested in is gamified buying and other things that make the shopping experience more fun, more spontaneous, and more engaging.
This is particularly important for the next generation — a generation of more influential tech-driven consumption. The gamified buying is one kind, and I think another kind is what we call "content-based search." Rather than simply using the keyword search, if I’m looking for a speaker, the content-based search would be like: What’s the best speaker for an apartment that’s 3,000 square feet and has five bedrooms? It’s a more detailed narrative around their shopping intention. I think these two areas are still substantially undeveloped today.
Nowadays, a lot of people are rushing to these autonomous stores or hybrid online-and-offline channels, but I think we’d like to look at something beyond just improving the efficiency of our day-to-day experience but look towards what really is going to reshape the consumer experience.
Q: What are the fields you’re looking at in the artificial intelligence (AI) sector?
A: Some people call the AI as a sector, but I prefer to call it "an enabler." For us, it’s more about how our business is incorporating data science and machine learning into their daily operations in order to aggregate hugely-valuable data, which they could monetize themselves, or they could use to optimize the efficiency or improve the shopping experience, software experience, or transportation experience.
We look at AI across multiple verticals like mobility, enterprise, retail, life sciences, and healthcare, such as autonomous driving solutions provider Fabu Technology and healthcare big data analytics firm Yidu Cloud. We also invested in other companies in the mobility space like Didi Chuxing, Uber, and Mobike, which are using daily traffic and algorithms to optimize routes, offer geo-located advertisements to customers to get them to buy something here or change their route to go there.
Q: How are the valuations of AI start-ups nowadays?
A: We stay outside of the mainstream types of businesses — AI is one of them — because there are just buzzwords around AI. Anyone who has that kind of thing incorporated into their businesses, however applicable or not, seems to get a higher premium. Certainly, companies like Face++ and SenseTime have huge valuations today, but we’re looking for companies like Songshupinpin.com where we can uncover them at series A round; we can help mold their strategy; we can write the first institutional check; we can continue to follow on in their subsequent rounds. Because we know the company so intimately that we really understand their upside potential and the economics of their business.
So yes, we’re trying to stay outside of the mainstream. I think the overall valuation of AI companies is still inflated. In 2019, I think the overall market climate is cooler than one year ago. We are seeing a little bit more humility among entrepreneurs who take a little bit more time to make decisions upon more reasonable valuations. But as you can see — some of these big guys still have no problem to get astronomical valuations when they’re raising new rounds. So we’re trying to steer away from those mainstream companies. We would need to see an upside case of at least 10 times [return] in a deal before we could underwrite it for our venture fund. I think we struggle to find most of those companies today in the mainstream.
Q: Do you find it difficult to identify those companies outside of the mainstream?
A: There are a lot of great opportunities. China is a much more exciting place than the U.S. for AI. Given the concentration of datasets and mass platforms like Alibaba and Tencent, these things enable companies to gather huge amounts of data through an application or sub-app or mini-app, and then to monetize the data. So we’re certainly not having any troubles finding these opportunities, but we have to keep pushing that boundary and look at more frontier technologies.
For example, we invested in Yidu Cloud three and half years ago when not a lot of people were talking about big data and healthcare. But now everybody’s saying things like big data is the new oil. So I think we should be investing in things today that no one hears about for the next two or three years. And then, they come to the market and become a really big deal.
Q: How is the current cooler capital environment impacting your strategy?
A: 2018 is an awesome year for us. We had four IPOs (initial public offerings), including Pintech, Meituan-Dianping, Mogu Inc., and NIO. We had a big merger and acquisition deal, which is selling Mobike to Meituan-Dianping, as well as a lot of big funding rounds in our portfolio. I think 2019 will be a little bit quieter, but we will probably have three to four IPOs by the end of this year or early next.
We’re really choosy about the types of deals we do. So historically, we’ve had 1% loss ratio even for our investment in the venture stage. Because we invest a small amount of capital in a company to get to know it, and then we’ll scale up our position as we go along. The result is that our downside risk is very low, so we end up being very capital-efficient.
Q: How are your portfolio dealing with the macroeconomic slowdown and uncertainties?
A: It’s not a bad time at all. We haven’t really felt the pinch. For example, one of our portfolio companies, a job search website called "Boss Zhipin," just raised a massive round of funding. I think the company is perfectly positioned to take advantage of this kind of macro environment when a lot of companies recently laid off 5% to 10% workers. That means there are a ton of employees looking for their next jobs. Our hiring platform is seeing record-high business opportunities as we saw more downloads in the last few weeks on WeChat.
We’ve invested in some counter-cyclical businesses. We’ve invested in some businesses that are thinking globally about expansion and looking more towards emerging markets. I guess the biggest negative impact we’ve had is that couple of our portfolio companies listed in 2018 have gone down a little bit since their IPOs, or their IPOs weren’t as highly valued as we had expected. But overall, we haven’t been feeling that negative.
I actually prefer a market like this. I hated the market one and half year ago when we were having this crypto bloom and bike-sharing craze. Everyone was running around with a pitch deck, thinking they could reach a US$10 million or US$100 million pre-money valuation in a series A round. What happens when we have a downturn or a cooler period like what we have today is everybody gets a lot more rational. Good companies bubble up to the top and people who were trying to start a business that didn’t make sense have gone to work for other people. We’re happy to see a more mature and subdued market.
Q: What’s your outlook for the Nasdaq-like new high-tech board in Shanghai?
A: It’s still a little too early to say despite we know what it’s supposed to do in principle. I think the Chinese government understands the pain points that are caused by the Chinese A-share market in the last few years while people are being stuck in the queue, and the A-share market is oriented toward more highly-valued companies. I think certainly China has been taking very aggressive steps to improve platforms to increase foreign ownership and access to the Chinese financial market.
The new board certainly provides us with a new liquidity avenue. I think the liquidity avenue is already maturing in the right direction as we can see a lot more mergers and acquisitions over the past two years in China. We already have a few of our portfolio companies that are potentially exploring IPOs through this new board.
We’ll start to see some positive signs that China’s innovation is on the way or accelerating at a faster pace than they were a year ago partially due to the favorable political environment. But it’s unrealistic to say that China will become a tech leader in the next couple of years.
Q: How do you access China’s ability to innovate nowsdays?
A: If you think about the fact that most of the biggest funds in China are 10 or 15 years old, while American venture capital companies can be dated back to 40 or 50 years ago, it’s only very recently that this innovation development has started China. So we can’t expect that tomorrow China is going to be able to generate the same kind of output as the U.S. But I think China can leap on the evolutionary ladder and accelerate at a much faster pace. The Greater Bay Area in southern China is an example of how China can create an integrated creative international ecosystem that connects foreign engineers with their local peers.
The reality is a lot of companies that we see everyday today are companies jointly built in laboratories globally. Their engineering teams are not just sitting in China. They’re also sitting in other countries like the U.S., Europe, and Russia and operating under differentiated systems from China. So I don’t think it’s fair to say just because of the political environment in China, it doesn’t enable them to have free thoughts to be creative. For example, Fabu Technology has half its team in China and another half in America. But I don’t think the Chinese environment has curbed its creativity.
It’s a little bit like the joke about why America isn’t good at soccer. The reason why we’re not good at soccer isn’t because we don’t have incredible athletes. It’s because incredible athletes play other kinds of sports like American football because they are much more lucrative in America. I think it’s somewhat similar in China where top scientists are working in a consumer-internet context. The Chinese political environment could be their challenge but it’s also their strength.
Q: Are there any great misses you’ve made in your investments?
A: Yes, we were the first investor of NIO to complete due diligence, finish the documentation, and sign for the investment in its series B round. But we should have invested in its founding round — we were offered to invest in it at the founding round. At that time, it was just an idea on paper.
The same thing is for Mobike — we could have invested in Mobike since its series A round, instead we invested in it a little bit later. Because we just didn’t understand the sharing economy. At that time, we didn’t buy the broader vision of how it would turn into an on-boarding mechanism for people using these larger platforms.
There are a lot of companies that we looked at since their series A rounds but didn’t end up investing in them: One of our partners was actually an angel investor in Horizon robotics but we didn’t invest in its subsequent rounds thereafter. We also looked at VIPKid in its series A round. Even Pinduoduo — the founder of Pinduoduo is one of our limited partners — we didn’t invest in that company because at that time, the valuation just rocketed up so quickly.