“Sniper investor” Allen Zhu is slowing his investment pace, focusing on in-depth project research
The managing partner at GSR Ventures talks about VC investment in China, large language models, consumer market, Chinese companies going global, etc.
Good evening. Perspectives from investors on the Chinese economy are always popular among the readers of this newsletter. For this year’s first piece of GRR on investor views on China, I choose to translate a recent interview of 朱啸虎 Allen Zhu Xiaohu, a Managing Director with GSR Ventures, a company that invests in early-stage start-ups. The interview was conducted by the magazine《中国企业家》China Entrepreneur and was posted on its WeChat account on Dec. 25.
Allen Zhu focuses on investments in early-stage companies in the internet, mobile and new media sectors, according to the website of GSR Ventures. He has rich experience in software product development and global expansion. Prior to GSR Ventures, he co-founded eBaoTech, a leading provider of insurance core back-end systems. He has also worked for McKinsey Greater China, where he earned deep consulting and implementation experience.
In the interview, Zhu shares his perspectives of his investment experience in recent years in China. He explains why 2023 became the most comfortable year of his investor career, and why he made investment decisions more slowly than before. From his interview, you can tell the shifting focus of Chinese investors. The topics include:
Large language models (LLM): those who invested in them in early 2023 faced losses later on
Entrepreneurship: "less money" is a double-edged sword
Investment in Chinese consumer market: a more prudent approach
Going global: Chinese entrepreneurs' work ethic is unrivaled globally
Please note that the translation has not received Mr. Zhu's personal review and was edited by GRR for clarity.
When Zhu Xiaohu (Allen Zhu), managing director of venture capital firm GSR Ventures, attended a meeting in the United States in October, 2023, a limited partner (LP) was surprised that Zhu didn’t have gray hair on his head.
In fact, Zhu even thinks his hair is darker than in the past. One reason may be that he’s not so worried about how the projects he invested in fare. “I won’t lose sleep as long as the businesses can make profits,” Zhu said in an interview with China Entrepreneur in his office at GSR Ventures.
Zhu remains sensitive to profitable opportunities and new trends. In April 2022, Zhu made an investment decision shortly after a 15-minute video call with William Li, founder and chief executive officer (CEO) of Fancytech. He sensed an opportunity: AI-generated content (AIGC).
At that time, AIGC hadn’t become a hot trend yet. Fancytech specializes in helping e-commerce websites with content aggregation and brand growth with higher conversion rates. After Fancytech incorporated AIGC technology into its business and exhibited rising performance, Zhu invested more in another three funding rounds.
To some extent, however, that sensitivity comes his own way, in explicit disagreement with others. For example, when most others sought after large language models (LLMs) — advanced artificial intelligence models that uses deep learning on large data sets to understand and generate new content — in the first half of 2023, Zhu suddenly emerged from a period of silence and poured cold water on their enthusiasm. In a public speech, he mentioned that ChatGPT is not friendly to start-ups and they should let go of their funding fantasy in the next two to three years. His remark triggered an objection from Fu Sheng, a well-known Chinese entrepreneur and chairman of internet company Cheetah Mobile.
Today Zhu still sticks to the view that LLMs would prove a boon only to tech giants and start-ups should focus on vertical market segments such as sales and marketing. In his opinion, the same logic applies to investment. “I think investors must set their eyes on what they’re good at, because the market cycle keeps changing.”
AIGC and consumer market have always been the focus of Zhu. He is very familiar with the two areas, yet his investment decisions are made more slowly. In 2023, Zhu spent more time doing research before investing into a project. Quick investment has almost become a thing of a past, like the one in 2012 when Zhu decided to fund Chinese ride-hailing company Didi after only half an hour’s chat with its founder Cheng Wei. “Without much competition from other investors, I had as long as three to four months on average to consider whether a project is worth investing in.”
At such pace, 2023 became the most comfortable year of his investor career, in a shift from racing to invest in a project to investing without haste.
As of November 2023, Zhu had invested in only one third of the number of projects compared with past years, although several projects were still under consideration. “Investors find a sense of achievement in the success of the businesses they invest in, not in the amount of money they invest.”
In the interview with China Entrepreneur, Zhu repeatedly mentioned a word: discipline. In his view, disciplined investors know how to control the size and pace of investment and project valuations properly.
Large language models: Investors chasing after LLMs in the first half of 2023 mostly took a hit in the second half
China Entrepreneur: How do you see the rise of LLMs impacting technology cycle and entrepreneurship cycle?
Zhu: I think it’s a good thing. It marks the start of a new cycle. For entrepreneurs, the key is figuring out what you want to do. The era of computer and mobile internet each lasted for 10 years. The new AI era will probably last as long, 10 to 20 years.
One thing is for sure, though: it is unrealistic for entrepreneurs to develop a new iOS app of the mobile internet. Instead, they can find a lot of opportunities from applications built on LLMs. We have invested in many companies working on AIGC and they are doing very well.
China Entrepreneur: Do the changes in the cycle of entrepreneurship mean new requirements for investors? With fewer entrepreneurs and thus less demand for money, how can investors find good projects?
Zhu: In fact the key remains unchanged, that is, make your own judgment and never follow the herd. In the first half of 2023, most investors chased after LLM projects and poured massive amounts of money into them. But now they are all at a loss.
What we’re seeing now is similar to what happened when the operating systems first appeared. In the early stage, LLM companies may have no time for some minor features, leaving opportunities for start-ups. Once the LLM companies have time, however, they will be able to add them immediately. By then start-ups can get nothing.
An article online mentioned that the launch of a package of GPT-4 tools destroyed a slew of start-ups. In my opinion, start-ups should avoid directly taking part in the LLM race, or they would struggle to survive.
Today LLMs can boost programming efficiency by 50 percent. This can help companies cut labor costs from junior programmers. But to deal with the bugs created in the process, more senior programmers will be needed to check for mistakes. Many start-ups try to use AI to optimize codes or improve debug process. This is viable only in the short term. If GPT-5 is released in 2024, LLMs may be able to speed up a developer’s code generation by as much as 90 percent. By then, those start-ups will be in danger.
China Entrepreneur: You earlier had a debate with Fu Sheng on LLMs. Is there a clear result now?
Zhu: I honestly don’t think LLMs itself can create opportunities for start-ups. First, the cost is high. Second, without data and scenarios, they would not be able to optimize.
So ultimately it’s still tech giants that can benefit from LLMs. Start-ups should focus on sales, marketing and other vertical market segments they’re good at. Opportunities are huge in these areas.
China Entrepreneur: What does the commercial LLM landscape look like in China? In which stage are companies in?
Zhu: LLM companies are looking for where to use their LLM tool. To be honest, I think China’s top LLM companies can reach the level of GPT-3.5 by the end of 2023, and hopefully GPT-4 by November, 2024. But its commercial use is another thing. Since OpenAI released GPT-4, it has had limited commercial applications to date in the U.S. The situation should be similar in China.
China Entrepreneur: In which areas do you think are LLMs most commercially viable?
Zhu: For now, things like creative writing are the easiest to do. LLMs can write an advertisement, a work report, a speech, and so on, and improve them. They can also do a good job in generating images. Midjourney is a powerful image generator. In China, we feel that LLMs can be better used to generate short videos based on the large amounts of short video data. The situation is different in the U.S. Nearly all of its e-commerce websites, including Amazon, are still based on pictures. In comparison, China’s e-commerce businesses have all shifted from pictures to short videos in the past three years. All the e-commerce websites mostly use short videos, rather than pictures, to increase traffic.
Companies can use product data generated through short videos to train their own vertical models. Fancytech is a successful example of product transformation.
China Entrepreneur: Does this mean a big opportunity for investors to make money? Image and text generation is still a niche.
Zhu: We suggest investors begin with a niche market. Why? Large platforms don’t pay attention to or spend heavily on it. There will be more room as the market grows in several years.
Now is a very good time to spot cost-effective projects from non-consensus areas.
China Entrepreneur: Are leading LLM start-ups still sought after?
Zhu: They were in the first half of 2023, but investors all regretted their decision in the second half.
China Entrepreneur: So is embracing AIGC the only choice for entrepreneurs?
Zhu: Yes. And they must. In two years, any company that does not rely on AIGC would lose competitiveness. Even consumer companies and home appliance companies need to reduce costs and increase efficiency through AIGC.
And that’s very easy to do now. For example, we invested in a company with more than 1,000 employees. Four employees in its human resources department are responsible for answering questions about employee benefits; by using AIGC, the company just needs one robot to do that. How? Feed its benefit policies into the robot and train it for one week, then it will be able to answer all questions from the employees. You see, AIGC helps reduce costs and increase efficiency immediately. AIGC can also be applied in other scenarios, like image generation in games.
Entrepreneurship: Lack of money is a double-edged sword
China Entrepreneur: Recapping 2023, how many projects have you invested in?
Zhu: About one third of the number in past years. The number seems small, yet the truth is many other projects are also on my list of consideration. I’m actually optimistic about several of them and considering whether or when to invest.
If anything else has changed, it is the way I invest. In the past few years, there were times when I invested in a start-up without even meeting its founder in person. I’m unlikely to do that now. Instead, I would spend three to four months researching a project, like what I did 15 or 16 years ago. I would look at the monthly forecasts provided by the founder and compare them with the real data. I would be willing to continue only if they were consistent for three to four months.
China Entrepreneur: Have there been any changes to the pace of fundraising and investment for GSR Ventures in the past two years?
Zhu: Not much. We raised a new fund in early 2022.
Of course, there have been other changes in the past year or two. The biggest one was the share of the secondary market dropped too much, causing a serious mismatch between the primary and secondary markets. New funds will not be raised in the primary market until the situation improves in the secondary market.
Second, the U.S. risk-free rate is now 5 percent plus. When investors can get 5 percent plus current interest from their bank savings, why would they invest their money in the non-liquid primary market? This would affect the big picture, and the amount of money investors are willing to allocate to the primary market.
Last, financial liquidity. Investors who love the Chinese market most are unable to re-invest in China because their money is still locked up.
All these have restricted the flows of money. If RMB funds can’t return proceeds to their investors, known as distributed to paid-in capital (DPI), the latter will unlikely invest more. The same holds true for U.S. dollar funds. Their money will flow back to China only when they see that money flows again in the Chinese market and some excellent businesses can exit. I don’t expect a market recovery until at least 2025.
China Entrepreneur: So from this perspective, will the macro environment pose a variety of challenges to entrepreneurs?
Zhu: Yes. Lack of money is a double-edged sword. It’s a good thing for true entrepreneurs as they don’t face much noise or competition. Firms investing in VC, of course, can’t get money now.
China Entrepreneur: Throughout your investor career, did you feel any different in 2023?
Zhu: Yes. 2023 was the most comfortable year for me. I had a feeling that I was reliving a life from 15 years ago. I could just take my time to investigate a project.
China Entrepreneur: RMB LPs were active in the primary market in 2023. Has GSR Ventures considered expanding the scale of RMB funds?
Zhu: No. Discipline is very crucial. Expanding scales now is the best method of suicide.
I must control the size and pace of investment, and investment valuations properly. This is very important.
China Entrepreneur: How well did institutions do in 2023 in exits?
Zhu: It all depends on profit. During the capital winter (a significant funding slowdown and drop in fundraising activities), my hope is that every business can make a profit. Exits will happen when conditions are ripe. As long as every business can make money, there will naturally be a way to exit when the Federal Reserve starts to cut rates.
China Entrepreneur: When might be a peak time for exits?
Zhu: The key is when the Fed will cut rates. Before that, I think all large-scale exits are unrealistic.
Investment in the consumer market: Offering a PE ratio of up to 15
China Entrepreneur: What areas did GSR Ventures target in 2023? Have there been any changes?
Zhu: We have always focused on two areas, AIGC and consumer market.
China Entrepreneur: It seems difficult to spot good projects in the two areas, especially the latter, right?
Zhu: There are still good projects, of course. The problem is whether they want to raise funds now, given that valuations are too low. The chemistry between founders and investors is critical for success. When a founder is really in urgent need of money or needs an endorsement by a renowned investor, a deal will get done.
China Entrepreneur: Currently consumption potential in China’s third-tier and lower-tier cities is enormous. Many big brands are marching into those markets and some brands even focus on opening chain stores in those cities. How do you see opportunities for these brands?
Zhu: We have always been bullish about China’s overall consumer market. In the third quarter, China’s economic growth was mainly driven by consumption. During a talk with some LPs in the US a few months ago, I told them that China is the world’s second-largest consumer market, with a 1.4 billion population and 300 million middle-income earners. Every one of them is a consumer. The size of the market is unmatched by any other country.
Entrepreneurs now face less competition to win over Chinese consumers if they can offer them high-quality products at good prices. Total gross merchandise volume (GMV) on China’s e-commerce platforms has grown 30 percent so far in 2023 from a high base in 2022. A large part of that came from a 60% rise in online group discounter Pinduoduo. Some brands will definitely stand out and offer consumers better prices.
Besides, the Chinese mainland has a much lower chain store rate than regions like Japan. So opening a chain of stores in lower-tier cities is an inevitable trend.
China Entrepreneur: How do you choose a project in the consumer area?
Zhu: We look at two indicators. First, to develop a new product or brand, whether a single product can create an annual turnover of 10 billion yuan. The ceiling must be high. Second, whether the varieties of products are standardized enough to open 10,000 stores in lower-tier cities.
Profit matters too. Today all consumer companies pursue profit and seek to grow while ensuring profitability. In this case, some consumer brands may not grow as fast as we imagine, but in a healthier way.
China Entrepreneur: Is the pursuit of profit mainly caused by a lack of money they can raise?
Zhu: Yes, and that’s just one reason. Another reason is that the valuations system is completely different now. As it becomes more expensive to borrow money, companies will find it not worth their while to raise funds. They would rather pursue profit and develop with their own money.
China Entrepreneur: How has the valuations system changed?
Zhu: It reached a peak in 2021. At that time, valuations were 10 times the amount of sales for the future 12 months.
China Entrepreneur: Is the sales number an estimate?
Zhu: Yes. Usually we assume it doubles a year. So the final valuations could reach 20 times the amount of sales for 2023. But today valuations stand at 10 times the amount of profits for 2023. The difference is huge.
China Entrepreneur: How did that change happen? How was the standard set up? Did someone come up with it or is it a consensus?
Zhu: The market bubble was too big during the peak period. With everyone chasing after excellent projects, the price-to-sales (P/S) ratio reached as high as 10. At that time, you could find new opportunities from all consumer projects.
Things are different now. RMB funds are unable to exit the A-shares market and thus unable to invest in the consumer sector. So investors have no choice but to turn to profit and seek to get a dividend each year. They can only afford a price-to-earnings (P/E) ratio of 10 (We calculate PS by multiplying PE and the net profit ratio).
China Entrepreneur: Are you also more cautious toward investing in consumer projects?
Zhu: Yes. Now I can only promise founders a P/E ratio of 10, or 15 at most. If they didn’t accept that condition, I wouldn’t waste time meeting them. I may still follow truly promising projects when the founders really want to develop them, but with more caution.
Going global: Chinese entrepreneurs are the hardest working people in the world
China Entrepreneur: What’s your view on Chinese companies going global?
Zhu: Chinese companies have two advantages to go global. First, supply chains. Chinese products, especially when it comes to hardware, are at least 50 percent cheaper than what we’ll find in other places.
Second, Chinese entrepreneurs are the hardest working people in the world.
In a recent interview, Indian tech billionaire Narayana Murthy said young Indians should work for up to 70 hours a week. His suggestion prompted backlash from some Indians. Well, maybe they should look at how hard working Chinese entrepreneurs are.
When I attended the annual meeting in the US in October, I met a US dollar LP who has invested in GSR Ventures for more than a decade. We talked about China’s 996 work schedule, work 9 am to 9 pm, 6 days a week. He told me he had never heard it before. I think the hard-working spirit of Chinese entrepreneurs is a big asset for them.
Combined, the two advantages help make China’s start-ups a powerful presence on the global stage. Today global competition is becoming one between Chinese people. In social media and short video platforms, you compete with TikTok; in e-commerce, you compete with SHEIN and Temu; in games, you compete with Hoyoverse (miHoYo), Tencent, and NetEase Games. In the financial technology sector of Southeast Asia, Africa, and South America, Chinese companies are also occupying a large share.
Another example, warehousing and logistics robots exported to Indonesia are all made in China. China’s e-commerce companies deal with 300 million orders on average every day. Behind that are powerful robots, including robotic mowers and robotic pool cleaners. No other country has as many orders as China.
And Chinese people have an unrivaled ability to innovate. Take for example a shotgun sight company in Wuhan. The company exports high-tech shotgun sights to the US. Americans love to hunt. Can you imagine that this tiny gadget can bring the company an annual revenue of 2 billion yuan and a profit of 200 to 300 million yuan? The laser beam emitted by the shotgun sight can make a deer faint. All these things are made in China, and Americans could not possibly produce them.
Today we’re seeing some companies move manufacturing out of China. But what they moved out is just the lowest level of supply chains. Lululemon, a Canada-based athleisure apparel company, for example, moved some manufacturing to Bangladesh and Sri Lanka. When it comes to hardware, it will be extremely hard to move them out. They are all part of China’s supply chains. So it would be impossible to move them to Southeast Asia. I just talked about shotgun sights. The manufacturer iterates fast and launches new products every year.
China Entrepreneur: How are the companies backed by GSR Ventures doing in going global? The competition is already excessively fierce in Southeast Asia, especially for consumer brands.
Zhu: If a start-up offers enterprise services, I usually suggest that it first go to Singapore or Middle Eastern countries before marching into European and American markets. I say so for two reasons. First, Chinese products are more acceptable in Singapore and Middle Eastern countries. Second, living in these countries is a good way to improve English. You know, English is a big problem for Chinese companies. After staying in Singapore for six months or so, a founder will be able to communicate with locals in English.
As for consumer companies, I suggest entering the European and American markets directly.
China Entrepreneur: In 2023, Middle Eastern investors have been active in China’s venture capital landscape. How do you see this playing out?
Zhu: Well, they come to China with a purpose. They hope that industries could complement each other or Chinese companies could set up factories in their country to boost local employment and economic development. A company must align with their strategy to get their investment.
China Entrepreneur: Will Chinese investment institutions try appealing to their needs?
Zhu: It depends on whether your strategy suits their needs. Although we also work with LPs from the Middle East, we might not voluntarily spend more time on the Middle Eastern market. It’s not our priority right now. We focus on the Chinese market. More importantly, we want to help Chinese entrepreneurs go global, not only to the Middle East, but to the world.