Chee-We Ng is founding partner of Oak Seed Ventures, where he invests solely in early-stage enterprise and core technology startups and offers strategic advice and his network in order to help them succeed. Chee-We studied Electrical Engineering and Computer Science at MIT and has been programming in C and C++ since he was fourteen. Prior to founding Oak Seed, he was a managing partner at CRCM, a venture partner at Redpoint, led Cisco Investments in China and was a McKinsey consultant. He was the seed investor in many enterprise and core technology startups including Kyligence, Fastone, Oushu and Shangyong/Xinji. With Chee-We, we talk through several developments in China’s enterprise and cloud market, including trends across SaaS, data analytics and AI, data infrastructure and more. We also discuss the impact of geopolitics that has brought about an opportunity in “import substitution” away from global chipsets and core enterprise software. Finally, Chee-We shares his views on how to assess technical deals and founders and offers advice to aspiring technical entrepreneurs.
Adam: Welcome Chee-We, before we dive in… could you please introduce yourself and tell us more about how things got started.
Chee-We: Sure thing. I started Oak Seed Ventures in China to focus on enterprise and deep technologies. I’m a geek! I taught myself C when I was 14 to develop a word processor for my mum because WordStar, WordPerfect and Edlin were all too complicated for her. Going to MIT was a dream come true, where I participated in early research in AI and chatbots as well as integrated circuits for low-power communications. I interned at Xerox PARC, where almost everything we know about computing was invented, including object-oriented programming, GUI, mouse and Local Area Network; and Bell Labs, where UNIX and transistors were first created.
I started my professional career in information security where I led a team to build encrypted phone and hard drive which are still award-winning at CES after twenty years. Right before my investing career, I spent two years at McKinsey after Harvard Business School, where I advised telecom operators on mobile OS and Computing as a Service (before they called it Cloud).
I began investing in enterprise tech in China at Cisco. We were probably the only team solely focused on this in China. Our mandate was to invest across the enterprise stack from integrated circuits, storage, security, cloud, big data, AI, internet of Things and SaaS. We did so with a global perspective, looking for world class companies in China, and also with a value-added approach, where we sought to help our companies with our partner and customer access. I spent a lot of time taking our startups to meet our sales team, channel partners and customers. We were also value-added LPs in venture funds focused in areas of interest to us. After Cisco, I transitioned over to CRCM and Redpoint where I learnt portfolio construction and fund management. It was there I decided to focus entirely on early stage.
I founded Oak Seed to focus on enterprise technologies. Unlike elsewhere, there are few if any VCs focused on early-stage enterprise tech startups in China. Yet, when I was at Cisco, we began seeing truly world-class enterprise tech startups in China, for example Kyligence. Due to a number of factors, there are now world-class engineers and in particular technology architects and innovators in China, which are building the best enterprise technologies anywhere. This may not cut across every field yet, where Silicon Valley still has the lead, but in selected areas. Later stage funds and public markets are beginning to recognize the value of these startups, and I believe having an early-stage enterprise focused fund will be good for the industry and deliver great returns to investors.
Oak Seed’s Investment Approach
Adam: Clearly you’ve done a lot over the past 20 years and I think being a geek is super helpful for investing in some of these areas. In terms of Oak Seed Ventures, we’d love to learn more about your fund. Can you tell us a bit more about your fund’s overall approach to investing, what you focus on, what stage, what are some of the criteria you look for when it comes to interesting opportunities?
Chee-We: Oak Seed is looking for ambitious, talented entrepreneurs starting companies to tackle very difficult problems that enterprises face. Our promise to you is we will help you with introducing key partners and customers, advise on marketing and strategy, or simply being your cheer leader. You will not be dealing with a VP who’s trying to size you up, but a potential thought partner who understands technology and business in a deep way.
We are focused across the entire enterprise stack, from SaaS and tech-enabled business services on the top, 5G and IOT, to data infrastructure and information security. You don’t need to meet partners who would rather spend time with consumer investments because we don’t have any!
We like investing early, as early as just having a business plan or a prototype, because that’s where we can apply our technology understanding and market insight to identify potential winners and help make them successful. For our investors, this drives the upside potential which is important because the road for enterprise startups can be very long.
We apply a convicted approach to investing. This is possible because we can reduce market risk and technology risk through careful due diligence. This allows us to spend more time with companies to help them succeed. Our goal is to deliver 5 to 10x for our investors through patiently waiting for the right investment, working hard to win it, and committing to helping them be successful.
China’s Enterprise Cloud Market
Adam: Can you tell us more about these areas you’re investing into… it’s a mix of cloud and enterprise, some deep technologies and some semis as well. There’re few large companies in China in the enterprise space. There is no Salesforce, for example. There are a few companies that are over $10B in market cap in China whereas in the U.S. there are many such companies and enterprise and cloud has been all the rage for the past 10 years in particular. How would someone who’s not in China look at the cloud opportunity in China? Where do we start first to understand this opportunity?
Chee-We: Indeed, the IT spending has been small in China even as a percentage of GDP. There aren’t that many big companies (in enterprise software), but there are in hardware, such as Huawei, Lenovo, Inspur, or HTC. There are also a whole bunch of system integrators, many of them listed domestically. But software in particular as a percentage of IT spending is only 3% to 4% whereas in the rest of the world is usually 15%. That explains why, historically, you don’t see that many software companies at all. However, what’s changed is that with the mobile internet, today you see mobile internet, big data, IoT and AI becoming the top agenda for many enterprises. This started with internet companies leveraging all these technologies and then encrouching on the markets of traditional companies, for instance, in finance and insurance.
And now enterprises have no choice but to reinvent themselves. They have to use state-of-the-art technologies to have a mobile and omnichannel experience and they also have to employ big data and analytics in order to stay ahead. So what you can see now in the startup space is a whole school of startups in enterprise space across SaaS, applications, business services, infrastructure services, data analytics and so forth.
There is actually evidence that there will be more and more of these companies listed on the stock market as well. You can already see, for instance, over the past two years in China, in areas of public cloud, other than AliCloud which is like AWS of China, two companies have successfully listed, Kingsoft and UCloud. We’re also seeing big data companies like Kyligence and PingCap that have gone from seed to Series C in less than 2-3 years. We expect these companies to become the next set of companies that will become public and gradually change the entire IT industry. Overall, it is true that IT market in China is definitely not as big, by any measure, compared to the U.S. or the west, and it is true that it’s largely hardware and system integration, but many many startups are now emerging to provide SaaS, analytics, cloud and infrastructure solutions to enterprises which are now under increasing pressure to stay ahead and facing increasingly competitive market environment.
Adam: There’s been a slew of Chinese enterprise cloud-based companies going public. Kingsoft has been doing well recently. Agora just went public a few weeks ago and their valuation is quite rich. I realize this is a very large topic. There’s a lot to tackle when it comes to cloud opportunities and we’ll probably spend multiple episodes (to cover them all). Perhaps today we can focus on a few specific areas that you’ve been looking at. You were just talking about databases like PingCap and Kyligence. I think you mentioned SaaS as well. These are two things that global investors and founders know very well. Can you talk us through what you’re seeing when it comes to SaaS and databases in China?
Chee-We: Yes, SaaS is a very hot topic. It’s driving the value of many companies from Microsoft to Zoom to Salesforce which eclipsed Oracle in its market cap, that’s amazing. Yet SaaS is very nascent in China, but there are many startups that are trying to be the Salesforce of China. There have been many waves of SaaS companies in China. There were waves that targeted SME’s. There have been SaaS companies offering “freemium” model. Some of these startups in the past have been challenged when tackling SME market. This is a market that, although it looks big because it counts for a significant part of GDP and there are a large number of SMEs, it turns out that they don’t have deep pockets and also often don’t have a long lifespan. The average SME in China only has a lifespan of 2 years. And with the freemium model, these companies face the challenge of converting free customers to paid ones; typically, the converting rate can be as low as about 2%.
But we’re seeing new types of SaaS in China. These type of SaaS companies do not just digitize a workflow. It’s not about recording or digitizing a workflow. These are really the SaaS of the past. The new SaaS really combines analytics, AI and external sources of data. They really engage and truly automate workflows. In China the boundaries between SaaS and tech-enabled business services are also blurring very quickly. Sometimes you’re won’t see a distinction between the two. I might even say that this is happening faster in China than in the U.S. where SaaS are usually software companies and business services is very distinct. Here in China it’s blurring very quickly.
We are also seeing new trends in what’s called low code or no code platforms. This really enables end users to build their own applications that enable the workplace of the future. This is what VCs are actively looking at in the U.S. and in China we’re seeing new companies in this area as well.
Talking about databases and data infrastructure. This is an area where a lot is happening. With data infrastructure, often times there will be significant bottle necks that show up when you have volumes of data increasing exponentially. New innovation is always required to keep up with large volumes of data: new ways of analyzing and making use of data whenever the underlying hardware changes. Our observation is whereas in the past, many technologies were imported from the U.S. by multinational companies, data infrastructure is an area that, going forward, there will be many more domestic companies, whether these are the big companies like Huawei and Lenovo putting together solutions or these are startups. For instance, data analytics have gone from using Hadoop to cloud and Kubernetes to the next thing which is called serverless. We see both big companies like AliCloud and small startups trying to offer AI on serverless architecture.
Data warehousing and databases are now seeing new growth areas around graph and time series databases. Compute storage fabric has gone from conventional to converged and hyper converged and now composable. Many of these areas will see new innovations. I do not anticipate startups in the U.S. being able to enter the Chinese market going forward. I expect many of these new disruptions to be led in China by Chinese startups.
Adam: Got it. There’s a lot to unpack here. Just focusing on databases in China, we brought up Kyligence and PingCap and there’re many more projects. I’m trying to get a sense of how well these companies monetize. In the U.S. some of the database companies have done quite a good job with their open source open core approach where they offer up the code for free but they offer other services, could be security or something else, in order to monetize very well. In China, some of these companies I’ve heard aren’t really generating any revenue. How do you see them developing over time?
Chee-We: Different companies have taken different strategies. Companies like Kyligence have actually taken a rather similar approach, having open source version – the Apache Kylin – that everyone can use and extensions and enterprise versions that have more features like security, better performance and also working with end customers to deliver the actual commercial value of using analytics. That’s one strategy. Because Kyligence is delivering something that’s truly world-class, the company has also gone abroad and worked on not just the Chinese market but also opportunities elsewhere, leveraging a global traction to raise money and grow.
There are also other ways that many of these startups have succeeded. You’ll find in China, as you move from the more generic data infrastructure to providing a service, providing an actual business analytics that’s closer to the enterprise being able to generate revenue, then you will find it much easier to monetize. A lot of the spending these days in enterprises is held in hands of the business unit leaders, not so much in the CIO’s, so to the extent that a solution provided by a startup can actually drive the topline of a business unit, you will find much more willingness to pay on the part of the company.
China’s Aspiration for Tech Independence
Adam: Since we’re on the topic of some of these companies that have world-class talent or are pushing products that could compete against solutions in the US or other countries, it gets me thinking about the elephant in the room, so to speak. There’s quite a bit of tension going on between the U.S. and China. There’s geopolitics going on and I was thinking, there are certain areas in China that seem to benefit from that. It could include some of the semiconductor companies, chipset makers. A lot of these public companies in China have seen their market cap skyrocket and have pretty incredible P/E ratios like 100x. Pretty ridiculous stuff. And not just semiconductors and related companies, but also even some companies that we’re talking about right now, databases and alike enterprise. Could you talk us through what you think about this continuous tension and how that might impact the growth of these companies that you’re looking at?
Chee-We: Sure. To set the context, China has been trying to be independent, or at least less dependent, on foreign technology for many years. This started out in, for instance, semiconductors, being one of the major causes of trade deficit. In fact, the trade deficit due to semiconductors has been larger than that of oil for many years. However, what has really happened in the last 2 years, starting with ZTE and now Huawei, is that the very survival of Chinese IT companies and even the entire Chinese economy has been threatened. When you have exports of certain critical components to China being entirely limited, when you have a blacklist of companies being drawn up, and now potentially limiting the ability of companies to list in the U.S. Basically, going forward, Chinese companies and government will try to independent on foreign imports if they can.
There is phrase coined many years ago called “自主可控”. It actually came about when Edward Snowden revealed Project PRISM but now it’s on hyperdrive, which means Chinese government really wants to have its own chips and…in the past there was a phrase called “removing IOE from China”, I stands for IBM, O representing Oracle, E representing EMC. I don’t think that’s so explicit these days, but you can see the Chinese government’s desire to be truly independent of foreign technology, and this really cut across the whole stack, from chips to basic software and hardware.
Now, not all of that is possible. It’s probably very challenging to replace Intel and ARM processors, Microsoft operating systems and Oracle databases from legacy applications, but for new applications you will find Chinese companies and the government very cautious of whether it is dependent on a foreign technology. You will find that Chinese government and enterprises want to use open source if they can or a Chinese vendor. So for every new innovation in the U.S., whether it’s a new type of database (for example a graph database) or a new deck of data infrastructure (for instance composable infrastructure), going forward its very clear that these markets in China will be entirely serviced by Chinese players.
Adam: Got it. Chee-We, how does that impact your investing? Or take a step back, not just import substitution but in general, given what you’re seeing, how does it impact your investing and how do you go about finding good investments before due diligence given the circumstances?
Chee-We: That actually has been a good change for us. It means there’s a protected market. It means there’s also a lot of talent coming into this sector, which in the past is considered very challenging, because you have dominance of foreign players. In the past, trying to create a new infrastructure startup is impossible; that’s entirely done in the valley or, to some extent, in Israel. Now, with the market opportunity and also with some Chinese talent that has been in the U.S., whether it’s working for startups or just doing their Masters and PhDs at Stanford or MIT types… they’re coming back, sometimes forced to come back. They will be starting startups in areas of enterprise technologies because this is an area that they feel passionate about and they have expertise on, so that’s actually very good for us.
Now, having said that, it does not mean that you can deliver good returns by just backing every company with an import substitution theme, because you still have to spend the time to do the due diligence and make sure that, indeed, this is a company that has the best technology, the best commercial team to execute and you’re going in at the right valuation. Often times in China, when something gets too hot, then their valuations go crazy and it may not be a good investment.
What Makes a Good Investment?
Adam: What else do you look at when it comes to investing? With regards to the team? With regards to the solution that’s being developed? How do you go about conducting commercial due diligence on very technical products?
Chee-We: Sure. If we take a step back, later-stage venture in China places a premium on two things: the ability to scale really big and the perceived scarcity of what’s being built. That means our entrepreneurs need to (1) be very ambitious and (2) have a world class technical background and has or is developing something special. I like the teams that have a combination of strong technical and commercial founders. The commercial aspect is often neglected, but it’s often needed because you need to build a clear marketing message for enterprise customers. The product managers need to make the difficult choices of what to be and what not to be, also what use cases to go after and what features to prioritize. This side of things has to be driven by a clear commercial plan, not so much of a technical plan.
I also place a strong emphasis on the team being at the forefront of technology and pushing the envelope. For applications, the team must understand the business challenges and have available at their disposal, the ability to implement state-of-the-art technologies with excellent user experience. For data infrastructure, it means the team understands the bottlenecks of conventional solutions and are making architectural choices that enable a solution that is orders of magnitude better. It is also highly crucial to link technology to the use case. One must never fall into the trap of a solution looking for a problem.
To complete commercial due diligence, it is important to spend time with the end-uses, the CIOs to understand use cases, but also partners, for example system integrators which are needed to put together solutions for end customers.
I often get asked a question, how do you know how strong is a CTO or founder technically? It really boils down to this. In every area for enterprise startup, there is an envelope of what technology can achieve beyond which there are serious trade-offs. Sometimes there is even a computer science theorem, for example the CAP theorem, CAP’s short for consistency, availability and partition tolerance, that says you cannot achieve all three at the same time; you must have a trade-off.
When I do due diligence for a company, I try to get myself educated at where that technology boundary is by talking to experts and learning myself. Then I spend time with the CTO or founder and see where his understanding is. There are CTOs that don’t understand the limitations of technology or where that “state-of-the-art” is or when things will get challenging, these CTOs are not the world-class founders who are pushing the envelope. They haven’t gotten there yet and that’s why they don’t see where the limitations are. That means, whatever he’s working on, unless there is another knowhow that’s required from outside of technology, it will be quite easy for a competitor to build exactly what he’s thinking of building.
Then there are the CTOs that appreciate the limitations and trade-offs that technology forces them to make. They are already thinking of solutions and are making architectural choices. These CTOs are at the forefront of technology envelope.
Then there are, now and then, CTOs that not only appreciate the limitations and trade-offs that I am aware of (after speaking to a host of experts), but also understand other challenges and issues further down the road. These CTOs are truly pushing the forefront of technology envelope. Startups that are founded by the latter two types of CTOs, in my mind, stand a higher chance of being successful because what they build will have fewer entrants or copycats.
Entrepreneurs in the Space
Adam: Could you walk us through some of the investments you’ve made prior and give us examples of some of these founders. Why did you decide to back them?
Chee-We: Sure. Let me share with you the stories of two entrepreneurs in the data infrastructure space, hopefully they serve as inspiration to future entrepreneurs. Luke Han with Kyligence was with e-Bay in Shanghai. E-Bay, as everyone knows, has no business in China, but they have an R&D team. E-Bay wanted to perform multi-dimensional analysis on very large sets of data, which were stored on Hadoop. Think of it like performing pivot tables on Excel, but for billions of rows of data. So, they got their Shanghai R&D team to work on it and build, initially, this internal tool for themselves. Then e-Bay decided the open source what they built, which became Apache Kylin, the first ever top-level Open Source project from China. What happened next was customers ranging from Citibank to China Mobile began to use it and began to ask Luke and his team at e-Bay for technical support. The team decided to quit and create Kyligence to build an enterprise solution around Apache Kylin and help customers with their most challenging analytical problems.
Another example, Chen Xi, the founder of Fastone, was with EMC. He was probably the youngest technical director at EMC, leading the team that built EMC’s Hyper-converged Infrastructure product line from nothing to becoming No.1 worldwide. He paired up with his college classmate and former colleague who became a sales leader at Amazon Web Services in China to start Fastone, which provides a proprietary high-performance computing cloud solution. This is truly proprietary and world class, not just a Copy to China kind of strategy. He really thought through where the technical bottlenecks were and created an entirely new solution. Now the solution is now being used by industries in China cutting across genomics (which needs a lot of high performance computation to do genome sequencing) to semiconductors (for simulation and IC design) which is currently seeing a huge uplift, given the US-China tensions.
Luke Han of Kyligence and Chen Xi of Fastone, in my mind, represent a sample of world-class technology entrepreneurs. They are not doing a Copy to China strategy that many predecessors did. To me, what’s most surprising is that this is happening in the enterprise tech space in China, where what they are doing is not just best in China, but best everywhere in the world.
Adam: Definitely impressive founders. Having worked with some of these top-class founders, do you have any advice for other entrepreneurs looking to build businesses in these areas?
Chee-We: My first piece of advice is be different and choose a difficult area. The scariest thing about starting a company is after one year of diving into doing it, there are 5 to 10 startups trying to do the same. When that happens, you will find it very difficult to stand out and have to compete on price. I often hear entrepreneurs say “that’s okay, the market is big enough” and that’s one of the biggest fallacies – yes, the market may be, but the customers will get you and your competitors to outbid each other in RFPs and tenders. That basically means it’s a competition of price and a race to the bottom. This is very different in the consumer space.
My second advice is focus on commercial value that your solution brings to a customer. Technologists are too often inward looking. They try to make the world better, not a particular customer. But frequently, the world may be better, but nobody is willing to pay big money for it. Companies pay for someone to solve a pain point, not for removing inconveniences. Try to work on the top one to two priorities for your customer, which could be the CIO or, nowadays, could be the HR or finance department. You may be working on an analytical tool, but at the end of the day, there must be a commercial benefit to someone, e.g. the marketing department or the finance department. It must mean something to them.
Finally, move fast. Really fast. Be willing to make mistakes and learn quickly. The difference between a winner and the rest often boils down to being the first few and speed. In China in particular so because Chinese VCs will throw money at the first few quality teams and totally shun the later ones because they feel “the window is over” and there are players in front that have lots of money. Deliver quick results in short cycles, and always be in the market to raise money, which is a rather unique feature of the Chinese startup/VC market.
Adam: Got it, that’s really helpful advice Chee-We. Thanks for sharing your views on what’s happening in this market to me, to the investors who are listening to this podcast and also the founders who look to draw inspiration from your experiences. Chee-We, thanks so much for your time.
Chee-We: Thank you!