In this episode we speak with Eric Rosenblum on how to think about strategies for deep tech investing. Tsingyuan Ventures is the successor fund to the TEEC Angels, a Silicon Valley angel group of graduates of China’s Tsingyua Univeristy (China’s most elite school). The TEEC Angels were early investors in Zoom (NASDAQ: ZM), Ginko BioWorks, and others.
Tsingyuan Ventures is a $100M USD fund investing in disciplinary technology across enterprise, life sciences, and deep tech. Notable investments include WeRide, a leading autonomous vehicle technology startup in China, Otter.AI, an artificial intelligence live voice transcription company, the precision medicine company Mission Bio, and the no-code platform, Bubble.
Disclaimer: [00:00:00] This is the Propelx podcast, a discussion on investing in all things. Startups, startup investing is highly risky. Please listen carefully to the disclosures at the end of this podcast.
[00:00:15]Andy: [00:00:15] In this episode, we’re joined by Eric Rosenblum from Tsingyuan Ventures. Our topic of conversation is how to develop strategies for deep tech investing. Eric is a leading figure in the venture community and his insights warrant our full attention. Let’s get started.
[00:00:28] All right, Eric, thank you so much for taking some time to speak with me today in coming to the Propel x podcast.
[00:00:33]Eric: [00:00:33] Thank you for having me.
[00:00:34]Andy: [00:00:34] What is the history of Tsingyuan and what does it mean to be a cross-disciplinary early stage venture investor?
[00:00:41]Eric: [00:00:41] I should start by saying that we are a US fund. We have about a hundred million dollars under management and we invest really only in North America.
[00:00:49]And as you said, we are early stage. So we focus on the seed stage. We normally make $1 million checks into the first institutional round. Now, when we talk about cross-disciplinary. That has to do more with the kinds of ventures that we are attracted to. We tend to believe that there’s great opportunity that occurs when one discipline collides with another.
[00:01:13] So for example, genetics is really a lot of computational biology colliding with traditional biology. But just to go back to the fund’s history we’re successor fund to the TEEC angel fund, TEEC stands for Tsingyua Entrepreneur Executive Club. That was a set of funds founded in 2010 and there were three funds, a TEEC Angel Fund One, which was the first money into Zoom video and was extremely successful. TEEC Angel Fund Two, which invested in companies at the seed stage again, like Quanergy, which is LIDAR, Gingko Bioworks, which is synthetic biology, and several other subsequent unicorns and take into fund three, which was the first check and plus.ai, which is a leader in self-driving trucks. In 2017, a number of the partners for TEEC Angel Fund, decided to launch a new fund called Tsingyuan.
[00:02:03] Which roughly translates to from Tsingyua. The founders, as the name implies originally came from Tsingyua University in China to the US, got PhDs, founded companies, and then start investing. Long story short: the fund as it’s grown has found a lot of success investing in highly technical founders in the U S. Because of the background of the founders and of the fund, a lot of those founders have some background in China, normally that they came to the U S for their PhDs. And so, as a result, we’re fairly heavily exposed to what would be called Deep Tech, because of the nature of many of the founders that we’re investing in.
[00:02:41]Andy: [00:02:41] What do you mean by Deep Tech?
[00:02:43]Eric: [00:02:43] Normally we mean something that takes a fair amount of traditional R & D. So lab based, university-based, something that requires a technical breakthrough to be achieved. So it’s not a pure application of a technology that’s been developed elsewhere.
[00:02:56] Although there are times where the line between those gets a bit blurry. The areas that really are driving. And I think just using examples, is sometimes quite helpful… Probably the biggest area is around life sciences. So therapeutics, genetic engineering and synthetic biology, all kinds of health tech, diagnosis, diagnostics and imaging. Automotive is another sector that’s this driving a lot of deep tech, including batteries, the navigation systems for self-driving cars, the sensor suite that you need to, be able to have the intelligence for the car to know where it is,and a number of other areas. In general, deep tech requires some kind of technical risk, in order for it to be considered part of that category.
[00:03:41]Andy: [00:03:41] When you’re talking about the development of deep tech, and you had mentioned lab and university work, how do you distinguish between, something that’s novel research, versus a meaningful investment opportunity? Where do you see that line being drawn?
[00:03:56] Eric: [00:03:56] Yeah, it’s a great question.
[00:03:58] And it actually bleeds a little bit into a related question about why this isn’t a really popular category in the US. And, and I’ll answer it a little bit backwards, which is first why this is not a very popular category in the US and this should kind of get to your initial question.
[00:04:15]Most VCs like to see some evidence of market traction . And so even at the seed investment stage, they want to see some proof of concept, some early customer adoption and then work forward from that point to say, how big could this be? Is there evidence that product market fit will be achieved within X number of years? And therefore, how do we value this investment?
[00:04:36]Because of the nature of fundamental research and development – you’re often talking about years and years in the future before you’ll have that kind of proof point. And so it is quite difficult to make this assessment. Something that we’ve learned is that it is incredibly important to have, research scientists that have some idea of what the commercialization path is going to look like -even if it’s several years off in the future. They can’t just be pure researchers. They have to have an idea of the pathway from a small scale lab experiment to scaled lab experiment, to get into manufacturing, to get into a joint development, et cetera.
[00:05:14]And that’s hard won knowledge. I think that we have had many investments with scientists that probably didn’t have enough experience or understand the commercialization path ahead, and it can become quite difficult. And so I’d say to answer your question directly, the difference between things that are just pure science experiments versus things that are investible is having a pretty good understanding of the commercialization path and whether or not it fits within an investment time horizon.
[00:05:41]And you, you are taking on additional categories of risk. On the flip side, compared to many other types of investment, there are a lot of non-dilutive funding sources.All kinds of government grants, university grants, and other kinds of support that can help mitigate some of the scientific risk, and some of the technical risk, and so there are some advantages to being in this area. But in general, your time horizon is pushed out much farther, and you’re thinking about many more forms of technical risks.
[00:06:08]Andy: [00:06:08] How do you gauge when a technology is ready or when a deep technology is really ready for a market? When you’re thinking about new technology trends, how much are you thinking about, this is really the right time for this type of technology versus this is really the right team for this type of technology?
[00:06:26] Eric: [00:06:26] Yeah, it’s a good question. We’re a ten-year fund. And we do need to think about when the first exit opportunities are going to occur from every investment that we make. Whereas in software, we may be willing to invest in a pre-seed where it’s just an idea – sometimes the person hasn’t even left their dorm at Stanford and we’re still willing to put a check in even before the company has formed, because the nature of a software startup is that once they get going, maybe, half a year, a year of development before they have their first alpha and beta customers. But the scale can be quite quick after that.
[00:06:59] In a lot of areas that are under this heading of deep tech, it may be years before they even get to the point where they’re working with the external customers. And as a result, we do not come in that early. So we don’t come in when, when it’s just a professor and an idea. Often we’re coming in when the first lab experiments have been quite promising and they’re ready to move out of the university lab and they’re starting, to engage with the first potential customers, for example.
[00:07:25]Even thoughit may seem like it’s even a little bit earlier than many software ideas, they’ve actually generally gone through years of work before we’ll even come in. There are firms that come in much earlier than we do that work again with a professor and an idea and help them with the early stage of commercialization and really thinking through building up the core team, from nothing. And so we have to be a little bit careful about when we’re come in to make sure that the probability of it exiting in a 10 year time horizon is very, very high.
[00:07:54]Andy: [00:07:54] Is there a different framework or set of risk factors that you evaluate as part of your overall investment methodology? How does it look different than how people think about venture investing?
[00:08:04]Eric: [00:08:04] It is quite different and I’d say in some respects it can be a more complicated evaluation and in some respects, somewhat easier.
[00:08:12] So one of the areas that we really liked about deep tech is that we think it’s an area that’s very difficult for most US VCs to evaluate and part of that is talent. So if you look at the composition of partnerships at most us venture capital firms, they’re normally not that technical themselves.
[00:08:30]And so I forgot – I should look this up – I saw a study of deep tech VC investment firms, and it’s something like fewer than 20% of the partners have PhDs themselves. And these are the ones that are investing in deep tech. So forget about those that really don’t have this as a focus area. So even among the ones that are focused on deep tech, they tend not to be that technical.
[00:08:51]First, it’s hard to evaluate. The technical evaluation component is obviously much more rigorous here than it is in many kind of software investments where you’re making more of an assessment around a pain point and a business model. And even then of our nine partners, eight of them are PhDs in various fields, but even then we can only cover so much.
[00:09:09]And so we’re fortunate that over the many years that we’ve been in business we’re able to draw upon several hundred entrepreneurs -75% of the founding teams in our portfolio have PhDs – and so we’re able to pull on a pretty big pool of people to help with the evaluation.
[00:09:24] Number two, that is a little bit different, is there’s often a global perspective to some of these investments. And again, because the U S moved away from manufacturing, many years ago or many decades ago… oftentimes, if you think about like material science, there may have to be a moment where the startup eventually gets into supply chain and say China or another manufacturing centric economy. And so it’s important for us to know, will this company be able to first succeed and get those overseas partnerships?
[00:09:56] And second, would they be able to attract overseas capital? Because once they go down that road, often the investors that are going to go in for their B round or their C round are actually going to be Chinese experienced VCs that actually have quite deep pockets for things like battery technologies, or, other kinds of material science in a way that there aren’t really us investors playing at that level.
[00:10:17] So sometimes after we’ve done the evaluation and we think that this is technically feasible and we think that the founding team has a good idea of the commercialization path, we also want to have a good sense of who the series B and series C investors might be. So if this does succeed, when the big check has to come for large-scale manufacturing, for example, that we already know who is likely to be.
[00:10:37]Andy: [00:10:37] Oh, that’s interesting. And so do you think about the returns profile, for these investments differently than you would in say your software portfolio?
[00:10:47]Eric: [00:10:47] Ultimately everything gets compared against everything else. So you always want to think what is the best use of a marginal dollar, right? And so you have the same hurdle rate that is applied to everything. On the other hand, one thing that I think has been really successful for Tsingyuan in the past is having a portfolio that includes, our mix right now is about 20% pure what we would call deep tech, which actually excludes life sciences, 30% life sciences… so those two together, 50%, and then 50% software. And so having a portfolio that’s pretty diversified, we think has been really a big part of our success. And each of these areas we’ll come up with some big winners.
[00:11:26] So Tsingyuan Fund One, probably the most promising single company in our portfolio is Ambilite, which is smart glass. Which is very much in the deep tech portfolio.That’s , chemistry and physics, into thin layer, film applied to glass that allows you to face change glass. But we also have some big winners in the life sciences area, including like Mission Bio, which is positioned medicine, platform and single cell sequencing. and then we have some big winners in our software portfolio.
[00:11:52] And so I think that it’s quite good to be able to diversify risk cause different sectors go through different cycles, but we try to apply the same hurdle to all of them. We don’t have like a lower bar for potential returns for deep tech, for example.
[00:12:05] The difference is really again how we do the evaluation. So we’ll be bringing in a lot more people at the early evaluation stage on the technology side versus a software deal where we’ll spend a lot more time talking to potential customers, for example.
[00:12:19]Andy: [00:12:19] I’ve heard you talk about looking back andreviewing the portfolio of the TEEC And you kind of came in at a moment where you had a unique perspective on what were some of the key drivers of success. Just briefly speaking about the deep tech portion of that. Were there any key insights from that study that were surprising to you?
[00:12:37]Eric: [00:12:37] Nothing that I would say was overly surprising, but I’ll tell you a couple of our conclusions .
[00:12:42] Our team has always been a group of technologists, except for myself to be blunt. I’m the only one without a PhD. I’m the business guy slash product person, and my skillset is for the complimentary to my partners. And just briefly,I spent my career in software. And most of them, spent their career… some of them are software PhDs, some of them, we have one material scientist, one physicist, so pretty diverse set of, of scientific disciplines.
[00:13:08]But our strength is clearly that we’ve been operators and have acted as fairly senior engineers and product managers for most of our career. And so learning number one is we realized that first: when the venture itself has technical innovation, we tend to have done very well, when the venture is a pure business model innovation, we’re pretty hit or miss.
[00:13:31] Now, some of the biggest companies in the world are business model innovations. So Amazon was initially a business model innovation as is Uber. And so what we learned is we’re just not very good at picking those, we’re much better at picking what has a true technical breakthrough and accurately assessing whether or not it has a shot of making it,
[00:13:52] But I’d say more particularly… I think that it’s been important for us to align each part of the way our firm operates and what I mean by that is our strategy and our staffing model and our sourcing model and our support model all should be well aligned with each other. So our group of partners, we have specialties in different areas that helps us understand what kinds of deals we’re able to source and evaluate. That helps us with our evaluation model. That helps us know where to invest, to help support the group of companies we’re investing in. and these things are all pretty well aligned. Our big learning was it’s really important for a small fund to have a focus and alignment along these dimensions, you know, your strategy, your staffing, your sourcing, your evaluation, and your support have to be well aligned.
[00:14:46] Andy: [00:14:46] What I really hear you saying is that you have a certain almost value creation perspective and that’s premised on the network that’s been developed and then the utilization of that network and perspective. And so it’s really trying to find a fit for where can that be the biggest driver of potential returns? Is that kind of how you would characterize it?
[00:15:08] Eric: [00:15:08] Yeah, I think that’s right. And I’ll, I’ll be a bit more specific about this. So most venture firms are quite local and quite localized, meaning they invest in their backyard and they invest in what they know. And so the venture firm is quite old school in that respect. It’s very network-based. And so if the founding team of venture firm originally came from Google, then there’ll be investing in a bunch of engineers that came out of Google. And this thing tends to self-replicate. Now our firm came with a group of people who got their PhDs in the U S who came from one university in China, which is the quote unquote MIT of China.
[00:15:44]As a result, their initial investments were often in other very technical people who originally came from China. And so you start to get other referrals from that group of investments. And so your portfolio starts to, self-replicate and you start to go deeper and deeper and deeper.
[00:16:02]So, what we realized though, is this group of founders… so 18% of US PhDs have Chinese passports. And so it’s a huge group. And we realized that a lot of venture firms actually don’t know how to penetrate or evaluate this group. But furthermore, this group is skewed towards more scientific disciplines. many of them are doing fundamental physics research. many of them are doing material science research. And again, this is an area where the U S venture ecosystem was not necessarily set up well to understand. And so as a firm, we then have to invest in making sure we’re able to evaluate and source, these kinds of founders.
[00:16:42] But then, many of them, frankly, have an issue themselves with these founders. They have often not run sales teams before. They don’t know what go to market strategies might look like they may not even be good at communicating. And it’s often not an issue even of Chinese versus English. Even if their native language is Chinese, they may not be great communicators in Chinese either. It’s just not the pathway they took in life, right? They, the path that they took in life was being in a lab, and so suddenly you have to help with a narrative and communication, and strategy. How do you package yourself for the next round of investment?
[00:17:17] And so as a firm, you start to line all these things up around what capabilities do we have to develop to be able to help these firms?
[00:17:25] It really just starts with sourcing and evaluation. if you want these companies to be successful, you also have to invest in the capabilities, to allow them to grow as companies. So everything from management to communications, to branding, go to market, et cetera.
[00:17:40]Andy: [00:17:40] And so when you think about your own fund, how has the group of people that you’re investing with changed over time or the group of people that you’re investing in?
[00:17:51] In particular, what I mean is you said that there’s this aspect in the beginning of self-replicating and yet there’s also this aspect of looking for interdisciplinary and cross border and cross-disciplinary, so you’re always this edge of change.
[00:18:07] How haswho you’re investing in changed as the fund has matured?
[00:18:11] Eric: [00:18:11] Well, that’s a great question. And, and just to go back and make sure that this is clear.
[00:18:15] Absolutely right, that we invest behind anything that is early and has technology. It doesn’t even have to be deep tech. We have deep tech as a part of our portfolio, but we invest in things that are early and technical in some fashion.
[00:18:29] And it doesn’t matter what university they’re from. As long as it’s North America, we will look at it.
[00:18:35] Now, in terms of the kinds of venture firms that we’re co-investing with, part of our strategy is to help bridge these kinds of founders to the world of quote unquote, a traditional VC. And so we have a lot of VCs that trust us and our evaluation abilities and our sourcing abilities.
[00:18:53] Because we may be penetrating a different, entrepreneur step than they do, and they trust our ability to really look at certain areas.
[00:19:01] Now there’s some clusters, so like automotive, for example, we’ve done quite a lot in self-driving. So we’re the seed investors in WeRide, which is the leader for self-driving cars in China, plus.ai, which leader for self-driving trucks in China, again, but both these firms are based in the US. We’ve invested in five different battery companies at this point that are all looking at, all really, trying to take advantage of the ride is of electronic vehicles . Sensor companies, so both LIDAR and radar. And so in those, there are a group of companies and firms that invest in the automotive sector, that we’ve co-invested with quite happily and are finding more and more things to look at with them.
[00:19:44] The next big area is therapeutics and genetics, and we have a pretty big cluster in synthetic biology and genetics. And this is again, highly cross-disciplinary kind of core deep tech, and, it’s also a very much a cross border type of deal where most of the medical innovations that are here in the US you’ll also see them in China and vice versa.
[00:20:07] By the way, I should’ve mentioned automotive is also a global market. So battery companies that succeed here. Also, we’ll find a market in China and vice versa. Automotive sensors that find a market here will also find a market there. And so these are things that our global perspective also comes in handy. We do tend to co-invest with a lot of sector specialists, that may value, other aspects of our ability to evaluate and then help with the cross-border aspects.
[00:20:34]And then there’s another cluster that we have, around materials, and chips where you have lot of strategics like Applied Materials or TDK, Intel, et cetera, that we’d like to co-invest with.
[00:20:47]We have a company that is essentially lens on a chip called a Metalenz, out of Harvard that we think is revolutionized all optics . So they’ve essentially shrunk a lens down to a chip form factor, which is just a huge breakthrough.
[00:21:01] Another company called xMEMS, which is the first true MEMS speaker. So the speaker in your headset, if you’re wearing those little iPod or AirPods speakers or headsets, there is a little, vibrating membrane inside of that. So this is a traditional speaker, just shrunk and therefore its sound quality isn’t great. It also takes a lot of energy. In the future that will be replaced by a chip, but no one had successfully done it before xMEMS. So these are two examples of essentially taking a conventional optics and changing it to a chip, or conventional, speakers and changing that to a chip where it’s both the technical breakthrough and the fact that this is a multi-national, supply chain, where we come in as experts in both and work in both cases with, the strategic investors that are very exposed to material science. so that is a big part of our strategy, connecting some of these very technical founders to the more standard world of VC and also very strategic investors.
[00:22:01]Andy: [00:22:01] So you’d say that you have a different global perspective and now thinking through not quite the nuts and bolts, but there’s a lot of other moving pieces that make companies great, including you’ve mentioned a couple of times kind of understanding where follow-on investments or the right investors will come in.
[00:22:20]What are the major failures of the past that have been learned by early entrepreneurs and early investors in this space.
[00:22:27] Are there, are there certain things that you think are different now than 10 years ago when TEEC first started?
[00:22:34]Eric: [00:22:34] It does go back to the venture staffing model.There simply isn’t that much deeply technical staffing at the average venture firm. And so first it’s not something that most venture capital firms feel that comfortable with to begin with.
[00:22:47] However, having said that, venture capital firms have a lot of herd behavior. And so there was a rush into clean tech, starting about 20 years ago, that blew up spectacularly. There were many high-profile failures in batteries, and various kinds of alternative energy. As a result, everyone got their fingers burned a little bit, and so backed off of some of these, but also extended the lesson more generically to say, well, highly technical ventures may just be too risky given the returns.
[00:23:18] And that is partially correct. Software companies may have less risk, higher return profile, higher margins, et cetera. And so it might just be a rational assessment of the economic risks. We think that it was an overcorrection. So we think that actually, no one really felt that comfortable with a lot of these areas anyway, and they had a decent rationale to not focus on the, on anything in this area and then kind of retracted from it.
[00:23:47] The things that are different now. A lot of these areas, frankly, are looking a lot more like software. So when we think about genetics, it is something that a lot of people can understand that came from a software or math background. Like I mentioned, the collision of computational biology with pure biology and genetics. When you open the hood on some of these firms, also, what you’ll see is a lot of people work in robotics and automation. Gingko Bioworks or others, part of their secret to their success is not just the ability to synthesize new proteins, but to run thousands of thousands of automated experiments with highly automated laboratories. A lot of these are areas that pure software people are quite comfortable with. And so I think you’ll see more and more people starting to reassess more deeply technical areas in the near future.
[00:24:32] The other thing that’s happened is, it’s just kind of a lot cheaper to run experiments, with the advent of various AI tools, for everything from pharma discovery to material discovery, you can be running thousands or tens of thousands or hundreds of thousands of simulations around, molecules and materials that allow you to skip some of them more, human intensive lab work that one had to go through before. So I do think that some of the infrastructure is changing and the setup costs are changing, and so there are some areas that are kind of ripe for re-evaluation as investment categories.
[00:25:08]Andy: [00:25:08] Within deep tech, broadly speaking, are there certain areas or certain intersections of, technologies that you’re most excited about in the next five to 10 years ?
[00:25:20]Eric: [00:25:20] The big platform is genetics. The applications of genetics are going to be profound. They’re already quite profound. But it has a long tail where there’s going to be breakthrough innovations from a five to 10 year horizon. And we’re only beginning to see that impact on everything from precision medicine to synthetic meats to all kinds of applications of cheaper and cheaper sequencing.
[00:25:42] On the even more generic side to collision of AI into dozens of fields will transform each of those fields. And so we look at our own portfolio. We have a company applying AI to MRIs and pet scans called Subtle Medical that we think has revolutionized the way hospitals do scans. We have AI applied to security. AI applied to, obviously, various automotive applications. And so in general, the AI is accelerating the kind of collision of software into many other fields.
[00:26:15]I’d already mentioned that it can dramatically lower the time and cost of new drug development or new material development. We have this applied into battery management. And so you have a big shift in software techniques that suddenly have relevance into a lot of areas that, previously might not have been considered investible by a software investor, but now, people that at least have strength in AI are able to evaluate a lot of other areas that are being influenced by AI.
[00:26:45]Andy: [00:26:45] Eric, thanks so much for your time today. It is absolutely fascinating to speak with you. And I know that you have materials that you’re publishing on your website and, there’s a bunch of other really good, speaking engagements that you’ve had that I’ll also share in thenotes for this podcast.
[00:26:59] So thank you so much for your time and for your insights.
[00:27:02] Eric: [00:27:02] Andy thank you. This was a pleasure.
[00:27:04]Andy: [00:27:04] I hope you enjoyed this episode as much as I did. For those of you who are curious, Eric has written several insightful articles on the Tsingyuan blog about topics of cross-border intellectual capital transfers and investing in cross-disciplinary disruptive technologies.
[00:27:18] If you’d like to learn more about Eric and see what he’s up to visit T S I N G Y U A n.ventures.
[00:27:29]Likewise, you can keep up with our team by following us on Twitter and LinkedIn. And check out our website, where we have all of our podcasts episodes. And recordings for all of the public events that we do as well. If you’re a credit investor, looking for deep tech investment opportunities do create an email@example.com slash Joel.
[00:27:50] Registration is free. And it’s for accredited investors only. Lastly, if you liked this podcast, please share it with your friends, rate it and subscribe on Spotify, Apple, Google, or anywhere else where you find our podcasts. I’ll see you next time.
[00:28:10] Disclaimer: [00:28:10] Propelx is a funding platform, not a broker dealer. Securities are offered through Hubble Investments, member FINRA, SIPC, and an affiliate of Propelx. Private investments are highly illiquid in risky and are not suitable for all investors. Past performance is not indicative of future results. You should speak with your financial advisor accountant and or attorney when evaluating private offerings. Neither Propelx nor Hubble Investments makes any recommendations or provides advice about investments. Additional information and disclosures can be found on our website, propelx.com.