Ep87: Bob Nelsen of ARCH Venture Partners on Biotech Investing

Yeah, I was swearing on Twitter a lot yesterday, so there's a good chance I'll swear on this.

I tend to swear at the extremes. Good moods or bad moods.

I think it's just incredibly motivating to be, you know, with people that are way smarter than you all the time trying to solve really, really big problems that are, you know, unsolved and have major impact. So it's kind of — to me it's like the most fun job that you could have.

Yeah. I think we're just in this truly revolutionary, up-and-to-the-right kind of scientific and kind of even industrial innovation realm where a lot of vectors are kind of intersecting that really will have just a fundamental impact on the human condition. You know, more than just human health or human maladies, but, you know, things like reversing aging and fixing depression and even thinking about optimizing kind of your mental state, human performance, those kind of things. You know, well beyond just fixing a malady like COVID.

I mean, that is the challenge. And so, you know, many funds that get big have lost their edge, and they tend to either — it's the kind of the old guys who have made most of the money hand it off to other folks who don't know how to make money or don't know how to build companies, or it's they start thinking about what we would call putting the money to work, which is like a bad phrase at ARCH. It should just be about doing cool stuff and having impact. So if you can have impact with a million dollars, that's awesome. If it takes a hundred million to have impact, that's fine. And so some of the things we do, we couldn't have done. You know, some things really do need $100 million or $200 million to have impact, because you need scale and you need to invest and you need multiple things. And some things, you know, really do just need to be brought to the next technical risk reduction.

Yeah, I think, you know, we've done it in, you know, Fund 8, 9, and 10, so I think we'll be able to do it in 11, I think. But you do have to kind of just keep the scientific bar high. That's really the key thing, is to just keep that love of the scientific innovation. One of the things that we're doing to innovate within the innovation model is starting to incubate some of these smaller things within the companies we already have. Because the biggest scarce resource right now is actually not capital. It's human capital. Like it's great managers, it's, you know, great company builders, it's great scientific innovators within companies, and that is actually the hardest thing. So we're doing a bunch of companies where we're actually in — we kind of create these innovation divisions within a company, like Sana has Sana X, Resilience has Resilience X, and those X divisions are meant to be kind of the R&D innovation, high-risk money. And we can take a new idea, incubate it within those companies, and actually invest in it from the other fund, not have it be controlled by the company. It can be a completely independent company, and then kind of in exchange for being incubated, it might give a pick or two — a market-rate kind of pick or two to the company that's incubating it. So we're experimenting with this model to be able to do more things and also be able to be more effective at company building, because we've just found that the biggest risk that we have is picking the wrong people and trying to build these teams over and over. And when every graduate student, you know, with a new paper thinks that they have a new company, you really need to figure out a way to not build, you know, a thousand different companies because you're suboptimizing. So we're experimenting.

Well, I think an example would be that Prime is kind of has incubated within Beam. And we have other, you know, companies that are unannounced that are incubating within some of our other companies. So while, you know, Flagship might incubate companies — a bunch of companies within Flagship — we have a model that's kind of incubating companies within companies, and then we still have our conventional model that is incubating, you know, just new companies and doing kind of the classic venture model.

It's not even scale per se. It's more just efficient allocation of human capital, right? We just — there aren't that many great company builders out there. So when you find one, or multiple of them, you know, the amount of effort that it takes into that, you know, building that first year-one or year-two team is immense. If you can apply some of that human capital across, you know, kind of multiple projects and they can handle it. And that's how we've done it within these X divisions, is really build the X division to be, you know, kind of great early-stage risk reducers.

Yeah, I don't think we've proven that we can make them there, but we've definitely are excited about reducing the risk, you know, in that context, right? Where we're reducing technical risk within these kind of X divisions or within companies within the X divisions. And then decide whether these are, you know, independent companies we should put a lot of money in later. So that's one of the innovation models that we're playing with.

I just think it's a following-the-science list, right? And following the clinical or kind of market need. You know, clinical trials need to be reinvented, manufacturing is a huge issue. You know, we think the frontiers of science will go into kind of in vivo gene therapy and in vivo potentiation of cell state, you know, rejuvenation of cells, organs, people, things like that. So some of it is kind of — and then in the mental health area, it's, you know, nobody needs explaining why mental health is a problem. It's — I think it's kind of the problem of our time more than viruses and cancer and other things.

Yeah. I was looking at this epidemic of suicides in Japan. I mean, there's just — I think the Japanese government just appointed a minister of loneliness. Sounds like the Trump post-presidency or something, but basically, I think it's this — viruses will be handleable. You know, the next pandemics, I think we'll be prepared. The science isn't the hardest science. And it's a lot more about, you know, some pretty good R&D and a lot of engineering and scale. So I actually think viruses will be handleable. I think cancer will be a lot either prevented, you know, with early detection, or treatable. So, you know, you start kind of going down the list and then you get into the hard ones, and they're mostly in the neuro space of Alzheimer's, Parkinson's, mental health. Really hard. And then when you think about rejuvenation and, you know, kind of reconditioning the body, then you can think about liver and kidney and heart, and, you know, then those are kind of the next frontier, with the brain being obviously the most complicated, both to access, you know, from a kind of blood-brain-barrier-protected status, and also to even understand from either a genetic, a circuit, or kind of phenotypic level. So I think there's a lot of redefining of the existing paradigms that need to be done to really kind of fix the brain.

I think it's both. I think there's, you know, you can approach mental health from what I would say kind of is the soft side of it, right, in terms of things that you can do to make you feel better. And then you can approach it from the hard science, which is really the kind of the genetics and the quantitative phenotyping and, you know, the real biology that's going on that drives, you know, a mental illness or something, you know, neuroinflammation, something that might drive a disease, or, you know, whenever we figure out what drives Alzheimer's. So some of it is risky science. Alzheimer's is the perfect example of that. At the same time, there is a hook in Alzheimer's, right? We know if you have an APOE44, you're probably going to get it. So that's huge, right? We know that there's, you know, at least you're going to get it in the definition of what it is. So in other words, Alzheimer's might be more than one — probably as likely more than one disease. So at least that version of the disease you get if you have APOE44. And so there are these hooks that you have. So it's not like totally in the dark. And the same thing with mental health. Some things are very specifically genetic and some things are polygenic, and it's just hard to figure out, you know, the weights and how you look at it. But, you know, it's linked, and some things we don't really know much about. And you can get into the very soft areas, you know, like how do you feel, you know, state-of-consciousness type of things. But, you know, there are some digital therapeutics and ways to kind of manipulate brain chemistry to make you happier, which is, you know, ultimately where we want to get, right? Which is we all want to be happier and more productive. And, you know, that kind of — and in an ideal world, we'll be happier, more productive, and rejuvenated and younger, and, you know, without disease for a longer period of time, or if you have a disease, we'll be able to fix it. So I think we're really in that realm. Within our lifetimes, you'll be able to fix a lot of these things.

I think it's, you know, the biggest one when we think about delivery there is mental health. But you can imagine in gene therapy or cell therapy. If you have high-impact therapeutics that are, let's say, you have, you know, things that cure solid tumors, you know, do you just have to go to the existing system with your cell therapy that cures solid tumors, or maybe you create a cell therapy center that sits next to a big hospital, but it's, you know, essentially owned by the company that's manufacturing the cells. So you can imagine as these complex biologics whose manufacturing is connected to the function of the therapeutic, you know, maybe there is no drug per se. It's kind of a service, right? You, you know, you want — you have cancer, you come in, we'll cure your cancer by giving you the cells. But you don't have to go to Stanford to get your cancer cured. You're going to that partner with Stanford that's across the street. We might partner with Stanford in a situation like that, because, you know, things might — you know, there's like reasons to go to the hospital, but, you know, getting your cancer cured might not be a reason to go to the hospital.

Pretty much what you said. I mean, it's just me being pissed off at observing the inability of our world to respond to the pandemic. And, you know, we did a great job in some context, right? It only took a year and a half and a half a million dead people, right? So it's like I'm not satisfied that we should be patting ourselves on the back and saying like, 'Oh, this is great.' And so, you know, I think that we can do things faster, better, cheaper, more effectively, more predictable, anticipating the next problems. And at the same time, you know, solve many of the problems in gene therapy and cell therapy and gene editing that make them very kind of unpredictable, hard to manufacture and regulate, and riskier. You know, you can systematize all of those things and make them a lot cheaper at the same time. So more accessible, and, you know, more consistent, decentralized, data-driven therapeutics versus the existing system that we live in right now, which is, you know, happens in CMC, happens in clinical trials. You know, there's — it's just a messy kind of first-inning kind of — we're in the first inning, and we just have never really invested in systematizing cells or viral vectors. Instead, everybody's got their own proprietary — they got their own proprietary thing, and they do it in small batches. And, you know, and every university has one, and every biotech has one, and every pharma has one. And so — but you're thinking of like a go-to resource where these are the guys that make, you know, AAVs at scale, and be the best manufacturer of cells and genes and edits in the world, right? With the nanoparticles, like it's like, you know, an mRNA vaccine comes along.

It's all under one roof, although that roof is distributed across the, you know, North America. But, you know, it's a place where people work out the processes to make the absolute best, fastest, cheapest AAVs. Like, oh, you need to make a billion AAV vectors in the next, you know, three to six months, you know, this will be the place to do that. Yeah, we want it to be obvious that you should work with us versus doing it yourself. And, you know, TSMC is a good model for that. But, you know, the general idea is we can learn some things from semiconductors in the sense that they actually invested a lot of money and technology figuring out how to make stuff way better, faster, cheaper. And that's what drove the entire computer industry, right? And so if you can systematize more about the biology, whether it's automating cell therapy or whether it's kind of being able to make rapid cell lines or scale viral vectors faster, you know, you can have just a more effective system. And then if you go take a step back and go into the basic biology, then you can rethink how these things should be made. And that's where it takes some basic R&D instead of applied R&D, really thinking about, wow, are we even making these things the right way? You know, do we have to make antibodies in big giant steel tanks? You know, maybe we should make them in the body.

Well, I mean, your body is making antibodies all the time without mRNA, right? So, yeah. So, you know, there are some really interesting ways of thinking about or rethinking the way that things are made. And, you know, mRNA is a great example, but there are many others.