Founder Stories: Bob Nelsen, ARCH Venture Partners

earliest whispers, even when the data are too early for just about anyone else. So UCSF people, take note. This is the guy you want to talk to if you're doing a startup. How many of you are doing startups in here? I didn't think I asked that. OK, great. So often, early companies were just gut instincts, was another quote from Forbes. So I didn't write this stuff, but someone else did. I've got to read this Lee Hood one. He appears random and scattered and marvelous at avoiding focus. But in the end, he has been a genius at creating companies like Illumina that have really transformed whole industries. So I think there's a little more to randomness than that.

That's probably correct.

That's probably correct. So Bob grew up in Washington. He went to the University of Puget Sound, majoring in biology and economics. Which is an interesting combination. He did an MBA at University of Chicago, and from what I can tell, that really, that really made this path happen. So he is co-founder and managing director of Arch Ventures, which is not a Silicon Valley firm. In fact, its roots are out of the University of Chicago. And the thing that we should all remember the university is this was a university's venture fund. So we don't have one yet at UCSF, but maybe we'll get one. The Arch Venture Fund in 2007 fund had a 47% return, the 6th best performing venture fund of the decade. 24 of his companies had IPOs, 16 reached billion-dollar valuations. So he's just had an amazing run in this 30-year career. He co-founded Icaria, which he sold for $2.3 billion. Kythera was sold to Allergan for $2.1 billion. He was a first seed investment for Illumina, which has a $21 billion market cap. He founded Kybella, or invested in Kybella, which sold to Allergan for $2.1 billion, and the list goes on. So what we're all fascinated about is sort of the newer investments, the grail which is a blood test for all major cancer types. He actually pulled off a $1 billion Series A and B. And when I read that, I read it before the B had been raised, and then I had to go back and say, well, did he actually do that? And then there it was, he raised that. Juno, we all know, is cancer immunotherapy, raised $310 million and did a $265 million IPO. Denali Therapeutics, some of you probably saw Ryan Watts here not very long ago. Was a $217 million A round at the time, the biggest A round ever done in biotech, and attracted Marc Tessier-Lavigne as chairman, who at the time was Rockefeller University and now is Stanford. The syndicates who invest with Bob and Arch Ventures read like the who's who of biotech venture capital. There are pension funds, family foundations, and even a sovereign wealth fund.

Fund.

So this man is quite extraordinary, and we're very fortunate to have him here. And so thank you for coming, Bob. So we're going to do this like an interview, see how this goes, um, mainly unrehearsed because I had 10 minutes with Bob a couple days ago while I was driving. It's like, okay, I can pull over and write notes for a while, but it's hot in this car and I need to get going. So what I learned is about your childhood strawberry picking. Why don't we start there?

First, we have to have a pool on how long it's going to take me to say the F word. Oleg, you have any thoughts? Sorry, I swear a lot. I will try to cut it down in this public forum. I actually, yeah, I used to once have this theory that everybody that worked for me should have worked on a farm. And it turned out that that narrowed the applicant pool significantly. But I kind of, so I started working on a farm when I was 9. And so my Social Security statement, when they send you that thing, you know, starts when I was 9. And then I started saving that money and started investing in the stock market when I was 12. I was 11. So I did have a certain risk-seeking gene that was kind of ingrained in me from the beginning. So nothing has really changed that much over that time. But I do think that it engenders a strong work ethic. And if you're going to do seed venture capital as opposed to later-stage venture capital or private equity where you just kind of sit around and take a couple meetings, you actually have to work really hard to try to make these companies go, and it's not without, you know, if you're not in for kind of 80 or 100-hour weeks, you shouldn't be in it.

So just go back to the stock picking thing. At age 11, what on earth made you decide you wanted to get into stock picking?

My mom encouraged me to, like, because I had $1,000 that I had saved up, and so she encouraged— or $900 or something.

That's pretty good for an 11-year-old.

Yeah, I know, it was better than some. And so she encouraged me to explore the stock market as a place to put it. So I bought— first I bought White Westinghouse, which was kind of a conservative, you know, down-the-middle bet. And then I actually— where I made a lot of money was back in the first MR company called Phonar, which actually invented MR and then GE ripped it off. Or there's— I'm sure there's a technical term for improved on the technology by stealing it and then, and then, you know, made more money than the original Phonar guys. But I traded that stock for a long time. Wow.

Okay, so you grew up and you went to college and then you went to grad school, and then how did you end up in venture capital?

I read an article one day. I was at University of Chicago in my first year MBA, and I read an article in the paper, actually it was the Wall Street Journal, about a fund, not a, a company that had been created by the trustees of the university to help commercialize the technology out of the University of Chicago. And they had kind of described this, you know, person that they wanted to hire, which was the first employee, and I just remember seeing the salary, which was $40,000. And what year was that? That was like, I don't know, '86. Was that a lot of money then? That would be a lot of money then, yeah. Although I set my original salary when I got the job at $1,200 a month because I was a bad negotiator at that time. So I remember turning down like $82,000 a year when I was— so I went straight out, so I would have been 23 or 22, 23, and turning down $80,000+ a year, which would be, you know, equivalent today of like, you know, it was higher than McKinsey at that time or Goldman Sachs. And to work for $1,200 a month. And literally every single person that I knew except for one guy told me to take the And this one guy told me that I should do what I wanted, which turned out to be all right. But not as good as the money. It took a while. But, you know, so basically that's how I started, by essentially volunteering, and then I eventually asked for for a salary. So my first year of business school, I was volunteering, kind of hanging out with professors to try to figure out what to do with their technology. And that's actually how we created this program that we called the Arch Associates. So we had up to 100 business school students working for us as volunteers, including like George Kadifa, who then went on to run the Hewlett-Packard software division and Corio and a bunch of other places. You know, our alumni have kind of gone and done some really interesting things.

So you tried to get in touch with Brook Byers someplace, was it looking for your first—

Yeah, I wanted to be a venture capitalist, but I didn't really know what that was. I actually didn't know what an investment banker was either. So on my application for Chicago, I put on there that I wanted to be an investment banker because I because I knew they got paid a ton of money. And so, and then once I figured out what they actually did, I decided I didn't want to be an investment banker. But I did buy this thing called, so there wasn't the internet, obviously, or cell phones or voicemail. So there's this thing called Pratt's Guide to Venture Capital, which is the most expensive book ever made as far as I could tell. Like $250 or $300 back then. And so I bought Pratt's Guide to Venture Capital, and I would just go through there and like look at the different firms, and I decided that this firm called Kleiner Perkins was cool, and they had done Genentech and a bunch of other stuff, Tandem, et cetera. And so I called them up, and I called this guy, Brooke Byers, who had, so Brooke would have been, I don't know, not that old at that time. It's not that I had dinner with him yesterday or something, he's older than me. But, and so I had to call him 35 times actually before he took the call. But I wasn't, I think when I got rejected from Stanford, Stanford, or I actually got on the waitlist at Stanford. My adjective, they make you choose some dopey thing where you have to choose an adjective to describe yourself, so I chose persistent. And so I was persistent and eventually he called me back and we had a good talk about how to get into venture capital, and of course there is no algorithm to get into venture capital.

And this was in the '80s?

Yeah, this was like '85, '86. Super early.

Yeah.

And there was nobody even knew what venture capital was back then outside of like a few places in the Bay Area. Like you would go on a plane and sit next to somebody and you would say, well, you know, I do venture capital even in the late '80s and they're like, you know, a business person, what's that? Wow.

So you were persistent. You tried in the morning and at night and in between and eventually picked up the phone. Excellent. Stays with you.

This— yeah, irritated Mark Tessier-Levine long enough to start a neuro company that eventually decided to start Denali. So he's like, you know, actually Steve Gillis said that, but basically that's how he decided he had to work at ARCH to get me to stop calling him.

So tell us about ARCH. What's the philosophy? Um, you know, you said it started out of university.

Yeah, so we were kind of bred in the, you know, the bowels of the University of Chicago ivory Tower, which was basically a science-driven, faculty-driven institution, that it really is faculty-driven. So it's all about kind of the science and the data. And so we were, you know, nobody there admitted to wanting to make money. You know, here, kind of now in these days, like it's kind of out there, right? I mean, people, I was negotiating with a professor, a very famous professor, a professor on the East Coast recently, and you know, I mean, it's really out there, right? I mean, they're very—

It's all about the money?

No, it's not all about the money. It's just, it's, you know, it's some about the money. I think it's first about the science. Almost every, almost all the people we deal with, it's first about the science. But the money, you know, is a marker. It's being treated fairly. I remember when people— the reason ARDS was created at the University of Chicago was that Gene Goldwasser, as a scientist there, purified a molecule called erythropoietin, filed an invention disclosure on it, which would have dominated Amgen and GI. And the university decided not to patent it. So the NPV of that decision is actually greater greater than the entire University of Chicago endowment today. Okay, so it's big money that they lost. That's probably one of the biggest drugs ever, right? And take a royalty on that. And so the trustees actually, the reason ARBS got created is the trustees decided that they were just as smart as these Stanford and Harvard people and how come we didn't have startups around. So we were given an incentive. We got to keep 20% of the equity and 10% of the license royalty in a bonus pool. So I think when I was 27 or 28, I made more than the president of the University of Chicago. Wow. And—

So giving up that earlier deal with $80,000 was a good decision.

Yeah, no, it was, it was okay. I mean, that company actually was a K-6 math curriculum. Company that allowed me to do okay that year. But it was out of the University of Chicago. But I think it just— the money's a piece of it, but the science is really what motivated me. And, you know, being kind of learning in that place was a really good place to learn about good science and to just, you know, hear the history of the place and kind of, I think once you're kind of forged in that, that stays with you. And in biotech, I was just explaining this to a very famous tech investor about a couple hours ago, you know, biotech, the data and the science matters, right? If you make the second best dating app, you know, you're still gonna probably get you know, successful, probably acquire an STD just like the first, the first best dating app. But like in tech, it very rarely matters who has the best technology. Sometimes it does, and you can argue like semiconductors in some places where it really does, but most of the time it doesn't. In biotech, you know, nobody wants the second best cancer cure, right? I mean, it's just one of those things where if the data really does matter and the science matters and there is science that's better than other science and you can't make that up. So we grew up starting with the science without management teams. So it never was this kind of, you know, A management team, B science. I mean, we usually had A science and we're pretty good even to this day at picking A science. Sometimes we picked B management team. Or C. We try not to do that. We try to consolidate our errors actually in one company. So we have certain companies that we use as examples for that where we actually picked the wrong science and the wrong person and like everything was wrong.

Do you want to tell us about one of those?

Well, yeah, let's see. Probably could learn from things that you've I'm trying to think of a good biotech one. Elixir was a really good idea. It was just too early. And we also, see, I almost said the F word there. Technically, the word is fucked up, the management decision. So we hired the wrong guy a couple times. And, you know, it doesn't mean they're bad people. It just means that they're not appropriate for the situation. And in that case, actually, somebody else kind of ran with one of the other co-founders and the same idea and sold their company for $700 million because they knew how to sell. And when you're doing an anti-aging company, which was the first anti-aging company, you had to be able to sell the dream. You couldn't really sell, you know, part of the dream. You just have to sell the big picture. And even then, the science was too early. But, and the people that sold the company for $700 million, their science was too early too. And the pharma figured out that, you know, it was only maybe 10 or 15 years too early, but, you know, from an NPV basis or however you do the math, it's probably not a good investment decision. But that same science is coming around again.

So you deal in very early. How do you know when it's too early?

You don't really. You know, I mean, you just have to— there's a certain amount of faith that occurs. I would say that you have to have a certain amount of scientific judgment, and you can't get too mired in the science either. You kind of have to have an algorithm of knowing that if you apply smart people to problems, you can solve them without getting too mired. So I think most ventriculologists have like MD PhDs, if they use them, are not very good. If they don't use them and they have other instinctual, because if you get, it's like in your own lab or the lab next door, right? You probably don't even know what's going on in the lab next door. So it's very hard for people to think about applying their technology to a different problem or a different discipline when they've spent the last 30 years thinking about one aspect of their technology. So I'm reasonably good at kind of templates and thinking about, you know, where things might go. And I think to do early-stage venture capital where you're actually starting the company like we do about half the time, you know, you have to be able to picture what the future is gonna be like, right? You have to be able to say, okay, I think we can cure a bunch of infectious diseases. How come we haven't? And, you know, we should put a bunch of smart people and And I think the science is at the right place that we can do it differently than all the other people have done it before. All the other very smart people have done it before. Or the same thing with Alzheimer's or whatever.

So how do you figure that out? What are your sources of information that lead you to a hypothesis?

I usually start with the problem, right? And figure out if the problem is very big. And then, well, actually, half the companies, Usually the bigger ones start with the problem. And about half the companies start with an invention. So like gene editing, right? So we're doing a company now in nucleotide editing. And that's just a really hard thing to do. If you could do it, it would be really cool. And so it's, you know, get it. Somebody called us up, said, I can do this. And we said, that's cool. Let's start a company. And then, you know, you pull in the best people and try to manage that into some, you know, funded entity. The other way is to say, gee, I think pharma and everybody else have kind of screwed up the search for Alzheimer's drugs. And things are different now in the sense that science has progressed in our understanding of, you know, maybe how to deliver drugs to the brain or how, you know, how to measure things in vivo to be able to actually know, you know, to get feedback loop on what you're actually doing. So let's take a broader approach, kind of a systems approach to that with really smart people. And if you do that at scale, i.e., with a lot of money, maybe you could approach people who normally wouldn't join an effort like that and kind of get this kind of amazing— and Denali's a great example of that. Like, if you meet the people at Denali, you would not want to compete with those people. Like, they will kick your ass. Absolutely. I don't care who you are in the world, whatever big pharma thing you have, you know, they will kick your ass. There is unbelievable resonance between not just one or two A pluses. Like, you go around the table at the board meeting, everybody there is just like, you know, 3 standard deviations exceptional. And when you get them all together, then you're 5 standard deviations exceptional. You know, there isn't anything like it. And then, you know, it doesn't take long for people to recognize that, either people with money or even pharmas, and they're kind of like, Oh, okay, can we give you our stuff? You know, or how can we collaborate with you? And that, then that is it, it becomes a self-fulfilling prophecy. You know, when you think about value or you're thinking about kind of innovation, right? So what you want to do is, you know, try to take risks in a smart and calculated way, but you also want to try to take risks. My biggest criticism of venture in general is that venture capitalists sometimes don't take enough risk. Really?

What a surprise for all of us.

So, you know, I mean, I think it's just one of these things where that's why they call it venture capital, right? You're supposed to lose money. I remember I was meeting with some famous LBO dude in New York, and he's a really nice guy, but he kind of starts this conversation with you by saying, you know, since 1981, he hadn't lost money in a deal. And I went to University of Chicago, so the first thing out of my mouth— and this guy's known to be kind of, you know, strong-willed— I said, oh, you must not be taking enough risk. And he was used to people saying, oh, isn't that just the greatest thing, right? And I told this vice president that I said that. He said, you said that to him? Yeah, because like, it's true.

You didn't use the F-word though.

Probably did, actually. It's possible.

So you sort of started on the Denali story. Why don't you just fill in some of the gaps? How did this all come together?

Well, Denali was a weird intersection of multiple people thinking the same thing. So I had been pestering Mark for a long time. That's Tessier Levine. Mark Tessier Levine. That it was, you know, we should do something in neuro, we should start a neuro company because the clinical problems were really big. And he had kind of said, you know, it's not time, you know, and I would call him up and tell him, let's do it, you know, and he would say, no, it's not the time, you know, and at one point I called him up and he said, okay, now's the time. Great, let's do it.

Was that 35 phone calls like, fucker?

Yeah, a few. And you know, some of them were scientific discussions on why, And then I have been having the same conversation with Doug Cole about let's do a neuro company someday. And then Fidelity had been thinking about and actually funding a little neuro company. And I kind of convinced everybody that we should pull all of our efforts together and fund it with kind of more money than any of them were comfortable with. And I was really taking the lesson on how to do that from Juno, which was to try to assemble— you know, we go through and basically try to create an aligned interest group of investors. And there's nothing that hurts startups more than misalignment, either in the board or the investment group, investor group. There's too many multiplicative probabilities that are working against you in the startup. You don't want to screw around with dissonance in the board. You don't want to screw around with investors that have different interests. It kills companies. And it's a manufactured risk versus an actual risk. And so what we decided, I kind of was interested in creating a long-term aligned group of investors that wasn't comprised of people who wanted to flip in and out. I don't see a lot of value, as much as I like some of the later-stage venture capital friends that I have, probably less recently after making speeches like this. [LAUGHTER] I don't see a huge need for or value in later-stage venture capital. It's just money, and if you're smart in the early stage, you can create a board and expertise that you really don't need more investors telling you what to do. So, and the same thing with some of these guys that flip in and out of companies. So the idea behind Gino was that we would pick people who wanted to hold it for 10 years, and that was their timeframe. So actually when we closed the B round of Gino, I think the average holding period for the investors we got was like 11.3 years. 3 years or something. It was crazy. And it turns out that that has no impact. It actually increases the probability of an earlier financial return. So if you think long-term, it's just like trying to sell your company, right? If you're trying to sell your company, you're gonna be screwed. If you're not trying to sell your company but you're just kicking ass and running your company, then it gets bought, right? And that the probability of getting bought is higher if you if you actually just do well. And, but the minute you, so it's the same kind of thing, is if you think long-term, good things happen and your optionality increases. And, you know, the way to think of startups is linked option theory, right? It's not NPV. It's all about increasing your optionality. The only thing you need to know, I think, about thinking about startups is you wanna just maximize your options at all times. And you want to be able to pick options, right? You have to be able to pick a plan, but you're still maximizing optionality within any plan that you pick, whether it's financial optionality or any other way you define optionality. You just don't want to get down the end of the maze and there's— you're stuck. Because, you know, then you really are kind of screwed. So, so, um— Do you model this?

Like, do they teach you it? University of Chicago?

No, I mean, I once used to know option theory because I actually worked on the OEX. Actually, I forgot that chapter.

Tell them what that is.

The Chicago Board Options Exchange. So I actually didn't want to be a runner because I wanted to kind of trade in my own account. So when I was 17, I got this job because I was on a chairlift with some dude who owned a seat, and I said, can I come work for you that summer? By the time I was off the chairlift, I had a summer job. And so I went there and started trading options in his account and my account. I have no idea why he let me trade options in his account. And so I guess maybe I was 19, I don't know. But anyway, you know, I learned a little bit about the theory and I think that MPBs, you know, one of the reasons pharmas are so bad at this decision-making is because they use NPV analysis on, you know, what should be—

You better tell them what that is.

Options. Well, I mean, just, you know—

Net present value.

I mean, they don't—

A lot of scientists in the room.

They don't appreciate the uncertainty, the multiplicative probability uncertainties involved in, like, making a drug. And so, it's a bit— but eventually it gets to some financial guy who has, you know, who comes up with a number and whenever you have a system that's very complicated, you know, you tend to want to believe the analysis. You know, whether you're talking— it doesn't matter which model you're talking about. If it's complicated and you can see all these lines, And it comes out with a number, you're like, oh, that number is better than this number, even though it's just all bullshit. And so, you know, you can pick drugs all day long that where the market analysis, you know, whether it's Gleevec or any other, you know, just pick drugs all day long, market analysis always wrong. And sometimes wildly wrong on the blockbusters. And so that's where you kind of have to I kind of look at the clinical, the size of the clinical problem. So I'm doing a big infectious disease company right now. It's the biggest one ever done. People are like, you're crazy to do infectious disease because you can't make money in infectious disease. I was like, you mean, you know, those big fancy mansions in Hillsborough are not owned by Gilead executives? I don't know. I mean, somebody's making money somewhere. But I think if you cure things, if you either prevent things and can capture the value, which is hard, or you cure or disease-modify things, you'll make money. So I don't model that out. I just need to know that it's a big clinical problem, and that's enough for me. I think you could get lost in the weeds. That doesn't mean there's not a lot of nuance in how you might do it. So one of the things that we think about is like, hey, Can you make money giving people prophylactic antibodies that treat disease or maybe prevent disease? And that's complicated, right? That could be, maybe it's a cocktail of them. 5 years ago, you wouldn't think you could do that profitably, but now antibody costs are going like this. Maybe you could make a cocktail of antibodies that you could, you know, there's ways to do that. You have to think about the business, but you don't want to let that overwhelm the kind of sign. What we always used to say at University of Chicago, which was kind of, we'd make fun of the Harvard people actually, was the Harvard people would have this really complicated model that would show like that if you increase taxes, unemployment went down. And the Chicago people were like, huh. I was like, that doesn't make any sense. It's like everybody we know, like, you increase the tax a lot, they don't go out and hire more people. And that, you know, it turns out if you throw like growth into the equation or something like that, it reverses the sign. But I had all kinds of arguments with people. So sometimes it's just better to like pick really big problems and then focus on what's the probability of the technology actually being able to solve the problem. And that is actually tricky.

So you sort of alluded to BEER. You want to tell us a little bit about it? George Skangos, famous CEO from Biogen, just joined as CEO here.

So BEER kind of arose from— it's really kind of simple. I just decided that if you look at the world of of kind of infectious disease, there are lots and lots of unsolved problems. And I kind of looked around at like, who is the competition? And I thought, huh, you know, I wonder if all this wave of immunology, et cetera, that is being applied to cancer can be, you know, applied to infectious disease. And it turns out a lot of the cancer Stuff was actually invented for infectious disease, it just hasn't been applied to infectious disease. And so it's kind of a simple idea that maybe immunology meets infectious disease from host perspective would be a cool company to start. And then one model it could be Genentech and the other model it could be Celgene or Gilead or maybe it's a combination of all, some aspects of all three of those. Business models. It's R&D-focused, science-driven, and so I just decided to start it. And I decided to call it VIR because that was like virus. And then luckily I had already recruited a female chairman before somebody told me, you know, VIR means male. And I was like, oh. But anyway, so we have a chairman named Vicki Sato who is a female which gives me some cover calling my company mail. But it also fits on a license plate, which is the most important thing you should remember about naming any startup. It has to fit on a license plate.

Which has to be on a Tesla.

Yeah, it doesn't matter what it is, just don't put it on a Prius. Unless you know how to drive. You can put it on it. Anyway, so Vyr, I just decided to start it. One of the luxuries of having a few successes is that I can kind of do stuff now to some degree, even if people think it's crazy. So most of the stuff I do for like 1 or 2 years, people think it's pretty nutty. Like an Alzheimer's company, people thought that was actually quite nutty for a while. It's catching on now. And then, so, and that one people definitely thought it was, and still do kind of think it's a little nutty. Tell us the concept. The concept was just immunology meets infectious disease. Science is at the right, there's a wave of science hitting in understanding the host, and that hasn't been applied very well in all the nasty infectious diseases that you can think of whether it's resistant TB or HIV or herpes or hepatitis B or just go down the list, there's just a lot of stuff. And then not even saying anything about the emerging viruses and stuff. And then there's this computational component that we are in the dark ages, right? You go into a hospital and they're like, oh, let's try this, let's try this, let's try this, let's try this. Oh, we have no idea what we're treating in terms of anti-infectives. Or even viruses. They didn't know whether it's a virus or, you know, they can, you know, oh, you have something, right? Let's guess what it is and then we'll try this. And then so, you know, the idea is to have a more quantitative approach to profiling what it is and that informs drug discovery. We also have a computational immunology component which is is actually direct. So we're actually using computational data to think about new therapeutics in an entirely different way than it has been done. And then we're still collecting technology. So one of the things I had a thought about, which I didn't really know, which is, you know, it's a little bit of if you build it, we will come. If we build this big thing with lots of money with really good people, that we'll be able to go get cool stuff that's out there because there probably is some cool stuff out there. I'd say the stuff out there is cooler than we thought. Most of it's in the bowels of biotech companies and some pharmas, and even there it's underfunded. And so it is accessible, which is the surprise. And it's also rare that when you go into the pharma time and say, you know, I'm willing to give you $50 million, will you give me your stuff? They're like, you're going to pay us? And, you know, and more interestingly, I'm willing to take your R&D budget off your, you know, your P&L. And then they're kind of, that gets their interest. And so we're able to do stuff at a bit of a different scale than has ever been done really in the space. Space because of the resources that we have.

Who's investing with you? Or do you have a normal syndicate you like to pull in?

It tends to be sovereign wealth funds and big public funds, and then high net worth, very, very, very high net worth people. So with Veer, we haven't really talked about who the other investors are yet. But we have talked about the Gates Foundation. So they actually invested in the company. And then are funding some of the programs.

So it's quite unusual that you've got the Alaska Fund involved with Denali anyway, and Juneau, I believe. So how did that happen? So first tell them about the fund, which used to invest in bonds or something 30 years ago.

Yeah, I mean, there's a bunch of— it is a symptom of something that's going on in the world, are— there's a disintermediation of intermediaries. So fund of funds and other folks are being disintermediated by big pools of money going direct. And they don't go direct, you know, willy-nilly. They go direct with very few relationships. So they have somebody kind of guide them in their efforts, whether it's CIC in China or or the big Middle East sovereign wealth funds or the US sovereign wealth funds like Alaska Permanent Fund, which is I think the only, or one of the biggest sovereign wealth funds here. And then you have Temasek in Singapore, GIC. And then the same thing is happening with big public investors. So some of the investors that used to be very hard to get data, and I was very active in kind of pulling some of these folks in early, 4 or 5 years ago, into Kythera as an example. These funds that manage a trillion dollars or 3, you know, kind of anywhere from a couple hundred billion to a couple trillion are, you know, they want to own 15% of you in 5 years or 10 years into hold. And that's a really good investor. So I want somebody that buys in the A, the B, the IPO, the secondary, and then in the market. And that's the definition of a long-term investor, and that's what you really want. And it turns out those two things are coinciding. The bigger money wants interesting things to do. They're not interested in putting $5 million in something CIC, I think their entry point is $200 million. Maybe they'll go smaller to $100 million, but it's big money. Most of these folks don't want to write a check less than $25 million. So what you can't do is rerun the error that venture capitalists made in the '90s and first decade of the 2000s, which was if you just add up all the money and then put it into one big round, and don't change any other systematic, if you don't shift the curve of probability set of the company, i.e., better management, better opportunity set, better technology, you've got nothing. You end up with a 1x. And a lot of VCs got that. I mean, I had deals where we successfully prosecuted the company to an IPO and had a 1x. Which is generally not good. It's better than a less than one, actually. I have some though. So if you're going to do the kind of the big money thing, you have to do it in a way that fundamentally shifts the opportunity set and therefore the value kind of sets. So the option, if you want to get quantitative about it, you could probably run a model of this series of linked options and say, you know, we've changed the probabilities of this. And you could say you're changing it because you have better and different human capital. In some sense, you know, one of the things I'm proudest of Denali is they have a really good kill algorithm, right? They know how to kill projects. So the very first platform that the company brought in, they killed. 6 months later. It wasn't, it was the entire platform. I mean, it's killed kind of down, I don't know if it's, you know, somewhere, you know, in the negligible burn rate, but, you know, watching the science still, because science is cool. But, and very few companies have the ability to do that. So if you recruit management teams like George Stambros or Ryan, then you can deal with a portfolio of assets where normally Historically, biotech companies haven't been very good, and pharma companies are definitely not very good at killing stuff, right? The whole culture of a pharma is like to preserve the project, 'cause that's your job. You don't see people getting promoted for like killing their clinical trial in phase 2. You should. I mean, Ageos has a party every time they kill a DC. And give out bonuses. So they all get drunk, they all kind of celebrate the effort, and that's very rare but very cool because that's what should happen, right? You should be kind of rewarding people for being disciplined and admitting that the data sucks. And, you know, but pharma 100% does not do that, and a lot of biotechs, like, if it's your lead product and you got one product, you know, how many people want to be the one to go into the sea CEO and say, you know, your baby's ugly. And so, you know, we've tried to do some of these companies like there's multiple, you know, there are multiple assets. So you aren't dependent on a single event. And one of the things that these big pools of wealth fear and don't, never liked about biotechs, Even in public companies, they buy a public company, some shit happens and it goes down 80% the next day. That's not good if you're like a sovereign wealth fund, you just bought 10% of the company. And so, they want some kind of way of balancing risk, but not in a way that's just putting a whole bunch of shitty assets together in one place, which is what people have done historically. So you have to do it differently and you have to do it in a way that either has some scientific insight or some different way of looking at the problem or some better way of killing something.

So I'm going to open this up for questions in just a couple of minutes, but I really want to know about Unity Biotech and your anti-aging approach and your personal N of 1. Do you want to talk about, start with that?

Self-experimentation. So, you know, I think if you're going to be in the pharmaceutical business, you you should be willing to take pharmaceuticals. So including prophylactically, as long as you run your renal and liver function often. So I do run my renal and liver function often, but I do think that there, you know, I tend to be a little bit ahead of the, I'm willing to extrapolate from mouse data to humans fairly rapidly. 'Cause I, you know, I actually fear death. It's like the Pirates of the Caribbean, you know, I actually fear death. And so I want to live longer. And I would prefer to live longer in a healthier state. And so we started one company, it was a, you know, wow, unsuccessful one. And we lost, I don't know, $15 or $20 million 15 years ago, which is a lot of money. In that context. And then decided to do it again with better data with Unity, a company founded by Ned David, which is based on deleting senescent cells. And out of the Gladstone and Judy Campisi's work and Mayo. So some really interesting data. It's kind of hypothesis-driven, which is a little scary, but then And now the data is getting better and better and better in terms of the real biology underlying it. So when we started, it was kind of truly a hypothesis-driven, it's not usually what we invest in, but it was such an interesting result that we said, okay, we'll give you a little bit of money. So I think we invested $7 million over 5 years. And just getting better data, getting better data, getting better data until it appeared that we had, you know, the ability to drug, you know, to have a small molecule that would delete senescent cells. And then we put in RealMed, which is, we just raised, I don't know, well north of $100 million to make drugs that first start locally. So they're gonna start in, Areas like the knee or the eye or the lung or something. And then eventually, you know, and it's not just limited to senescence. So then we broaden the mission. So we're looking at other science other than senescence. But, you know, the idea is to have drugs which increase healthspan, right? Which can prolong— it's always fascinating to me if you just look at kind of the data about risk for disease, right? So pretty much anybody in this room, anything you do, whether it's your genetics or your behaviors, are still less related to disease than your age, which is pretty wild if you think about it. You can have all these genetic risk factors, you can have, you know, you could smoke, you can drink, you can do all this stuff, but, you know, the probability of you getting one of these big nasty diseases diseases is correlated more with your age than any of those other things, unless you're, you know, kind of a 55-year-old, you know, ApoE4/4 unlucky person. I mean, I guess you're unlucky if you're an ApoE4/4 anyway. So I think that's the kind of thing that motivates me. I do take, you know, metformin prophylactically and lipitor prophylactically and nicotinamide riboside because, you know, I don't know, it seems like the thing. And then if you're going to do that kind of stuff, you should definitely get your renal function and liver function checked. But I don't think it's any dopier than a vegetarian diet, right? I think that's much more dangerous than to then taking metformin. I would trade my high-fat, high-protein diet, but you couldn't get me to trade it for a vegetarian or vegan diet.

I hope you eat also. You're not just taking pills.

No, no. Okay, good. I eat. I try to eat. I'm not ketogenic, 'cause I have not enough discipline to be ketogenic. I like wine and ice cream. But I do try to have a very high-fat diet, basically.

OK, I'm sure there are lots of questions here. So who wants to start off?

Go for it. Great talk. I was really interested by a comment you made early on about the concept of trying to swing for the fences by hiring a really big-name team. Put in a lot of money and then good things are going to happen. You also noted that typically pharma is really bad at killing projects and has all these kind of bad behaviors that have otherwise not necessarily always led to good outcomes. So can you talk a little bit about how you square those two things with each other? How do you find— But I don't think— so I think big name team is the wrong way to think about it. I mean, I think Like with Vir, we happened to hire a big name guy because he's really good. But it wouldn't have mattered to me whether he was coming from Biogen or some other place. So it's not about them. It's about the— the way I would say it is we hire DNA. So I want great DNA. And I don't even care if they're a cancer person going after infectious Is it a chronic disease or an infectious disease? You don't do box-checking. It's kind of like getting married, right? It's like if you try to do the whole box-checking exercise, it's going to be— I was absolutely convinced in my life that I was going to marry an MD-PhD Nobel Prize-winning economist, and I married my hairdresser. You never know. So I think you gotta go with the gut on the hiring part. I think pharma is a fundamentally broken incentive structure. It is as broken as it can possibly be. So you're the CEO of a pharma, you're lauded in the press and get to go on CNBC for some decision that the guy that got fired before you made 10 years ago. Right? There's no correlation at all with compensation and merit, whether you're in the R&D organization or whether you're in business management. The only people in pharma that are truly like, you know, have incentives that are aligned are the salespeople. People, right? Which is they're supposed to sell more stuff. And now that is controversial because you can't really, that creates incentives that are possibly bad, which is to sell stuff off-label. So like in Ikaria, we had a non-salesforce, I mean a non-sales-based compensation structure for our salespeople. But so I just think the incentive structure is broken. It's not a meritocracy. And there's a lot of really great people in pharma that we like to hire sometimes when they get pissed off. But it's a super frustrating thing because most of them know that there's a lot of this stuff that just is perpetuated and it's just impossible to kill. And nobody, you know, why, if you're the head of that program, you know, would would you want to kill essentially your job? In Microsoft, if you kill your job, you can get promoted. Especially old Microsoft. Like the first 20 years of Microsoft, that was what was so cool about it. You could be somebody with a pierced nose and a tattoo on the middle of your forehead and like lose $600 million and they're like, yeah, we thought that was the right decision. Decision and then get promoted. And Google and some of the places in the Valley, one of the cool things about those places is they recognize talent really well and they continue to give those people a shot, not indefinitely, but eventually.

Okay, down on the floor.

So something that still puzzles me, you seem to take a decidedly non-scientific approach to deciding on the potential for science. Now, based on that model, it seems like a lot of your investments would be Theranos-type investments. Yeah, we turned down Theranos pretty early. So we have a pretty good network of scientists, right, that we run supply. So we don't act in a vacuum. I'm acting in a vacuum when I'm deciding to do VIR, but when I'm making the science nest of VIR, it's Vicki Sato, who was CSO of Biogen Idec, president of Vertex, Jeff Bluestone from here, Tom Daniel, president of Celgene, Mark Davis from Stanford, Emilio Amini from the Gates Foundation, he used to run infectious disease at Merck, Larry Corey who was head of IAVI and head of the Hutch and, you know, good merger of infectious disease. So, kind of, and then Phil Sharp, who, you know, I called Phil and George called Phil and it's like, you know, why do you want me to do infectious disease? Infectious disease company and it's kind of like, you're really smart. Part of the problem with infectious disease, we need some orthogonal views to come in to challenge the existing dogmas, which I believe is always important in these things. So you can't just, you need to mix it up a little bit to get creative. So I mean, I probably have 20 Nobel Prize winners' cell phones. I don't pay any of them ever consulting fees. So sometimes I pay them in wine if they bitch at me long enough. But basically, and the reason we do that intentionally, we don't have a paid scientific advisory board, we do that intentionally so the bar is really high. If you call a Nobel Prize winner and say, I want you to do this for free, and they're used to getting $5,000 an hour or something, right? They have to think that it's cool science. And so it's an opt-in kind of thing. And it's a way of testing out, you know, how crazy is the idea. But sometimes it's just, you just need to have orthogonal other science, mix up the science because people are, kind of hanging out with each other too much. Follow-up on it, what about the valuation? You're sinking a certain amount of money, you must have some scientific calculation to decide what the value of the— No, it's just all, well, I can attest to this. It's, we just make it up as we go. And it's, valuations are actually not the issue. It's option theory. Again, that's the issue. How much do I own? How big is it going to be? That's all you need to know. And what's the probability of that? And so, what you don't want to do is miss the ones that are big. So, I'm actually willing to pay up in later— in deals that I think are going to be really big that I missed. Because there are only going to be, let's call it, I don't know, Let's say there's 5 deals that matter in the world per year, which is probably about right. Matter, I mean, that are going to be big and transforming. And maybe it's 1, I don't know, but it's not very many. And you need to kind of be in some of those or else you don't make excess returns. And so you can, you know, that's from the financial side. From the transforming side, you know, just following the clinical problems. And so about half of our stuff is invention-based, right? And then we need to check, like, is it gonna work? So somebody came and pitched me a deal. It was started by another venture capitalist who's really good. They came in and said, hey, I've got this gene editing thing. They could do in vivo, high fidelity gene editing, not CRISPR-Cas today. I was like, okay. I'll invest. If it works, I'll invest. 5-minute investment decision. I said that 5 minutes into the conversation. And it had been validated in 2 labs outside of that lab.

Do you have partners you have to take things in front of?

Yeah. Then I called Steve Gillis and said, can you make sure this fucking works? He's way smarter than me, and then he went in and like made sure that it worked, right? But there's a certain amount of, you know, it has to be good science, right? So I would say interestingly, the thing that we fail the least at relative even to other firms is picking the good science. We're probably 19 out of 20, 18 out of 20 in terms of picking science that works. Even earlier.

Is that because you surround yourself with the kinds of people that you just described?

It's a combination of picking the right platforms, so, and yeah, and picking the right people. So a lot of times you're betting on the scientists because the individual data that you might not— that you may be betting on a certain set of data that they have, but in reality it's a flow over 4 4 or 5 years that's gonna, and it may be year 3 that they come up with the really cool thing. That's the way, and so the company I'm creating right now is one of those where it was kind of really cool, really cool, we backed him in a deal that failed, then we backed him in a deal that succeeded, and then he came up with this new thing, and you know, it was a little bit of, you know, enough data to go, wow, that's really, really cool, cool, and if you could do these two other things, you know, then it would be super cool, and can you do that? And, you know, it's a bet. I don't— it could be wrong. So we're willing to take that risk. I think a lot of other people were not willing to take it, and we'll see if we're right or not. But it's— so there's real scientific risk in some of these ones that we bet on. It's a bet.

How important are the relationships with the individuals versus just the raw science? Would you take raw science from people you don't know at all and figure out you can make that work?

I think it's a trust but verify thing. So the one— the other gene editing thing was in the person that wasn't really well known, and so the other VC that was involved in it had it replicated in two labs before they wrote the check. I think there's a certain amount of, you know, if you're a young graduate student, it kind of depends on pedigree, depends on who's, and you know, we'll definitely look at all the way down into lab notebooks and making sure that it's real. We have backed a couple things where, you know, it wasn't real. Like, it was not real. And that was interesting. We were super nice about it. We didn't like insist on, you know, flailing them in public, but, you know, there were just, you know, retractions happened and off we went. Wow.

Okay, over there.

I see a lot more sort of traditional tech VC getting involved in biotech. They seem to develop an appetite for it. Do you think that's fundamentally misguided on their part and also maybe from the entrepreneur standpoint to take their money? And if they're doing something right or wrong, what do you feel they're doing right or wrong? Well, I think it's always a question of alternatives, right? I mean, if somebody's going to give you money and nobody else will give you money, then I would take that money. So, but, you know, I think it depends what the deal is. I think Basically what happens, you know, there's two things that are driving that. One is there are a lot of advances in data. And, you know, if you say the word, you know, kind of machine learning, big data, and healthcare, you can get funded today. If you say them in the right order, you know, near the right door. You have to whisper it into the little box. And then, So I don't, and the other thing is the aging of rich tech people, right? So, and aging is a relative thing. Could be 28-year-olds turning 32 and realizing that one of their relatives got sick or that they just want to live a long time. Or it could be, you know, 50-year-old, you know, deciding that, you know, they aren't going to live forever. So I think it's all of those things converging. I think it's a good thing generally. And we've reached out to and have co-invested with some of those folks. I think they have something to add to some deals. Some things they're just observers. They're just learning about, like, gee, what's the FDA? How come you can't do this? And all that kind of stuff. But I kind of like the thing I like about outside of pure therapeutics where they definitely have less to add, you know, in some of the healthcare things, you know, they ask really hard questions like, you know, what a fucked up system we have today. Why do we have that system? We're all kind of like looking around like, I don't know, go try to talk to your doctor about preventive health. They just look at you like, what the hell are you talking about? Just lose weight, your blood pressure should be lower, see you when you're sick. And so I like the tech VCs kind of asking the very difficult questions about the system and why things aren't.

Do you touch that digital health arena?

Yeah, so like with GRAIL and and other things. So we're quite interested. We have a company called Alidate that's in the kind of ACO. But we're not as much like Ben Roth or somebody who would kind of go really long in some of those things. Although lately we've been thinking about kind of how screwed up the system is and whether there's ways of making it better. This lady. I'm curious to know what you think What does he think about your exit strategies, timing, IPO, conditions?

I'm going to repeat that because I'm not sure the back heard it. What does he think about exit strategies, timing, IPO, conditions?

So I think it's back to the optionality thing. So what I want to do is be able to control that completely inside. So, and you know, that's hard to do. I mean, very few companies can do that. So I want to walk into the investment bankers and tell them, we're going public, here's our range, and here's our book, and do you guys want to do it? And we have done that. Works. It's difficult, you know, so you have to have a pretty strong internal syndicate to be able to come up with that kind of money. In advance, but all of the big companies that we have, I want the ability to do that inside. I think IPOs are fine as long as you have your long-term shareholder base, right? So if you have the shareholder base we have in our private companies and they maintain the public shareholder base, it's gonna be really stable and long-term thinking. Doesn't mean you're not gonna be subject to, you know, shorts coming in and, you know, if you have a clinical failure, your stock's gonna go down 50%. But it doesn't matter. The long-term kind of is baked in there. So we don't mind selling companies as long as it's for a lot of money, right? So we're not in the business to sell companies for $300 or $400 million and really declare victory. I mean, we do do that. It's not necessarily a bad thing. But it's not our goal. Our goal is to hit it out of the park on every single deal. We don't try to balance the portfolio, we don't try to kind of do some things that are less risky. We just invest in cool stuff and just keep investing in cool stuff. That's really the algorithm.

How do you know when's the right time?

In terms of to go out? I don't think anybody knows that. So anybody tells you they know that, it's bullshit, including investment bankers or anybody else, right? They don't know. I mean, there's, I had this one deal on the road once and so we're kind of almost ready to go public, like on the road show in a week or two and they're like, okay, you know, window's gonna be open for another couple months, getting a little bit choppy out there. And then we have these calls every few days, right, getting ready. And so we have another call later and they're like, oh, you know, it's really getting choppy out there. The window may close in, you know, just a week or 10 days. We better get on the road. 2 or 3 days later, market goes down another 6, 7%. And they're like, oh, you know, the window's actually closed and it closed 2 weeks ago. So who knew? Right. So investment bankers are very good at reading the data after it's happened. And so I think, you know, you need to have your— they don't raise money for you, you raise the money. That's the thing. There's this idea that somehow investment bankers go out and raise the money for you. They are facilitators of you getting to the right people to make your pitch, but you make the pitch and you raise the money. And that doesn't mean that there are no differences in investment bankers. There are in terms of their network and access and positioning and how they think about it and how long-term they're thinking. I don't know what the hell that is. It's not an earthquake or protesters outside. But any case, I think you have to be careful because when the— It's the musical chairs game, right? When the party stops, you better have a plan because everybody else is gonna be running for the chairs. And so there's like, today, if there was a war in North Korea or something, or pick your event, and the money dried up, there's more companies chasing the money than there is money. And somebody loses. So you don't want to be that, you know, you always need to be contingently planning for that. At the same time, you have to play offense. So you can't just play defense. And there are some of these companies who are just like, you know, so defensive that they actually never do the experiments to generate the data to create the value, right? They're just kind of almost hoarding their cash. And so, and they end up living with the probability of one product. And, you know, the good thing about one product that works is you get Receptos and you sell it for $7.8 billion. And the bad thing about one product is, you know, sometimes it doesn't work. We have plenty of examples where, you know, it goes to zero. And if you don't have a portfolio or something. So a bunch of— wow.

Okay, I'm going to give everybody numbers. So 1, 2, 3, 4. Okay, you remember your number?

Go. 1. So you mentioned MPV a few times. Are you— would you say you're using it for IRR than other investors that are on Sand Hill? And then also kind of perspective of the vintage of your funds and the amount of time that you've been putting in your portcos? Is that different or similar in relation to your investing style? And on top of that, taking on crossovers and investors, especially early on? Um, I don't think we— we don't really think in IRR. We think in cash on cash. Okay. And we don't actually care about IRR at all. Some of our limited partners don't like it when we say that because their bonuses are based on IRR. Um, some of them. But we were just cash on slinger. So when I think of like a good fund, I want to make a 3x or better cash on cash. And, and so, and I think to do that, you know, this is all the modeling I need, right? I just think, okay, you know, some of the companies need to make a 10x to 20x, and if I have a few of those, you know, then, and I have a reasonable amount of money in those, then I'm probably gonna do okay. And I can take, you know, 20% losses. And so, you know, people are like, oh, what's the percentage? I don't know. Probably 5 out of 10 succeed reasonably well, 2 crater, and, you know, 1 or, you know, of the 5, maybe 1 or 2 hit it out. And, but some funds, I mean, we had one fund where we returned 35 cents on the dollar. It sucks. That's it. Matter of fact, I had this one meeting while I was fundraising, 'cause I started with that, and I said, you know, here's, I start fundraising by going through our errors. And so, I kinda said, you know, here's what we fucked up. Here are the 3 things we've screwed up over like 20, 30 years. And they said, yeah, that's one of the worst, like your period in like '96 to 2003 is one of the worst records we've seen in venture capital. I was like, yep, it sure is. And you know, if it was a non-systematic error, that would really suck. But we know what we did and we corrected it. And so that's what you— and we did a bunch of internet investing, turns out. So it turns out, like, if you want an algorithm to lose money in the internet, you take our algorithm, which was it's technology-based, so it has to be a fundamental platform technology, and it's milestone-based, so you don't want to put too— you're kind kind of eyedroppering the money in. And you don't know anything about markets, which, you know, we didn't. So, if you put all those things together, we actually lost money in something like, in '99, you know, '99 to 2001, something like 17 out of 19 deals. In an era where you had to be an idiot not to make money on the internet. I mean, you know, you you could just randomly back some 25-year-old, and there are partners at Sand Hill Road firms who kind of randomly back some 25-year-old and made their career. And that was the only deal, you know, I mean, they got one deal. You can make your own assumptions about what that means. But, you know, we systematically chose an algorithm, which I think if you applied that algorithm today, it would also not work. And so, you know, with maybe a few exceptions. And then we're really good at semiconductor deals. It turns out you just can't make money in semiconductor deals. So it took us about 20 years to figure out, hey, you know, we're pretty good at this, but like the outcome just sucks. So let's not do that anymore. And so that's how we ended up most of the pipeline.

All right, number 2.

As a VC, do you believe that The valley of death is real where really promising discoveries happen in the university lab and they don't advance because people like you consider it too risky or the pharma companies consider it too risky? Or is it just a matter of scientists thinking their research is better than it really is? I mean, it's a combination of all of those things, right? So I do think that there's very interesting research at universities that does not get kind of that $50,000 to $250,000. And I actually think that's more of a university institutional problem than it is a venture capital problem. You can take that money from angels or other, you know, you can get it from philanthropy. I think venture philanthropy is a good way to get that kind of money to be able to, and a lot of times the academics don't know what question to answer. So we go through universities all the time just telling them like, here's the question you really need to answer, here's the value inflection point. That's incredibly valuable because they don't know what it is usually, and then once you know what it is, you can decide, okay, then I have to just change what I'm doing, or this data is a lot more important than some other piece of data to get outside venture money. But the universities need to have these pools of kind of, I think, discretionary kind of slush fund money. Maybe it's controlled by a few people. Not big committees, not peer review. I think one of the biggest problems, and I think at the NIH level that also needs to change. So I think there needs to be a section of the NIH that is, less peer-reviewed. So you don't already have to cook the data before you go in. Not cook the data, but I mean, you know, experiment it. So basically, you know, it's more like DARPA where there's, you know, maybe it's up to $1 million can be just done at the discretion of somebody who isn't going to be at NIH for life. And so that you get some of these young investigators or even old investigators with a crazy idea. I mean, I've had some, like, just across the spectrum of kind of age and experience, people come with some orthogonal idea that very much is worthy of being funded, that in the case of Icaria got funded by the director. So I carry it— actually, I asked Lee Hartwell a question which I knew he would hate and think was stupid, which is, who's the smartest person at the Hutch? And like, that's an impossible stupid question to answer, but I was trying to get a list, right? So he said, well, that's a stupid question. You know, you should go talk to Mark Roth. And I was like, okay, great, we have to talk to Mark. And Mark was doing suspended animation of zebrafish by using rotenone to kill mitochondria, basically. You can suspend the zebrafish indefinitely, basically just stop them. And that's pretty cool. So we, you know, so Lee gave him some money I think it was not very much, $100K or something. Because I said, you know, what you really need to do is do that in a mouse. Like zebrafish, you can do all kinds of weird shit to zebrafish. You know, freeze them in your freezer and unthaw them and they live and stuff. But, you know, mouse, like, that's harder. And so a couple years later, I mean, we kept in touch. A couple years later, he's like, okay, you guys, So he figured out using hydrogen sulfide that you could suspend a mouse. Like, you can make an animal a poikotherm. You give them this stuff and you just watch them. They slow down, their temperature drops to 15°C, they have one respiration a minute, heartbeat goes to 10, and they sit there indefinitely. And then you turn them back on, they run around, and as Mark says, do what mice do. And so that's pretty cool. So we decided to fund Icaria based on that, having no idea what the therapeutic outcome would be, like what the hell we would do. Although we did know that if you take a Montana brown squirrel and you hibernate it, and then, you know, somebody did this experiment, which is not a very good one, this So let's just say that you, you know, while the ground squirrel is hibernating, you put a needle in its brain and then take it out. And, you know, when they're hibernating, you know, they're functional. So there's a lot less damage to a Montana ground squirrel while they're hibernating. So the question is, hey, you know, would that, you know, if you could induce this state, you know, would it be good for surgery or other things. And it turns out that the benefit of inducing the state can be done without hibernation. And that was cool. Like, when we figured that out, it was like, okay, we have a drug company because we could induce benefit, you know, without actually putting you to sleep. And so, and there was all kinds of cool things you know, I mean, if you do it in a hypovolemic shock model, you can get, you know, survival in models where you have 100% death rate in the control group and, you know, 80% of these mice where, you know, you, you basically— it's kind of a nasty experiment, I won't describe it, but let's say, you know, normally it's, it's, it's called the— it's a battlefield wound simulation. And, um, and, and so, you know, you could have all kinds of, you interesting therapeutic effects. And so then we put more money in the company, and that was a good thing. I bet you were the exception though. There wasn't a lot of VCs chasing after— No, no one would do this deal. No one. Everybody thought it was crazy. And then I pitched to Bryan Roberts after we had seeded it. And I think Bryan did it because he thought it was cool. And, you know, so, but I think most people would not have touched that with a, yeah, a wand.

Okay, number 3, where's number 3?

So since most startups here do not start with ArchVentures, with multiple Nobel laureates and $10 million, but rather a couple graduate students and a professor and $750,000, could you talk about the transition of the skill sets that are needed in that first phase and then as you move on. I think that figuring out what the value inflection milestone is is super important. What question are you trying to answer? And is it a question that anybody gives a shit about? That's what Stephanie's classes do. So, Z-Startup and things like that. You define that. And often those are— Those are both scientific questions, but they're also kind of related to what other people have seen, right? Or what they're skeptical about. So, you know, if you're playing in the, you know, if you think you have an in vivo gene editing technology, you better have some, you know, something that shows that it really actually works in vivo or figure out the experiment experiment in vitro that's going to convince you that, you know, that data could be real and it's a big enough thing. So I think figuring out— and that would be good to get other people's advice who are not you, right? So whether it's somebody in kind of that's done it before or a business development person in pharma or somebody who has some perspective on Gee, you know, we've seen 20 of those, you know, I don't believe it, but if you could do this, you know, then, you know, I had a guy once, we started a company with a woman named Angie Belcher who can make all kinds of cool things out of phage, and they, it turns out it's good for cancer. Catalysts. Carl knows this company. Um, and, uh, and so we did combinatorial catalyst discovery with it. We said, oh, what's the biggest problem in catalysis? Because when— and, and so it's the oxidative coupling of methane making C1 to C2. And if you make C2, then you make C4. And so, so, uh, oh, okay, you know, we don't know anything about catalysis, but I called my friend who happened to be running Linde at the time, which is like the largest gas company, and he said, "Oh, that's interesting. We've had 50 German engineers working on that problem for 20 years, and there's no fucking way you're gonna solve that, and you're out of your mind. So don't do that." So we did it anyway, and 2 years later, Linde invested in the company because at least we had like the— He's like, "And here's the data." I said, "Just humor me. What would you need to see?" You know, and so I think that's the kind of thing where you need to figure out exactly the minimum amount of dollars that you can put to a problem to create the data that is not necessarily going to be publication data, it's not going to be FDA data. This drives scientists crazy, right? Or clinical people crazy. I just want to know if it works in humans. Like, for instance, if it's a clinical question. I just want to know if it works as an investor. It doesn't have to be a bombproof Phase 2 trial. I don't even have to— I don't care if you repeat the Phase 2 trial. I'll take it to Australia, put it in 10 patients, and see like, does it work? It doesn't even have to work to a p-value of, you know, whatever. I just want to know kind of directionally if— or another question is, does it not work? Like, are we wasting our time? So in a smaller lab sense, I think you just have to figure out like in the scale of the problem that you're trying to solve, you know, where are you relative to the rest of the world.

So will you fund that first $100K to find out if it works?

Like $50K in stuff, you know, but it just has to be an interesting problem. So, you know, if it's the Mark Roth problem, that still intrigues me. So we ended up selling that company for $2 billion because we bought some other gas technology that turned out to be valuable. The management team didn't know what to do with the existing technology. They kicked it back to us. We started a new company called Faraday doing the same thing and then discovered some other cooler stuff that we ended up working on. That company is like close to the clinic now. But so I think if it's an interesting scientific and high impact. I think the main thing is high impact. Now all companies don't have to go cure every disease, right? Some companies are just going to do some cool thing along the way and those things can get funded. We may or may not fund those. But there's a lot of people out there who are interested in funding things that they can sell for $300 million that only take, you know, $10 or $20 million or, you know, $30 million. And they'll do them. And it doesn't mean we wouldn't, it just means that our bar is high.

Number 4.

So I guess just to follow up on one of the things you mentioned before, like where do these deals come from for you that you've actually funded? And then I guess secondly, can you just talk about the importance of the relationship that you do have with these teams and these companies as they progress and grow as a company?

So the question is repeating. Where do the deals come from, and what about the relationship with the teams and the companies?

So the people, I would say Christina Barov is my partner here in San Francisco, and I were laughing the other day because we were in this presentation and we were saying, wow, when's the last time we actually were in a presentation together where a company was pitching with PowerPoint. We couldn't remember it. And that, I mean, could have been 10 years. So like, it's not like Sand Hill Road where there's a bunch of people coming through. So we're much more likely to have phone calls with scientists, and maybe they're walking through a deck of science, but it's not usually companies. Every once in a while it's companies companies. Like, we just, we took a meeting, you know, so I found out about a deal a week ago or 10 days ago, took a meeting, sent 2 guys out there, 2 days later committed $25 million, and total turnaround 6 days. But that's rare. So, but I do think the relationships, like, I talked to Juno guys 4 times a week, 3 or 4 times a week, public company. So there's a lot of relationship there and a lot of involvement from our side. It's not a passive— I'm surprised actually on the tech side it's totally different. In the tech world it's like you back a person and you let them run with it in spite of their flaws or whatever they might be, you know, imperfect and you kind of made your bet. In biotech, it's a lot more money. I was meeting with one of the bigger tech VCs and they're just like, man, you guys put serious money in these deals. And not just us, I mean all of biotech, it's a lot of money. So it's a different level. You can't back, you're usually not backing kind of a 22-year-old graduate student, although sometimes you will, and then you make it part of something bigger or, you know, help build it. So we're not averse to that, but a lot of times it may be it's 2 or 3 things coming together to be one thing. And the relationships matter because you're just having to— people have to trust that you're going to figure out a way to look out for their interests as well as your own interests. And I think that we're pretty good at that. I think venture capital in general is not particularly known for that, right? So we kind of figured out a long time ago that if other people didn't make money, then they're less likely to tell their friends to bring cool ideas to us. So universities as well. So we're not constantly trying to screw the university because then they tend not to bring us other things. And conversely, universities tend to bring us really cool stuff when they know about it because they know that we're not going to screw them. Whereas a lot of other venture capitalists really do spend a lot of time trying to minimize everybody else's pie. I'm kind of the view where I just look at my own pie. I want to make a lot of money. I don't care if other people make a lot of money. I want other people to make a lot of money. And so, but that's an education process with scientists. Scientists as, you know, I'm telling you, they're sure like it's a foreign thing, right? They just don't want to be taken advantage of and they want to be rewarded if it works. And more importantly, the thing I learned at University of Chicago very early on, which is you do not want their competitor to make money on their stuff. That really drives them nuts. So, so, so like they hated, you know, the only way I got people at University of Chicago to Literally, the chairman of immunology said that he thought patenting was unethical and that if his faculty spent money on patents, I mean, spent time on patents to the detriment of their scientific work, that should impact their tenure decision. Until he invented something about, like it was very shortly after that, like 6 months later, and he's like, I don't want a patent. I'm like, really? And I kind of forgot who it was, somebody at Harvard. I'm like, huh, that guy's working in the same area. I bet you he patents it. And then it immediately changed his entire worldview. He's like, well, he can't do that. I'm like, sure he can. He's going to copy your stuff. And then he's going to take credit for it. And then he's going to make money off your stuff. And it's going to drive you crazy. And then he changes his mind totally and says, oh, you know.

So what are the next hot areas that you're thinking about?

You know, we're very active kind of in gene editing-ish of every flavor because it's really cool. Although nobody's yet figured out how to make money out of it. We have a really cool company called eGenesis which is working in xenotransplantation. So they have You know, they're editing pigs to be able to do human, you know, to be able to transplant organs into humans. So they've already created the most gene-edited mammal on the planet. They also have the same team that's doing the woolly mammoth as well, so hopefully they'll do the woolly mammoth on this, you know, in their spare time. And then, so that's kind of cool. And we're very— I guess Infectious Disease is obviously kind of interested in alternative approaches to healthcare delivery and thinking about the whole system and how screwed up it is. So, you know, figuring out ways of using data to change, to make things more preventive and— or to discover things in a more rapid manner. Let's see what else. There are always just cool orthogonal things that come our way. So usually what we get is just somebody that knows somebody Sends us something, you know, and then it percolates through the— we just send it out amongst the group. We usually want non-confidential information unless it's, you know, somebody we know who says, you know, I have a secret, you know, you can't hear it. And then sometimes we'll sign an NDA, but it has to— we have to kind of know what it is we're signing an NDA around. And often with universities we do if it's something novel and orthogonal. And sometimes it's just something, you know, it's like, oh, I've got a new CAR T-cell idea. I was like, great, you know, you should talk to Juno. We, you know, introduce them and they, you know, have a— sometimes even if it's competing, we have a, you know, we've There are ways to have those conversations in a way that, you know, we're buying a company this week where, you know, the person was absolutely sure they wanted to do new code. They'd already taken $20 million. It was the best thing to do, blah, blah, blah, blah, blah. And I'm like, you know, no, it's not. Like, you should, you know, you should do it a different way and we can create a synthetic structure that gets you all the things you need, plus you have scientific control. And if you, you know, everything you do does not have to be a new co, right? So you both want to test the hypothesis of doing a new co and believe in what your idea. And you also can think about, you know, aligning with other people and not having to reinvent the wheel. And so there are kind of some novel structures now to be able to kind of simulate, you know, if you're sucked into the board, you don't necessarily have to be, you know, part of the board. So we've done things like, we've done a couple of these actually in VR where we're putting, we make a subsidiary. You know, so like you get your own playground and some money and off you go. But you're still part of a bigger entity. And so you create both the downside protection, which is you have stock in a more certain thing, and then some upside incentive that basically says, hey, if your stuff works, you know, you get an extra kicker in cash or stock, whatever you want. So that kind of is— there are ways to deal with that. But, you know, every professor, every graduate student, every wannabe graduate student has to start a company. And in the internet, I was riding on a plane with somebody who was a senior VP at Salesforce, and she's like, my daughter's starting her second company. I was like, oh really, how old is she? She's like, 12. I was like, you gotta be— actually, she's got a term sheet for her, an actual venture capital term sheet at 12 for her second company. And it's like, that's not going to happen in biotech. It's just like, you know, it's just unlikely. So the capital requirements are high. And if you're in diagnostics or you have tools or you have some widget or analysis, that can be a much lower barrier to entry. But if you're making a therapeutic, it's, you know, it's going to take money. And you'll look at diagnostics?

And tools.

Yeah, we look at all kinds of weird stuff. We look at computational machine learning, crazy things. You know, at some point, like 20 years from now, most— or maybe it's 50— but at some point, drugs are going to be discovered in computers. That doesn't mean that there won't be cellular assays and all kinds of other stuff, but like a lot of it's gonna happen there. And that, you know, and it's more likely gonna be, hey, and healthcare's gonna look like, I think, I get in trouble, and this is where I get over my skis in science, but I think I'm right. So I was having this argument with Jose Basalda, who's the physician-in-chief at Sloan Kettering, and he's a smart dude. I'm like, okay, here's what's gonna happen, Jose. We'd had some wine. So I think we're gonna add a preventive disease within 5 to 10 years. So you're gonna come in, we're gonna look at your circulating factors, whatever, methylation or whatever, pick your pleasure of what it might be. And I'm gonna tell you, you have 7 of the 8 real drivers for pancreatic cancer. And, you know, if it's me, I would like, hey, you know, I don't really want the 8th. I could wait around and then I'll have stage 0 or stage 1 pancreatic cancer, which is still quite curable, but generally would suck. So what if we took the sixth factor and edited it out, edited it back? Right, so you go from a 98% risk to a 1% risk. And if you could isolate that to the pancreas, why not, right? I think that will happen. So I think we'll be treating pre-disease. So if you think about that, if you're a pharmaceutical company and everything gets diagnosed early, I was having this, so I had this conversation with two different CEOs of top 5 pharmas, and one got it and one didn't get it. And one was like, I don't understand this GRAIL thing. Like, it seems like, you know, pan-cancer screening, like, I don't understand why that's gonna matter. I'm like, okay, you're an oncology company. I can tell you how it's gonna matter. You're totally fucked if everything is diagnosed at stage 0, right? Like, you sell drugs for stage 4. Okay, like, you can't figure that out? And I swear to God, they're like, oh, it's way out there, you know, I don't know. It's like, if I was in that business, I would put like 10 people figuring out like, if there's any probability of this happening, it's going to fundamentally, you know, that's Polaroid over here. And the other people are like, okay, I want to write you a check for $225 million. I'm like, first of all, you're my friend. And second of all, you must have vision, right? But, you know, that's super interesting because this company absolutely understands that it fundamentally changes everything if you can go to early diagnosis. And then you start thinking like, okay, is it gonna be conventional therapies that treat things early that exist today? Which maybe, maybe just needs surgery and radiation or something if you're in oncology, but, or maybe it's something entirely different. Like maybe if you're in psych, If you can identify a driver mutation like in some of my relatives, we know what it is, unpublished, 'cause we have access to interesting databases. If you can snip it out, would you do it? And so that's where it's gonna get really interesting. Treatment of pre-disease, treatment of risk factors. There's a big debate actually at this conference I was at recently which had mostly tech people. And the debate was, do humans have to edit themselves to be smarter to keep up with AI who's gonna dominate us and kill us? So there's a certain amount of probably weed being smoked at this meeting. But if you, you know, it is really interesting because Everybody said, oh, you know, it's terrible to, we were at TED last week and they threw this thing up, says, is it okay to gene edit embryos to make things smarter? And, or, you know, to select attributes. And like, you know, most of the TED people who are tech people was like, no, it's not okay. And I said, "You mean you selected your wife based on no attributes of physical beauty or intelligence or anything else?" They're like, "Well, that's not the same." I was like, "Really? It's not?" Because it kind of is. And it turns out Chinese are already doing that. It is out of the bag, it's happening today. Like, you know, there's very sophisticated analysis of embryos happening. My own view is it doesn't matter because it's gonna be a bell-shaped curve again. You can think you're picking the smartest person, but the randomness is, you know, so it's gonna be this, and everybody's gonna pick different stuff. But I think we're gonna be in the realm of these technologies rapidly going into data-driven treatment of pre-disease, you know, and treatment of things way outside of cancer that we didn't think. So the pharmaceutical business is gonna look entirely different 20, 50 years from now than it will today. And I don't, you asked me earlier, I don't go to conferences, I don't listen to pharma because I'm not supposed to think about things that they're already thinking about, right? That's the idea of, you know, It doesn't mean I'm not going to think about the science, but I don't— I want to think about the graduate student in the back of the room who has this crazy orthogonal idea, or I want to think about the famous person who is super smart, who really is 4 or 5 standard deviations, who has, you know, figured out how to do nucleotide editing or whatever. And it turns out, like, there is very much a 99-1 It may be 99.9991, but there's a disproportionate impact of some really smart people on this area. And it is not— you cannot just go down the list of degrees and stuff. So it really is hard to figure out where they are. If you go into the head of the departments at UCSF and say, who are the smartest people here? They know who they are. And they may not be even people with the most grants, or somebody might have just gotten a couple grants rejected, but they have some orthogonal way of looking at something, or they're the Mark Roth of the group where they made their career in one area, but now they're thinking about a new area in a different way, and that's where the really cool stuff comes. And you have to be kind of open-minded to that, at least I think.

Well, we sort of need to stop here, but just to pull a couple of themes out that are orthogonal ways of looking at problems. Get out of the sort of herd mentality and go try something completely new. Don't send in your business plan on a website. It's knowing someone who knows someone who gets connected to you. So it's all about relationships. And that you truly are one of the few VCs who will take risks around here. So that sandhill crab doesn't like to do what you do.

No, they might. There's like some of the— I think the angels are are worth talking to, just to realize what I always tell people about angel investing is just create flexibility in your financial structure. So take it as a bridge, make sure that you don't bake in some wacky high valuation that is gonna prevent me from, you know, I just won't do it, right? So you'll have to figure out, but it's real easy to get, you know, stars in your eyes. So you have to figure out a way to approach these things in a way that's flexible enough. But, you know, if you need some money to like reduce stuff, practice, or just get up and going, take the money but do it with a cap or a convertible or some sensible structure where you haven't kind of priced yourself out of the market. Okay.

Well, on that thought, we're going to thank Bob for coming and speaking with us today.