Arch Venture Partners co-founder and managing director Bob Nelsen is widely regarded as one of the most astute and successful life science investors and company builders in the game. The list of his portfolio company wins (if you think of raising money, going public, or being acquired as a win, which he doesn’t necessarily, read on) is impressive. In December, Denali Therapeutics raised about $250 million in 2017’s biggest biotech IPO. This January, Juno Therapeutics was acquired by Celgene at a valuation of some $10.97 billion—Juno had gone public in 2014 and Celgene paid $9 billion beyond the large stake it had already purchased. One of the most talked about new companies of 2017 was Vir Biotechnology, headed by former Biogen CEO George Scangos, which has already raised more than $500 million and done a wave of acquisitions. Nelsen is listed as a co-founder of Denali and Vir.
And that’s just a very short, recent list. Nelsen’s successes, using the conventional definition, were enough to vault him to number 5 on the 2018 Forbes Midas List of top investors; he ranked 16th in both the previous two years.
Based most of his career in Seattle, Nelsen moved in September 2016 to San Francisco. I caught up with him recently at a restaurant in his new home city, where we covered an array of subjects. Among them: his overall investment philosophy of thinking long term and not viewing IPOs as exits; how killing bad drug candidates should be celebrated; the convergence of gene therapy and gene editing; the rise of A.I. and machine learning in biotech; and his view that Big Pharma “should be very afraid” of what’s ahead, including the possibility that “insurance companies might become the pharmaceutical companies.”
Along the way, Nelsen shared the story of Arch turning down an investment in Theranos and some at his firm even betting when it would crash. (Here is just one of many recent stories on the spectacular flameout of the blood testing company.) And he declared that given the state of innovation in pharma these days, they should fire the CEO and board of any pharma that buys another pharma. “Fire ‘em all, that day. March them out of the office.”
All that and more, in this edited transcript of our discussion:
Xconomy: You mentioned once how Denali broke the traditional mold of biotech companies. I wanted to get your take on that—and what it says about your investment philosophy.
Robert Nelsen: I think that’s true. I mean, it did exactly what it wanted to do, which was get a bunch of shareholders in the company who want to build a really long-term franchise in neurodegeneration. The blood-brain barrier platform [that Denali has] is unlocking a lot of really interesting biology and allowing a lot of interesting new biology to be tested.
X: So when you say long-term shareholders, you mean people who don’t need a quick return and didn’t see an IPO as the exit?
RN: Yeah. We had 90 percent of our Juno [stake] at the acquisition. I think the problem with the industry is too much short-term optimization and not enough long-term optimization, which leads to poor choices, poor decision-making, too much focus on single clinical assets, too much focus on not killing programs that should be killed. Both in pharma and in biotech. In pharma, which is not a meritocracy, nobody ever wants to shelve anything because it’s like eliminating their jobs. And in biotech everything’s dependent on a single clinical asset. Even if you’re a platform company, whatever’s in the clinic first is what you get judged on, and nobody wants to kill that either, even if it should be shelved.
The biggest problem of the industry is the lack of killing stuff. But we’re in a highly innovative time, so if you feed the funnel with really interesting stuff you should have a lot more things to replace the stuff that you shelve. If you’re a company like Denali, or Vir, nobody’s worried about having enough stuff to work on, even if you kill stuff. So the idea is not just to have a portfolio. The idea is to have a better decision model, so that you make better decisions in general that are data driven. Then you’ve developed a culture like Denali or Agios have where they celebrate the kill. I mean, every time they kill a DC [drug candidate] at Agios, they have the same party that they have when they nominate a DC or they go into the clinic. Same party. Same speeches. Same everything.
X: Well, that gets to not being afraid of failure, and feeling confident that if you fail it’s not really a failure, if you’ve learned a lot.
RN: If you’re in our business, if you don’t fail, you’re not taking enough risks. It’s ok to fail.
X: Speaking of risks, there seems to be a lot of money thrown at new companies these days. What’s your philosophy on that?
RN: There’s a lot of money available–for good things and some for not good things. I think if you have the right science and the right team, they can use a lot of money and put it to work. People will always ask us, “How come you’re raising so much money for these companies—is it just because you can?” And I always tell them that we don’t actually think of it that way. We think of it as shareholder recruitment. So we’re just trying to find the long-term shareholders and then have them invest the minimum amount that they can to get them in, but for them to be invested in a way that we think that they’re going to stay with the company. In none of these companies, including Grail [Editor’s note: Grail is an early cancer detection company that has raised over $1 billion and has many high-profile investors, including Arch.] or Vir or Denali, did we take all the money that was on the table. None. It’s always a question of trying to figure out the optimum long-term shareholder base.
Even in the IPO, it’s not about taking all the money. It’s about figuring out how to allocate to the people who are long-term. And then the management teams are free, hopefully, to make the right decision. So, you know, if you just copy the phenotype you don’t get the strategy.
X: Can you elaborate on that?
RN: There’s a lot of people that are copying the phenotype [by] raising big rounds, and unless you have companies that can truly be a $10 billion company or a $15 billion company, you can’t justify raising $250 million or $500 million rounds and make a venture return.
X: It’s kind of staggering to me the way the dollars seems to go up, even for the Series A. Is that necessary these days?
RN: No. I mean, if the science is revolutionary, it’s fine. It’s an expensive business. But if you’re just aggregating all the rounds into one round and not changing the algorithm, you’re going to get the same result that they did in the ’90s and in the 2000s when they tried that, which is get a 1X [return] after five years. Which generally sucks as a return. Unless you can make 10X or 20X, that’s not venture capital.
X: Fair enough. Let me switch gears a bit and ask if Arch is raising a new fund and how big it might be? The last one you announced was in 2016. You’ve been pretty steady raising $400 million funds. That’s not going to cut it anymore, is it?
RN: We haven’t really commented on our fundraising. [But] we have considerably more money than that in multiple funds.
X: I’m looking for a little news here.
RN: I can’t comment on fundraising. But I would say, in terms of series As and series Bs, I think we’ve been investing at a pretty good pace.
X: All right. What percentage of companies that you invest in are what’s called venture creation, ones that you’ve created, versus people pitching you on their ideas?
RN: Probably 70 percent of the dollars, and maybe 50 percent of the companies.
X: So it’s still possible for an entrepreneur with an idea to come to Arch?
RN: Sure. In the last two years, we’ve probably done 10 deals that are in the $500K to $3 million range. People think that we just do the big deals, but we actually do quite a lot of small deals. And then, we’re pretty good at figuring out how to combine multiple deals into one bigger deal. So, you know, Juno is multiple companies. Denali is multiple companies. Vir is multiple companies.
X: Really? That you had previously invested in?
RN: Some that we’d invested in, but some which we just bought, or merged.
X: What has you most excited these days? What big trends do you see?
RN: Gene editing is going to be really interesting. I was a personal investor in Editas at the beginning. Other than that, we have Arbor, eGenesis, Homology [it went public in March], KSQ Therapeutics, and another one or two. And we’ll have something [new] in that space soon that we’ve been working on for a long time. I think you’ll see gene therapy, cell therapy, and gene editing converging into some new thing. Gen 2.0 cell therapy will be very interesting. I don’t think it’s just CAR-T cells, it’s other kinds of cell therapies and diseases other than cancer. And there will be some interesting things in machine learning and alternative approaches to drug discovery.
X: So do you see a platform like AI permeating into biotech?
RN: Absolutely. Machine learning will be a core of biotech within not very many years. The question has always been having enough data, and enough high-quality data, to be able to do machine learning. [But] using machine learning to make data on the therapeutic side is not interesting, unless you own the drug. So I think you need to be a drug company to use machine learning to reinvent drug discovery. And you’ll see companies like Grail using a lot of machine learning to be able to do early detection of disease. So I think there will be some fundamental change on the early detection side and fundamental change on the discovery side based on data.
X: Do you have any take on companies like Theranos [Editor’s note: the once high-flying blood test company has had a headline-grabbing crash, with founder Elizabeth Holmes charged with fraud and forced to give up control—here is a recent summary], and how that kind of super hype might affect your field?
RN: You know, Theranoses can happen at any time—if you have bad actors who aren’t science driven. We turned down Theranos in the early days, because there was nothing there. We used to bet people when it would crash. There were just like a million warning signs. There was nobody in the business that believed in that.
X: What about a company like Stemcentrx, maybe that’s a little different? [Editor’s note: AbbVie bought cancer stem cell drug company Stemcentrx for $5.8 billion in 2016, but its shares plummeted in late March when Stemcentrx’s lead drug candidate disappointed in trials, and the stock has only partially recovered.]
RN: That’s just also data driven. Somebody believed the data and they paid a lot of money for it, and they might have been wrong, I guess, based on the data, but that’s drug discovery.
I think the probability of AbbVie failing on their own program is very, very high. Possibly higher than Stemcentrx’s. I mean, most stuff in our industry does not work. There’s like an eight or nine percent success rate. So it shouldn’t be surprising. It’s surprising when people pay $5 billion for something, but that doesn’t mean it was a bad decision. Frankly, I think pharma should be doing more of that, not less.
X: That’s very interesting. Tell me more.
RN: They should fire the first CEO—and the board—of a pharma that buys another pharma. Fire ‘em all, that day. March them out of the office. Because the innovation is not in pharma, with a few exceptions, it is in biotechs and universities. There’s so much interesting innovation [in those places] now, and so much incremental shit in the pharma pipeline. It’s just unbelievable how much cool stuff we’re seeing. It’s going to totally change the business. If this early detection stuff, which we’re seeing in the data, works, and you’re an entire pharmaceutical and biotech industry that’s selling later-stage drugs, you should think about that one for a second. Because if there ain’t no later-stage disease, then how are you going to make any money? And if it’s early-stage disease, how are you going to treat it? What’s the business model look like? So pharmaceutical companies are going to have to become insurance companies. The insurance companies might become the pharmaceutical companies.
X: I’m going to have to think about that one.
RN: It’s not going to be the same industry as it is. People are going to intercept the patient with better and better data, earlier and earlier, even before they get to the existing healthcare system. Whether it’s CVS or Walgreens or Forward or some other thing. In China they have checkup centers that are treating 30 million or 40 million people a year. Consumers are going to be a lot more empowered with data on their own healthcare.
X: Do you think that finally either drives down the cost of healthcare or keeps it flat for a lot better performance?
RN: I don’t think the total cost of healthcare will go down. I think it will be reallocated. Preventive care will be more effective at saving lives and helping people live longer, and data-driven diagnostics will be probably the most important thing other than curative therapeutics.
So it’s interesting. Who loses in the new world is large hospital systems. Their bread and butter of ERs and late-stage cancer and other diseases are going away, so many will go bankrupt if they don’t radically change. The other folks who lose in that equation is anybody with incremental drugs or small molecules that aren’t targeted, which is 90 percent of the revenue of pharma. So if you’re pharma, you should be very afraid.