Y Combinator's Sam Altman Is Bullish on Biotech Startups

Among techies, Sam Altman needs no introduction. But among synthetic biologists, the president of the famed startup incubator Y Combinator passes pretty anonymously in his blue hoodie and sneakers.

Y Combinator is well-known (and rich) for funding Airbnb, Reddit, Dropbox, Stripe, and a passel of other more-than-familiar startups. But in 2014, Y Combinator founded its first synthetic biology startup, Ginkgo Bioworks, which recently raised $45 million in a Series B round. Since then, the accelerator has funded dozens more biotech startups.

At SynBioBeta, a synthetic biology conference that brought together scientists, entrepreneurs, and venture capitalists in San Francisco this week, Altman sat down with WIRED to talk about how YC became a victim of its success, what's up with YC Research labs, and why traditional pharma sucks.

WIRED: I imagine you usually get mobbed at these startup gatherings.

Altman: Actually not so much with biotech.

Hah, well just wait. Synthetic biology startups have raised a record $560 million this year. Why are investors so interested in synthetic biology now?

The two areas we look at are cost and cycle time. Because of tech changes like DNA sequencing, the cost and cycle time have crossed under the startup threshold. Roughly, you can make meaningful progress on single-digit millions of dollars and can you have an iteration speed that is measured in weeks or months, not years.

It’s remarkable how much the willingness to write checks on the order of tens of millions of dollars just blossomed in the last year. It’s the probably the area investors ask me about the most. Whereas two years ago they thought I was crazy: "What are you doing funding this stuff, it’s never going to be a thing."

Is the public perception of genetically modified organisms something you worry about? How does that factor into your investment strategy?

I don’t talk about this publicly, but I guess I will now. We make so much money on stuff that is relatively certain that we can take significant risk on the long game on other stuff.

With software companies, you often fake it until you make it, because you're reasonably sure you can make the code work eventually. With biology, shit sometimes just doesn’t work and you don't know why. How validated does the core technology have to be before you’re willing to invest in it?

We will take a look at it super super early. That’s unique to our stage of investing. We make a lot of investments and accept we’ll have tech failure on a lot of it. That wouldn’t work for Series B investing. We’ll take a lot of risk if there’s potentially a huge reward.

How do you know which ideas people can actually pull off, then? Ginkgo, the first synthetic biology startup you funded, has five MIT PhDs among its founders, including the father of synthetic biology.

We always bet on the people, whether or not they have the resumé. The resumé matters less to us the more we know about a space, and as we’ve gotten to be more of an expert on bio, as we’ve brought on more biotech partners, and as we have a network of biotech alumni that help us review these, it certainly matters less. At this point, we’re more able to take a bet on a true synthetic biology outsider if they actually have a great idea and know what they’re talking around.

What’s a bet that you might not have made a couple of years ago?

It’s funny that it’s not an insult in the world of startups anymore to say "this person dropped out of college," but I think it still hurts you in the world of biotech to say "this person doesn’t have a background." I don't think I can answer this without somewhat insulting a company.

Since Ginkgo has gone through the process, what you have learned about how to make biology startups work?

I think the mistake people make most often when they invest in other kinds of startups is they say, “This is totally different.” And so the things that matter like making a product that people desperately want, like talking to customers, they throw this out the window. That is a recipe for heartache and tears. What we mostly try to do is still evaluate whether we believe these founders can create companies that can change the world and scale for the next 10 years. Do we believe this is a great idea? Does it seem like a bad idea to other people?

There’s this famous observation that I totally believe: Great startup ideas are the ones that lie in the intersection of the Venn diagram of “is a good idea” and “looks like a bad idea.” So you want most people to think it’s a bad idea and thus not compete with you until you get giant. But for it to secretly be good. Unfortunately, most things that look like bad ideas are in fact bad ideas.

But some things are different—regulation, for example. An agriculture company might need two years before its product is approved by the EPA or many more years still for a medical device by the FDA.

It is true that many investors have an extreme short-term world view. Not us. Something that takes 15 to 20 years and the return is a $100 billion—that’s perfectly fine. I do think most investors don’t like heavily regulated businesses because of the uncertainty, because they're impatient. But we don’t view it as a bad thing. It takes longer but fewer people try it, so you have a little less competition.

Most investors hate uncertainty. In fact, I think there’s something a little bit wrong about me that I kind of love it, like it appeals to the gambler’s impulses, like there's this one moment that the FDA is going to decree it is or it’s not. It’s like a craps table sort of thing. So I think it’s a maybe a flaw. Certainly the rational thing is to prefer less uncertainty.

Can you afford to be careless about uncertainty that way because of the YC brand? You know that if a company goes through YC, other people are going to invest in just because of the YC imprimatur.

I wish they would stop that! The fact that Y Combinator has gotten to the point in the industry that it has is bad for us in all sorts of ways. The fact that if we fund a company it is significantly more likely to get other investors to invest without doing their due diligence is bad for them, bad for us, and bad for the companies. I know it’s ridiculous to say it’s a problem, but honestly it’s a problem for us.

It’s better for our ecosystem if bad companies die more quickly, and you have less talent dispersion and less competition for office space. If investors invest with no diligence in a bunch of YC companies that fail, they’re not going to blame themselves for doing not doing the diligence. They’re going to say, "I should stop investing in the good YC companies."

Do you see a tension between the culture of publishing and peer review in science versus startups keeping their technology secret? Theranos is a good example of that.

Actually, no. I think extreme secrecy is a bad sign in all startups. Very few startups die because they tell you exactly how their technology works. On the long list of startup killers, that’s pretty far down. Though on the list of entrepreneur fears, it’s pretty high.

Historically this is one of the things that has gone wrong with biotech. Traditionally it has been very focused on pharmaceuticals, and there’s been this cult of the molecule, the one thing that has to remain secret—because the thing that is valuable is the IP on this molecule. If you only have a single molecule to bet your entire company’s fate on, and it take seven years to find out if it’s real or not, then you’re going to delude yourself—intentionally, unintentionally, whatever—about whether or not it’s working. And you have to have huge secrecy about what it is because that’s your whole company’s value.

I have noticed biotech companies that are obsessed with secrecy are usually broken in a lot of ways. The companies that have a process that create multiple molecules of which they only need one to work don’t have all their eggs in this one fragile basket and are less secretive and less broken. We have a preference for people who have a platform to discover multiple molecules.

You recently announced YC Research, which sounds a bit like Google X, a bit like Bell Labs, and maybe even a bit like university research lab. What exactly is it?

YC is sort of in this unique position. We make lots and lots of money. We’re not public, we have a very long-term horizon, and we have a high problem flux. We have hundreds of companies coming through the doors every year. If you put all those things together, you can imagine in a world where we are more suited to have a really great research organization than a product company.

YC Research is an experiment to see if we can fund things that are not companies that are still important to the world. We’ve never looked at our mission as funding companies. We’ve looked at enabling innovations and turning up the output on innovation. I think startups have been the best way to do that, but they aren’t the only way to do that. It’s also a way try see if we can try to develop a new research model that is better than academic and corporate research.

The first group will be 15 to 20 scientists. It's what’s I’ve been doing last month. I went out and recruited them, and I’m actually going to run the first group myself. We'll be ready to talk about them soon.

This interview has been condensed and edited for clarity.