EP 02: Jason Kelly (CEO of Ginkgo Bioworks)

Yeah, thanks for having me on.

Oh, I know that very clearly, actually. I know because I went just this last weekend — there was a, at the Hollywood Bowl in LA, they did a Jurassic Park screening with like the whole concert orchestra and it was amazing. And that movie came out like when I was 12 years old, or 12 or 13 years old.

100%. Yeah. Michael Crichton fan through and through. I'd read like every book and then I saw Jurassic Park and I was completely sold. Yeah. So yeah, that — it's actually like among bioengineers in their mid-40s, it's like 80% are a consequence of Jurassic Park. Yeah.

It's huge. Yeah. Like, it really shouldn't be underestimated. I think it's a big pull. Yeah.

Yeah, so I'll give you like the intro to bioengineering, right? So the basic idea is, you know, inside of every cell is digital code in the form of DNA, right? So it's A's, T's, C's and G's, you know, not zeros and ones like in a computer. But you can read that DNA. So like, you know, you've heard of like the Human Genome Project, and that's a technology called DNA sequencing. So you basically like grind up a cell, put it into like a washing machine-sized device, a DNA sequencer, and then out comes on your screen the letters of the genome, right? So your genome would be like about four billion letters long. A bacterial genome would be like a million letters long. All right? And so that's reading DNA. There's a parallel technology called DNA synthesis or DNA printing where you do the exact opposite. You go on the computer, you type A, T, C, G, you hit print, and out of machines at Ginkgo or companies out here in California like Twist Biosciences, a piece of DNA gets printed exactly to what you ordered. And then you can essentially, if you've heard of technology like CRISPR, that allows you to find a spot in say that million-letter genome of a bacteria, open the DNA at that spot you defined and do sort of like a cut and paste, and you paste in the DNA you printed, and now you've essentially put in a new piece of code that makes that cell do something new.

And that's the technology of synthetic biology or bioengineering. It's really that ability to change the DNA code in an organism, make it do something new.

Yeah, I mean, so my experience was at MIT. So I was an undergrad at MIT and did grad school there. As an undergrad, I did all my summers like in the lab trying to basically clone one gene into a cell. Cloning is the term of art for what I just told you, like putting a new gene into a bacteria. And more or less failed. Like, I was not great in the lab. Probably why I'm not allowed in the lab. But it's unbelievably tedious. It's really hard to do. And so I was pretty demoralized by the end of my undergrad. I was like, 'This stuff just doesn't work.' And I met a professor, Drew Endy, who I ended up working for in grad school. And then he was close with another fellow, Tom Knight, who I ended up starting Ginkgo with, along with three other students I went to school with. And their vision was what I just said. It was sort of, 'Hey, we should really think of it like code, and we could program these things.' And so honestly, it was just the fact that they believed it that sort of inspired our generation of students to kind of like go for it, and that became the field of synthetic biology. I will say, if you fast forward now, you know, that was now like 20 years ago, 15, 20 years ago, and it's still hard, right? So I think this is where a lot of, I think what your organization and what we're trying to do with like accelerating science and some of the stuff that the National Security Commission on Emerging Biotech that I chaired, we're working on was, well, what are the bottlenecks, right? Because it is still like we don't see the pace in the biosciences like we see in computer science yet. And so what's holding that back? So I'd love to talk about that because I think it is still not a solved problem. And now...

But the upside is if you solve it, right? Like why is it so important to solve this problem, right? And I'll just nerd out for one minute.

Yeah. Like why should people care, right? And so like what makes computers such a big deal, right? You know, like if you think, if you go back, rewind, we're winding the clock 70 years, right? And computers were these sort of esoteric things, very small number of people were able to interact. Remember big mainframe, like that era of computing, like there was very few of them. And the reason that computers even back then you should have thought of them as a big deal was because they were programmable. So you take this device that you've invested all this effort in and you put in new code and suddenly it does a new thing instantaneously. And I'll just point out like your car is not really like that, right? Like you don't insert new code into your car and suddenly the wheels change or something, right? Like yeah, your dash system or it has a computer in it, but the physical object is not really programmable. Okay? But computers were programmable. So they could do all this different stuff by just changing the code. All right? However, it was limited, right? If you think about what computers can manipulate, okay, it is information. So what did computers disrupt? They disrupted media, telecom, finance, communication technologies, all the things that involve information bits. What did they not disrupt? Hamburgers, you know, right? Building materials, right? Like these are the physical world, right? Okay. And that's because they don't move that stuff. Well, a cell, all right, is programmable. Like I swear to God, like inside of a cell is DNA code. You can change it. You can make it do new things. You insert new DNA code and it can change color. It can make a new thing because it doesn't move information. A cell moves atoms. And the world we live in, you look around the room you're in, everything, it's all made out of atoms. It's made out of physical stuff. Cells are the only thing we have that are programmable in the physical world.

I mean yeah, it's interesting you bring that up. So one of the other things that's amazing about biology is like there's the existence proof of wild biology out in nature doing all these crazy things. Right? So again I'll give you an example. Let's like, over here there's a plant in this window behind me. Okay. Right. So let's imagine like we didn't have plants and you know Tim Cook got up on stage at Apple and was like, hey, you know, I've invented this thing, the seed. All right, and I'll tell you how this works. I dig a little hole in the dirt. I put this object in it. I cover it up with dirt. And then with just the addition of air, water, and sunlight, this thing starts self-assembling physical atoms out of the CO2 that's in the air. It builds solar panels to collect more energy to do more of this. And then it starts to do basically nanoscale manufacturing, producing these little fruits at a level of atomic scale resolution that exceeds what TSMC or Nvidia or Intel can do with semiconductors. It places atoms better than they do. All of that for free if you just put it in this hole in your backyard.

Literally insane, right? Well, you know, like what is that technology? That's crazy. So that's the potential, right? The potential is, oh my gosh, this, as a technologist looking at biology, so far exceeds what we can do with our human-engineered systems. It's crazy, right? And that's before you get to the fact that like we're made out of biology. Our food, you know, right? Like the atmosphere is produced all of biology. So there's just a level of capability, and we as humans, we kind of take it for granted because it's so familiar, right? Whereas like when we see new technology and inventions in computer science, electrical engineering, and electronics and things, they all feel new and different, and so we kind of ignore this thing over here. But it is in my opinion the most important technology for humanity in the future is biological engineering because it is the programmable technology that moves atoms, and we already know it can do freakish things, right? We know it because we see it when we just look outside as a technologist and look at something like a plant. And so that's all the potential. But oh my gosh, there's so many cool things if you go around and look at what biological systems are already capable of. It's wild. So yes, the chameleon is just one of them, but there are many things that biology does that's kind of incredible.

Yeah. And again, the way to think about us is we're like the tool providers. So the reason I can be in all these different areas that seem so disparate is that I'm providing you the tools to program and design biology to make it do new things. Okay. And so our customers end up going to market in all these different areas. But like a project we've worked on for years that's super interesting is providing the tooling and infrastructure in partnership with Bayer Crop Science, biggest biotech company in the world, and they're engineering a microbe to produce fertilizer for corn and other row crops. And the idea here is, I know you may remember like crop rotation from elementary school, like oh you plant this and then you plant the legumes or whatever, right? So the reason that they used to do that was that like crops like soybeans or other legumes, they actually have in their roots microbes that take nitrogen out of the air. Okay. The atmosphere has a lot of nitrogen in it and make fertilizer. Okay, and if you're like, I was, my undergrad is in chemical engineering, like the pride of chemical engineering is the Haber-Bosch process. The Germans invented it and it's basically pulling atmospheric gas into a big chemical plant, burning natural gas, and then making ammonia, synthetic fertilizer by the ton, okay? And it's magical, and we dump all this fertilizer, and crops like corn which don't have those microbes get synthetic fertilizer and grow like crazy. And that's like modern agriculture, but it's not great. It uses a lot of CO2 production. There's runoff. You have a lot of local environmental issues. And so the question that Bayer posed in partnership with us was, could you engineer the microbes that live on the roots of corn to act like the microbes that live on the roots of soy? And to your point about chameleons, like what's cool is we already know biology can do this. It can run Haber-Bosch effectively like in the roots of soybeans. So why couldn't you also do it in the roots of corn, right? Like we know it's got the capability, right? Does that make sense?

Totally. Yeah. So there's actually a pretty cool — so Senator Todd Young from Indiana, he's now chairing that National Security Commission on Emerging Biotech that I chaired previously, and he just put out a series of like four bills based on recommendations from the NSCEB, and one of them is on this project to catalog the genomes in the national parks. Okay. And to answer your question, what we do for this project is we do absolutely leverage the fact that people have previously sequenced the genome, read the DNA of the microbes that live on the roots of soybean. And that is what gives us the code that allows you to then go figure out how to program the microbes on the roots of corn. You want to basically like take the soybean microbe code and move it over. Okay? And like it doesn't work on the first shot. Like so you need to try a lot of different things. It actually helps a lot to have seen other organisms that do this too. So like soybean microbes aren't the only things that fix nitrogen, is what it's called. And so you want to have as many examples of that as possible. And so we really would benefit from bigger public databases cataloging all of the genomes out in nature that have all these incredible tools, right? Like the chemistry allowing you to just out in the wild grab nitrogen out of the air and turn it into like ammonia for the crop is wild, wild chemistry. It's only because you have these proteins, which are kind of like the building blocks of cells, that are able to do that, and that was invented not by humans, that was invented by mother nature. And what we've done is just catalog that code, learn to use it, and then as bioengineers we're repurposing it to something that could be, you know, 5% of global greenhouse gas reduction and a great savings in local fertilizer runoff. But that's only because people have studied the biology in nature. And so there's a really powerful feedback loop that the government could enable, which is to study more of that biology in a high-throughput way, generate these big public data sets so that people can then go design the next generation of biotech products using all that, and obviously AI and data science are a big part of that too. So anyway, that's where it comes from. Yeah, absolutely, we go and study those things. Yeah.

Yeah. Okay. So now let me get to what we've learned of 15 years of banging our faces into this because it's not easy. So, and I'll also bring up a little bit like what we're seeing across the biotech industry generally. So we obviously are not an academic lab. We're industrial. We do industrial research in partnership with companies that are trying to commercialize biotech products. So that's the translational layer where people are kind of making new discoveries but then they want to go commercial. That's the part that I follow, that I know the most about. Okay. And so I think we're in a small or maybe a decent-sized crisis right now, I would say, across the biotech industry that it's worth people tuning into this, and you know, folks that are developing these policies in DC or otherwise like to understand. So, all right, let's look at biotech. So I just told you we want to program biology. It's amazing. It's computers in the 50s or 60s. I think it's going to be the biggest industry in the history of the planet eventually. All right. Okay. So like why is it not today? All right. And so there's two problems. There's a technology problem and there's like an application problem. All right. So I'll get back to the technology problem. Let me just speak to the application problem first. So on the application side, genetic engineering got invented in 1978. There was a Nobel Prize, Herbert Boyer at UCSF moved the first gene from one species to another. Okay. And that was the birth of genetic engineering. And it was much older. It wasn't cool technology like CRISPR. There was no DNA printing. It was very old school like cutting and pasting literally. But that was the beginning. And he founded a company called Genentech. And Genentech, the very first gene that got moved from one species to another that they commercialized was human insulin being moved into bacteria. And once you put it in bacteria, you put it in a big tank right down the road here in South San Francisco. I'm in San Francisco today. And you produced insulin, and type 1 diabetics — my dad's actually a type 1 diabetic — who couldn't produce insulin since they were kids and were taking insulin from pig pancreases back then, could get human insulin because you took the human gene and you made it in bacteria and you grew it almost like you would brew beer, and out came human insulin. Okay? And so that was like the first application, like the first app on your phone basically. The first app of genetic engineering was human insulin, and it was commercialized by Genentech and Eli Lilly. All right. I still remember like the bottles that my dad had.

That's the beginning. That was the birth of biotech. Okay. And so you had through the 80s things like EPO, human growth hormone, a whole bunch of other human proteins started to be made in bacteria, and then you got antibody drugs like for cancer and so on. And then in the ag space you had companies like Monsanto and others pioneer putting genes in for things like protecting against certain crop pests and also for things like herbicide tolerance. Okay. And so those were like the big things. And then the last small niche application was like enzymes for laundry detergent. So you know like cold water laundry detergent. Remember, we used to have to always run laundry hot. Anyway, cold water laundry detergent is because there are enzymes that break up the dirt and they're made in the same way with a big engineered bacteria in a tank, and you add that to laundry detergent, you can save all this energy on hot water. Okay, so those are the three apps. And if you fast forward to today, overwhelmingly 90% of the applications for all of genetic engineering is in human therapeutics. Okay? And that's like okay. But human therapeutics take a long time to get to market. And they're just like a need, they're still like a small industry compared to the totality of what you could do with biotech. Does that make sense?

But I bring it up because they're kind of like the last line of defense for all genetic engineering. So when I gave you that vision earlier of like we could program cells and hey you could grow a tree, why don't you grow a house, you know, why don't you grow a new building, why don't we make semiconductor chips, you know, using biology, like it's like why don't we have all this abundance of being able to grow anything we want through programmable biology? Well, that all is predicated on improving the tooling of genetic engineering, but in order to get money to invest in the tooling of genetic engineering, the apps have to... you have to be able to make money on the end apps. Like you're not going to get Windows if you don't have Word. You're not going to get Apple iOS, okay, right, if you don't have like Uber and things. Like there has to be ways to gain value from the tooling if you're going to get investment in the tooling. Does that make sense?

So here's your problem right now. Like the US is basically in the process right now of losing the biotech, like the, sorry, the therapeutic drug app layer to China. Like it is like so, and just so you know how this industry works, basically a biotech here in San Francisco or Cambridge, Massachusetts, those are the two hot spots, has some smart scientists, they do some genetic engineering, they invent a new drug, and then typically what happens is one of the large biopharmas buys that drug, okay? And they buy it like when they see a good result in a clinical trial. So they'll spend $2 million, they'll buy the drug, they'll take it to market, and you'll get, you know, whatever GLP-1 drugs or some new cancer drug or whatever. Three years ago, less than 5% of the drugs acquired that way were from Chinese biotech startups. First quarter of this year, 30%. Second quarter, 40%. Okay. So what's happening is the actual technology, not for manufacturing, this is like the usual oh we're going to make low-cost manufacturing, but no, no, no, this is the technology of discovery, the technology that really drives the tooling and all the money that flows back to the companies doing that is starting to leave the United States. And so if we lose that, yeah, we'll lose drug discovery, which feels like a mistake to begin with, but what you're really...

going to lose is genetic engineering. And it's very scary to lose genetic engineering because...