Ginkgo Bioworks Holdings, Inc. (DNA) Earnings Call Transcript & Summary

Good morning. Thanks for joining us. I'm Matt Sykes, the Life Science Tools and Diagnostics analyst at Goldman Sachs. I had the pleasure of welcoming Ginkgo Bioworks to our conference today, Anna Marie Wagner, Senior Vice President of Corporate Development. Anna Marie, thanks for joining us.

Thanks for having me.

Maybe if you could just kind of start out by sort of setting the stage and going over some of the key highlights this year, the opportunities ahead for Ginkgo. I mean, I know you've announced recently a couple of pretty important partnerships and maybe go through some of those.

Sure. And I don't like to assume that people know what Ginkgo does. So I might -- if you'll forgive me, I'll spend 30 seconds just giving a quick background on how Ginkgo has built its business. So we are -- unlike most biotech companies, we are focused on building an integrated end-to-end kind of enabling platform for companies that are interested in using biology across a range of products, across different end markets. So our customers come to us with a spec, that spec might be producing a small molecule for a certain cost. It might be producing a protein that catalyzes a certain reaction really efficiently. It might be a live cell that's going to do something in a human body or in a field with a crop. And our job is to design a program for that customer that will meet their spec and deliver the products that they're looking for. The business that we've built around that means we are doing services for a customer, but we're retaining scientific control through that process. But in that process, we get paid services fees. And then when products are successful, we would get downstream value in the form of royalties, equities, milestones, et cetera. So just a quick primer on our business. But then as we look at what's exciting for Ginkgo, the sort of leading indicators that we look at and the kind of OKRs we organized the business around our combination of technical developments that we keep on trying to drive. And we've been consistently and again, this year are sort of excited to continue to expand our capabilities and our scale. One of the transactions we recently announced is an acquisition of an R&D facility from Bayer Crop Sciences, where they're effectively outsourcing R&D to Ginkgo in the ag biological field. But as part of that, we're gaining a suite of capabilities for agriculture, which is one of, as I think about the markets that we can deploy biology into agriculture, one of the biggest ones and so really excited about some of the platform capabilities that we'll be adding there. And then continue to be really excited about the new markets and customers that we're gaining across a diverse range of industries. One of the more recent announced transactions with Novo Nordisk. And for us, winning deals with customers like that is really impactful simply because it's -- like Novo is going to doing this stuff, right? They've got in-house capabilities. I don't know if they're here, but they're really good at biological engineering and design. And so to have a partner like that choose to work with us and trust us with some of their most important projects is a really important leading indicator on the quality of our platform and ability to get work done for them.

Great. Thanks for that overview. I appreciate it. Maybe we start talking about the platform and generating new projects has been very important at this stage. However, maybe you talk about how commercializing more projects down the road could actually have a flywheel effect in terms of creating a greater validation of the platform and then generating new demand?

Sure. Yes. So there's no better marketing than Ginkgo completing a program and having other customer -- potential customers see that our platform has helped enable a product get to market. We -- again, just as a reminder, we're not making our own products. But when our customers see success and that success becomes public, we invariably see inbound interest from other companies in that space wanting to talk to us about how we might be able to support them. We saw that last year with Motif's announcement of launching their first commercial products, including a myoglobin that we had helped them with. We've seen that in the cannabinoid space. We also see it when we sign new partnerships. When we announced deal -- our first deal with Corteva, it was a really interesting moment in the agriculture industry. We're now 2 of the major ag companies we're working with Bayer and Corteva. We got another call that week from a third major agricultural company because suddenly, it was like, well, all of my competitors seem to be working with this company, Ginko, what is it that Ginkgo is actually doing? And how should we get involved. And so both getting new customers who are well respected in their field and having products hit the market is really -- it's really important marketing for us, for sure.

Got it. There's been a bit of a mix shift in terms of project focus by industry over the past year. Has this been intentional in terms of sort of the program selection or more of a natural evolution as you expand into other industries?

It's a natural evolution. And so our general philosophy is if the demand is there and it's a worthwhile project to work on, we want to be working on it. What happened -- has sort of happened naturally over time is that as our kind of demonstrated ability to do this work and candidly, our credibility is increased in the field, scientifically, we've been able to successfully win deals with more technically sophisticated customers. Had we knocked on the door of Biogen or Roche or Novo Nordisk, some of our pharma clients 7 years ago when we were starting to commercialize the platform, they would have left us out of the room, right? Like they just got these MIT grad students with some robots like they don't know the first thing about doing R&D in biotech. And so we had to sort of earn our chops, and we had to prove that we could deliver in these less technically sophisticated fields like flavors and fragrances and specialty chemicals and food so that we could earn our right to go into ag and earn our right to go into pharma. I think what we're seeing now is because pharma is clearly one of the most established existing markets for biotech R&D, our ability to just gain market share there relatively quickly is going to naturally create a mix shift, such that we're not just diversified, but it means the growth in that sector is outpacing others.

Got it. Maybe talk about how you can transfer learnings from one project to another and kind of how you see this dynamic developing? Obviously, there's a machine learning component to it that's iterative in its nature. Maybe talk about some of the learnings that you've achieved? And is it a cumulative impact as you kind of move forward?

It's a great question. Yes. So our platform is comprised of 2 core components or assets. The first is a wet lab, we call it our foundry. It uses a lot of robotic automation to drive efficiency and scale, but it's just cranking through unit operations of wet lab work, moving clear liquids around, measuring outputs, mapping, genotype, phenotype, et cetera. The second core asset, which gets to your question, is what we call code-based and not as a collection of physical assets, strains, sequences and set of pieces of physical DNA, et cetera, and digital assets, data assets, mapping of genotype to phenotype, enzyme libraries, gene clusters, all sorts of things that we can then leverage whenever we're approaching a new problem, we can say, "Well, what have we learned about problems like this in the past. When we have tried to engineer enzymes that looked like this in the past, what worked, what didn't work?" Maybe we just have something off the shelf, promoter libraries or strains that have been optimized to produce high amounts of protein, for example, that we can just pull off the shelf and leverage for a project to eliminate the need to sort of reinvent the wheel on the lab operations every time we can reuse those learnings. The way that we got there is candidly being relatively stubborn and countercultural within the biotech industry, our IP terms are wildly out of market for biotech. A traditional CRO does not maintain scientific control of a project. They are doing work at the hest of the customer. The customer maintains scientific control. The customer retains the IP. When we work with a customer, we own the scientific direction of a program, and we retain the ability to reuse the IP. We, of course, give the customer an exclusive license for their product and their field. We're not enabling competitors, so the customers are okay with us at the end of the day. But we do retain the ability to reuse the IP in other areas, and that's important because we have the experience that this -- again, biology did not evolve by end market. And so learnings from flavors and fragrances could very well be and have been shown to be useful for projects that we've done in other spaces. And so that -- to me, that's kind of our most important long-term moat is the ability to reuse the learnings across these projects in new areas. And the last thing I'll say, actually, it's important, the scaling of our foundry drives the scaling of our code base. So one way to think about how our learnings grow is that every experiment we do in the lab is fundamentally creating an opportunity to learn about the impact of a certain genotypic change into a phenotypic change. And that information feeds into our code base and is information that we can then leverage in the future. And so as the scale of our foundry grows as we're able to do more and more experimentation over time, so too is the pace of growth in our code base.

Got it. Got it. You've talked recently about your ability to sort of toggle some of these partnerships in terms of the mix of upfront fees versus downstream economics. Maybe talk about that decision-making process and what factors into it? Is it the estimated probability of success? Is it the opportunity for a longer-term deeper partnership? How do you think about the decision when you're thinking about that mix on the economic side?

Yes. There are really 2 major factors that drive that. One is the customer and where they are in their life cycle, a young startup that may be very sensitive to its cash position, maybe more able to candidly just do a deal when there's more back-end weighted or when they're able to use alternative currencies to do a transaction with us equity or otherwise versus a large company with too much cash and trying to figure out what to do with the cash. So the type of customer can drive that structuring. But it's also a conscious choice that we'll sometimes make and that could be influenced by the market. Candidly, right now, we are in a market where cash is important. And so we are going to be very sensitive to maintaining our cash and our cash runway. There have been times in Ginkgo's history, where we've felt similarly and we've move the pendulum over to optimizing for cost coverage on deals and derisking our investments in projects. And there have been other times where we've really leaned into downstream value because it's where we think long term, the most value lies. I think in this type of a market environment, you may see us swing more towards some of the upfront in order to derisk some of -- derisk cash position over the next several years.

Got it. And maybe just touching on the comment you made about the market environment, just given some of the weakening in the funding market, particularly for some of the early-stage companies. Maybe talk about how that has an impact on your business. And if we sort of move through the year and we still have this funding issue, what can you do to kind of augment?

Yes, it's a great question. I think there's sort of an immediate impact, which is certainly customers are going to tighten their belts. They're going to prioritize R&D projects. They're going to be looking to maintain cash, right? I think this is certainly going to be felt more in young companies than mature companies. And so that's going to impact where we think about hunting for deals. So there will certainly be, I think, an immediate impact there. The longer-term impact that I'm really excited about is that anytime there's a market disruption like this, it tends to yield some amount of creativity. And so companies are not just looking to cut costs. They're looking for new ways to cut costs, right, to how do we keep our innovation pipeline while not breaking the bank. One of Ginkgo's value propositions is a more efficient way to do R&D. And you can think about that efficiency in a few ways. You can think about it as the cost of a project. You can think about it as big cost relative to probability of success. And we have the ability to market against both of those factors. And so companies may look at this market environment and say, well, rather than having this giant fixed cost infrastructure where we've got a lab and we've got 100 scientists who -- like that is a fixed cost. Maybe now we actually want to reduce some costs there and leverage Ginkgo as a variable cost provider with lower marginal cost to drive more efficient R&D. And so one of the things I'm hopeful for, it's too early to tell if this will happen, but I'm hopeful for is that this kind of a market environment may drive a shift towards outsourcing. And we've seen that happen in other pullbacks and other areas of outsourcing, IT outsourcing, et cetera, shared services outsourcing more broadly. This is obviously a bigger bet for a company to make. But that is the -- the single most important kind of market driver for Ginko is the willingness to outsource R&D. And so if this market environment allows people to experiment with that more freely, it could be really helpful for us long term.

Got it. Maybe focusing on the foundry business. I mean, can you talk about the long-term strategy for earning a margin on the foundry business? Do you plan to take additional margin potentially as an offset some of the downstream monetization risk?

Yes. Again, I think it depends on where Ginkgo is in its life cycle and the market, I think. Long term, our view would be the probability of success should get to a point where it's really underwritable. And therefore, we would much rather take most of our value in the downstream rather than the upfront and really play into like the elasticity of demand, make it candidly like free, not actually free. But like many people think about it the same way you think about it alias, like you're not thinking about your data usage or your compute usage when you're starting a website today or if you're signing up for Shopify or something, it's a toll on your success. And so I think as we think about what is the way to maximize our TAM and our potential value capture, it is being aligned with the customer and their success and having it be a toll on the products. There is obviously -- it's going to take time to get to that stable point where there's so much royalty flowing in, that we are cash flow positive without charging the foundry fees. But that's where I'd ultimately love to get is the foundry upfront fees are a nonissue for the customer, and we take value downstream. In the interim and especially in this kind of a market environment, we're going to be looking to optimize foundry fees, no question. I think we don't -- it's not -- I don't think we're going to go all the way to like the CRO model of the world, where we forego downstream value, but we're definitely going to be looking to offset more of our costs in the project.

Got it. I mean some of the benefits that you could actually achieve in the upstream has to do with the increased scalability of the platform, right, and taking down your cost as the platform scales. You've had some charts in some of your decks in the past looking at that sort of scalability and cost benefit that you've been able to provide. Maybe any update on sort of the progress you've made from a cost perspective as your platform has continued to grow.

Yes. We updated that in our Q4 earnings deck. So we continue to see kind of what we call Knight's Law, basically, we started measuring this maybe 6 or 7 years ago. The output of the foundry has roughly tripled every year and associated with that has been a 50% decline in unit cost. Now it is a complex and imperfect metric. Not every unit operation we do, not every service scales exponentially, and has costs come down. We tend to find the bottleneck just move. It's a little bit like guacamole, like, okay, let's solve the high-throughput screening bottleneck, now let's solve the sample prep bottleneck, and you're constantly trying to drive step changes in bottlenecks and then identify the next bottleneck as quickly as you can, but we do continue to see progress there. A lot of the recent progress has been driven by multiplex operations. So that's when we're able to run instead of having to run hundreds or thousands of experiments in what's called in a raid format, a 96-well plates or so. We've been able to use biological tools to be able to run millions of experiments in a single tube that dramatically reduces the cost of doing that work, and there are more and more tools that are now at our disposal. We've made some recent acquisitions in that space as well that have helped us be able to deploy that type of technology eventually.

Got it. Just on biosecurity, you guys have been very clear and upfront and transparent about the testing part of biosecurity and the volatility in that revenue stream. Jason mentioned in the last call some of the longer-term applications, potentially wastewater, passive air monitoring. Any thoughts on in terms of like the future commercialization of these product offerings? I know it's very early, but also sort of the potential longer-term sustainability of the biosecurity revenue stream as we look forward on those on more of a surveillance basis?

Yes. So I think it's inevitable that a biosecurity industry will emerge. I think unlike cybersecurity, it's likely to have much more public sector involvement, and it's likely to, therefore, be much more concentrated. I think the most important thing that our biosecurity efforts to date have done for us. You've got K-12 school testing business as an example. It has earned us a seat at the table with the U.S. government and with international governments to help design what that solution becomes. So again, I think our best guess today is that the bulk of this, at least in the early days, will be government contracts to do pathogen surveillance, basically pathogen monitoring. It will likely involve a lot of the same tools that we use for cell engineering, which is one of the reasons why it's interesting for us to invest in this area, next-gen sequencing, for example, to identify what's the DNA that's floating around this room, combined with machine learning to help identify if that DNA is something we should be concerned about or not. But in the process of monitoring for pathogens, we'll probably also just find a bunch of interesting biology floating around that might have useful cell engineering properties, particularly as you start thinking about monitoring in unique and interesting places like wastewater. And so we're very excited about what the future of biosecurity looks like. It's something we've been thinking about for many years. And I think we've more than earned our seat at the table through the work that we've done to date. I would say like it's going to take time. We would expect and candidly hope that the school testing revenues will drop off and then over time be replaced by, again, I would expect it to be kind of pretty steady contract revenue, big government contracts and every now and then you're going to have a response that needs to emerge. And that's part of the infrastructure we would build, and so you'll have some lumpiness as threats are identified and contained, but that you'd have a much steadier stream of kind of defense contracting style revenue in the business going forward.

Got it. You touched a little bit on your opening comments about the Bayer partnership. We thought from a validation standpoint, it was very interesting. And in terms of a long-term partnership perspective in an industry, ag chem where synthetic biology has had some success. Maybe talk a little bit more about what you think this brings to Ginkgo. Obviously, you have the R&D capacity you're bringing on. But just longer term, how do you think this positions you within ag chem, and your view, frankly, on the ag chem as an end market for Ginkgo?

Yes. So I mentioned the capabilities that we're acquiring earlier. We're very excited about that. And the other thing, though, that this acquisition and kind of new partnership enables us to do is to open up our capabilities more broadly in the space through our joint venture with Bayer. We had been somewhat limited in who we were able to work with in ag biologicals. And as part of acquiring these assets from Bayer and restructuring our relationship with them not only is Bayer committing to a much larger partnership with us going forward, but they are also allowing us to work with other folks in this space. And so we're very excited to be unleashed in ag biologicals in a way that we really weren't historically. And then as we think about potential applications for biotechnology, agriculture is one of the single largest and most important areas of biology today. Obviously, plant traits are probably one of the top couple of most impactful applications of biotechnology over the last 40 years. There's a lot more we can do on the biology of not just the plant, but also the environment that the plant is in, and that's what we're focused on, maybe there are multi-tens of billions of -- hundreds of billions of dollars markets when you put it all together. So it's worth going after.

Got it. One of the main competitors for you that we've always kind of seen is actually the internal R&D departments within some of the larger companies. You've lately been successful, and you've had Roche and Biogen, now Novo Nordisk, but how do you approach that potential sale or relationship? Is it a longer sales cycle? Is it different in terms of economics? Like just talk about that type of thing because we do think that from a validation standpoint, from a long term kind of deepening the relationship, it's a really important marker for Ginkgo to show.

Yes. No, I agree with that. It's a complicated sale. And as you can imagine, there are multiple levels in the organization that we need to talk to, to make those successful. You need the R&D department to validate the science, but you also need the R&D department to not feel like they're saying goodbye to their departments and their responsibilities within the company. So we work very hard to make sure that it's a real partnership with R&D teams to these companies. And given the scale of these relationships, typically, we are also talking to CEOs, sometimes Boards of companies as well because these are big bets that they're making. Yes, they do tend to take a long time. With some of our bigger customers, there can be a mix of, let's just dip our toes in the water and test Ginkgo out on a program or 2, but it can start as a reasonably small relationship. Sometimes they'll go all in and it will be a big upfront, multiproduct collaboration in an area where they've seen us demonstrate success, but these are definitely long and slow processes. What gets us really excited is when we open the door with the customer, then the most important sort of the next leading indicator we look like after just -- or look for after just signing them up is, do we see repeat business from them, because candidly, whether or not the program is successful, they're going to evaluate whether they think we did a good job and did better than they could do. This is a high-risk industry. A lot of these programs are going to fail just because biology is hard. And when we see them signing up to do more programs with us, that is for us the most important indicator of success.

Got it. Do you have -- you kind of discussed what that first project might be in order -- I mean, I would think you'd want it to be a project that there's a certain level of success you can prove that validation and get more programs. Is there a level of discussion that takes place in terms of the initial program, if that's how that -- I mean I know it's case by case.

Yes, it's case by case. We do tend to focus our outbound BD efforts on areas where we think we have a real rate to win because we've demonstrated success in that space or have assets that are valuable in that space. But if a customer comes to us after that call and says, "Hey, we want you to work on this really hard thing that you've ever done, we've never done, but it's worth doing." We're not going to say no to the project as long as we have a kind of trust-based relationship where we're being upfront on the challenges and the opportunities on the flip side. All of our programs are also -- it's important to note, governed by what's called a joint steering committee. So the customer and Ginkgo will assign technical leadership to program governance and oversight. And so we're constantly evaluating how is the program progressing along its desired milestones? Should we refocus it? Should we allocate this somewhere else? Should we stop the program. So there's that constant dialogue with the customer as well, which helps build that trust.

Got it. Maybe a little bit about the competitive landscape. I think one thing we've learned over the past number of years is synthetic biology is very hard. And you've taken the approach of being more of a platform company in sort of the picks and shovels versus single product risk. But maybe talk about how you've seen over the past couple of years, the competitive landscape change and how that has impacted the sort of value proposition that you're providing to customers?

Sure. So there are sort of 3 buckets of competitors, one you just mentioned, which is customers doing it themselves, and that is 99% of what we see -- there are -- there's another bucket which is platform companies in very specific niches, very specific application areas. Really, the main area we see about is in antibodies. So if you want to go make an antibody drug today, you can call up any number of companies to go screen -- build and screen antibodies for you. It's a large enough market with a standardized enough workflow that companies have been able to build up kind of platform business models in that space. Those platform capabilities are specialized enough that they wouldn't be able to port over to compete with us anywhere else. And it's not an area where we've really felt the need to compete. Our view and the sort of bet that we make is that the general purpose platform generally wins in a scale game. We saw the same thing happen in semiconductors, Intel, beat the industry-specific chips over time. We think the same thing will happen here. But that's an area where it's certainly harder for us to compete in antibodies and we candidly just choose not to because we've done antibody deals, but it's not an area we proactively go after because there is vertical-specific competition in that space. And then there's sort of the third category, which are -- or I should maybe just say which were, there were a handful of companies that did think about having platforms across a few different industries in what I would call like industrial biotech right predominantly. And those companies have all over time, focused on making product. And it's not an irrational decision by the way, like it's very clear why they made those decisions, and they were probably the right decisions. It's just -- it's a high risk but when you make that because an individual product typically has a high risk of failure. But what was driving those decisions is that it is very hard and very expensive and takes a very long time to build a horizontal platform that is competitive with specialists. And we crossed back [indiscernible] in like 2016, but that was 8 years after we were started. And so as we think about a new horizontal platform wants to emerge, it's got to slog through the process of scaling to a point where they have equivalent efficiency to you just doing it by hand. And most companies aren't patient enough or can't raise enough capital to get to that point.

Got it. Maybe talking a little bit on the financial side. I mean one of the benefits you have is a very strong balance sheet and the ability to spend, but there's obviously many things that you can do with it. And so allocating that capital is important. But first, when we start out, just thinking about the potential operating leverage in this business, there's a lot of fixed costs that are embedded in there. How should we look at sort of as project volume ramps, the level of fixed costs, I mean, you've added a number of different additional labs and things like that. But in terms of fixed cost base and operating leverage, how should we think about Ginkgo going forward?

Yes. I might answer a slightly different question, which is what is the piece of our platform that doesn't scale well today?

Yeah.

So certain things scale really well. The foundry scales really well. All the CapEx in that foundry scale as well, the space generally scales pretty well. We can drive more volume through the same footprint pretty effectively. The piece that we are most focused on, the hardest challenge for us to tell, and I'm not going to pretend we solved it. This is very much what we are focused on right now is decoupling program management from a number of programs sort of that human element of determining how are we going to approach this problem? What -- how do we -- what's the scientific pathway? What is the organism we want to use? How do we want to scope it? That is still a very artisanal human process that is run at Ginkgo by PhD scientists. One of the things we've been focused on is trying to create a set of more standardized types of programs. Our first foray into this is something called CDKs, all development kits which was basically just packaging up capabilities that we are quite good at and marketing against those so that we could sign up new programs that would be easier for us to execute because we'd already done programs that shared similar engineering elements. And so it took less of that human artisanal management to run it. So I think one of the things you'll see at Ginkgo is that if you think about the distribution of programs in terms of difficulty, time, human involvement, you'll see -- you'll start to see some SKU emerge where we will very proactively sell into programs that are easier for us to run. They might be faster overall, they will require less human involvement, and then you will still maintain kind of that distribution of harder programs that will push our capability set out further as well. But that is the -- the hard part is how do we continue to scale into that human element of program management that I think we and all of our investors should be watching over the coming years.

Given kind of back to one question I asked earlier about the learnings program, have you been able to, from a people standpoint, become more efficient with the staffing of the programs? Meaning, does it take less people per program, depending on the program? And have you seen a progress in that regard?

Yes. Less so for the average program. Although this is the core difference between how Ginkgo works versus a traditional lab. We would typically have a handful for people and anywhere from 2 to 5 people managing a program and each of those people is managing multiple programs, whereas a traditional program might be run by 20 or 30 people in a lab, and they're all focused on one program. So that is already our kind of core advantage is the efficiency with which we're able to run programs because we're able to leverage these learnings and the functional capabilities of the foundry. But this ties into what I was just saying, like I do think over time, what we'll see is we'll see certain types of programs to be able to run with one person managing it or 2 people managing it. And then those 2 people might be able to manage 10 programs that sort of look like that, where we can find sort of common archetypes of programs will make it easier for us to drive down the sort of number of people per program. That's exactly what I was describing.

Got it. And as you see the potential for new programs to come in, if one looks like one that you've done and had success, is that a higher probability that you would really want to take that programs so you can...

Yes. Absolutely.

Got it. Okay. In terms of capital allocation priorities, are there specific end markets that you want to be focusing on? I mean I know it's such a broad application, and there's varying degrees of success, but also probably partnerships and economics when you look at different end markets. Are you -- is that how you think about things when you're looking at it? And if not, how are you thinking about things?

No. We generally don't think about investments by end market. Every now and then we'll make one like this Bayer transaction, obviously, is helpful, predominantly for agriculture, although there are some cross functional benefits. I think you might see us make some more targeted investment in pharma. As I think about the world, I think about it as being in basically 3 buckets. There is biology being deployed to a tank, and I broadly describe that as industrial applications of biology. It could be in consumer, it could be in health care, it could be in anything, but these are industrial applications where you are deploying engineered biology into a micro organism, which is going to go into a fermentation tank and you're extracting something out of that, some sort of molecule. That's category 1. That's been our bread and butter historically. We're very, very good at it. Category 2 is deploying biology to a field into a plant. That's where this Bayer acquisition is going to be useful. Category 3 is deploying biology into an animal, into a human predominantly. And then so I think we could -- you could see us make some investments that help us with that sort of translational research between what we do in the lab and the application area in the same way we've invested in that translational research and industrial applications through scale up manufacturing, optimization and ag through this Bayer deal. You might do something similar in pharma. But typically, as we think about capital allocation, we're focusing on what are the general purpose, horizontal technologies that we can invest in to help us across different industries.

Got it. Maybe in the time we have left, just one last question, just on customer mix. You've done a good job of diversifying the customer base, in terms of the type of customer. Maybe talk about how you think about that and the diversity you want to achieve? I know it's not necessarily the goal to do that. It kind of happens naturally, but maybe talk about how your customer mix has changed over time?

Yes. So end markets is the area where it's most visible where we -- where again, as I mentioned earlier, we had sort of start in areas that were not technically sophisticated and then we earned our chops and went after more and more technically sophisticated markets over time. But we do try to get diversity in customer size, deal structure, type of biology, right? Is it mammalian, fungal, bacterial et cetera. And the way that we effectively drive that diversity is our BD team is segmented by end market. Our BD team is segmented by kind of the big elephant hunters versus startup partners. Our scientific teams are -- there is some specialization and therefore, capacity for us to take on programs in certain areas versus others. So some of the -- just team structure naturally drives diversity because there's only so much any individual person can do or go after.

Got it. We are out of time. Anna Marie, thank you very much for the time. Really appreciate it.

That's the way things are. Thanks.