Transforming biomanufacturing using software with Martin Permin

Welcome to the Axial podcast. Axial is an early-stage investment firm based in San Francisco. We partner with great founders and inventors investing in early-stage life science companies, often when they are no more than an idea. Axial is fanatical about helping the rare inventor who is compelled to build their own enduring business. Okay, Martin, thanks for doing this. I think I always kind of like doing some initial talking before recording. So great kind of getting some more context on Invert and talking a bit more about what you guys have been building. But maybe to start off the conversation, just tell us who you are, what Invert does.

Yeah, thanks for having me, Josh. I'm Martin. I'm the CEO and one of the co-founders at Invert. At Invert, we build software to help people who do bioprocessing of various different kinds do their jobs better. So very practically speaking, we build a software service tool that lets you manage bioprocessing data fully from upstream to downstream, from lab scale to production. It helps you analyze that data, helps you model out what you should be doing next to help improve your, your yields or your stability or whatever other whatever other things you care about. Cool.

And so how did you, as a tech entrepreneur, maybe on the surface, get into bioprocessing, get into biotech in general?

That's a good question. It's a, it's a long story, but to make it short, my background is in, in sort of more non-biotech. I was very fortunate to join Airbnb right out of high school, which means I never actually went to school after high school. It was very fortunate for me. I didn't enjoy school very much. Joined Airbnb right out of high school. That was, as we all know now, a very good outcome. And I got very fortunate there. After that, I was the first hire and built the go-to-market function at an enterprise SaaS business called Hive that also worked out very well. And about late 2019, 2020, I was sort of at a little bit of a standstill in my career, in my professional life. Didn't really know what to do. My now again co-founder Holger and I were working on a startup that wasn't really working out and we're sort of beating our heads against the wall. And COVID rolls around and both Holger and I feel fully useless for not knowing what to do with ourselves. You know, we don't, know how to invent a vaccine against this thing or build a respirator out of motorcycle parts from the garage or anything of use. And so we ended up building this COVID tracking app for Denmark and for Sub-Saharan Africa. This open source project that's— before there was such a thing as contact tracing or COVID testing outside of PCR, we would ask people twice a day for flu-like symptoms. And then hand that data over to government to try to help make resource allocation decisions on the margins. As it turns out in retrospect, that is not at all what ended up mattering in COVID, but at the time in early February 2020, that seemed very relevant. And so Honker and I end up spending 6 weeks building this project. We get thousands and thousands of daily actives. It's a very successful project for some definition of success. But more importantly, by the time it was done and we had handed it over to government and so forth, we sat down and it had been well over 6 weeks of being on the news and hundreds and hundreds of volunteers, lots of legal threats around liability and of GDPR and so forth. And it's the hardest we'd ever worked, made no money, it actually cost significant personal money. And it was the most fun we'd ever had. And so it became sort of very evident at that point that the next— if we'd spent the last 10 years or so of our careers using bits and bytes to coordinate pixels on people's screens, that the next 10 years or 20, or for my sake, the rest of my life, will be about using bits and bytes to coordinate atoms in the real world. Luckily, through that process, we got to know a lot of biotech people, especially the people in very senior leadership at the big Danish biotech companies. We're both Danish by origin. NovoScience, NovoNordisk, these companies. And through that, got to spend a lot of time in that world. And I spent all of 2020, 2021 angel investing in real estate, especially industrial biotech businesses, and fell in love with the world, fell in love with the people in this world, and started thinking about what types of problems could be solved with a MorphoTech mindset.

And so through that process of doing a project with your future co-founder or even past co-founders while discovering, you know, just helping science or the physical world, as you put it, atoms, is more rewarding or not. It's also very exciting. How did you do that process of angel investing and then doing that project, hone in on bioprocessing as a big enough problem to work on, maybe potentially devote a big part of your life to?

Yeah, well, so the very practical thing that happened here is that I kept being in these meeting rooms, either in board meetings or in pitches, and as I'm sure you share this experience of being— talking to these absolute heroes of civilization who are— a lot of these are synbio companies, right, who say things like We can make spider silk using fungal cells. And if we can make spider silk at a low enough price point that we can replace a lot of use cases for steel, and that will have a great environmental impact. Or in therapeutics companies, we can cure X, Y, Z. I'm an investor in a company that just sold to L'Oréal that very much seems like they have a cure for acne in a probiotic product. And kept being in these meeting rooms with these absolute geniuses and heroes, and then leaving the meeting room to cross the hall to go into either a lab or a production site or pilot plant maybe, and seeing all these like 2008 Windows computers, tower computers with the password stickered on, a little sign that says do not update because the mass spec machine doesn't talk Windows Vista. And just this fully arcane, technology stack, which I kept having this repeated experience of feeling like these people are like Bruce Willis in Armageddon, fully just trying to save the world, both from a human health perspective and a climate perspective. And the tools that we give them are bows and arrows and Stone Age flint axes, which just felt frankly disrespectful to these people's intellects.

So like, I think a lot of— I'm a biologist, and I think we all have that, those kind of common experiences where there are devices in the lab that run really old software and you don't want to update it because you want to mess up what's already working.

Yeah.

And so that's maybe, you know, the historical reasons why all this stuff is that way. But you're somebody who's new, you don't have all that baggage. That's kind of— maybe you can see this as ridiculous because I think as a scientist you think, oh yeah, don't update the software because that's how it is and that's how it's always going to be. But as an outsider, you can call out something that doesn't make sense. But that's kind of the easy part. The harder part, how did you learn all the intricacies of biotech bioprocessing, especially as an outsider? What resources did you use? What processes did you use to get up to speed fast enough? To be able to then be serious, taken seriously in that world?

That's a good question. And I wouldn't profess to be at that level yet in many ways. We've been very fortunate in that we work with some excellent customers and excellent investors. So in some respect, somebody must be taking us seriously somewhere. But, you know, we, the part of our teams who are coming from a tech background, not a biotech background, know just enough to be dangerous. And we're very sort of acutely aware of avoiding what I think of as a pitfall, which is a certain amount of arrogance or naivete, which comes from coming from tech where everything is very modern and where it's very easy to look at biotech and think, oh my God, this whole industry is in the '80s, or at best, and sort of bring a little bit too much of an arrogant attitude into that. We're very sort of acutely aware of avoiding that. So, you know, I know a lot at this point about the world of bioprocessing, about the market dynamics, about who the players are and how they fit together. But you couldn't in any way, shape, or form let me loose in a lab and have me be productive, right? That is just not what I do, and it's not what Michael van der Hulger does, who's our CTO.

So did like, I mean, just like, just like the basics of it. Like, did you just like read a bunch of books or like case studies? It just, I think it's great. I would say this, it's very easy to learn about science, get a textbook, but to actually learn about the industry, it's actually rather tough to learn about any industry because there's no really good books written about business, in my opinion. How did you go about practically? Did you like, maybe, maybe your co-founders or advisors you brought on bring the expertise? But even for you, like, were there any kind of standout experiences or materials that helped you like make a big leap in your like just understanding of like biotech?

That's a great question, and I'd say there were definitely individuals. I don't know that there were any materials in the sense of books or articles or whatever. We were very fortunate, as I mentioned earlier, to make very close friends with very senior leadership and especially recently exited senior leadership at this Danish company NovoScience, which is, uh, to your listeners who might be unaware, is a very large Danish bioindustrial company. And a few of our good friends used to be the CEO and Chief Scientific Officers there, and Between them, they've probably seen more bioprocesses get scaled over the last 30 years than anyone else on the planet. And so in many ways, the first few years of our entering biotech were just tutelage under those types of people, hearing their war stories, asking a ton of questions, and really trying to grok how each piece in this value chain fits together. I think that really was the path. I don't think there was any real shortcuts. And I also, just for the record, will not at all profess to have sort of traversed all that ground yet.

I think just having the right people around the table, that's the hardest. It's harder than you think. I think you got— from listening to you, it seems you got a little lucky. I think most entrepreneurs I interact with, they get the wrong advice. And then it's like they get the wrong person and they get brought down this wrong track. That's a whole other conversation of how to avoid that.

Yeah, but I think that's exactly right. That's—

you know, you're grateful you've had some great people in your network. Uh, how did you go about then defining your problem? Because you have this whole broad landscape of biotech, biomanufacturing, and it's daunting. Most do anything. How did you go about saying this is the narrow problem we want to solve and then what is Invert building to kind of solve that problem?

Yeah, I think initially when looking at the field, it seemed to me that the thing that was lacking from a tooling perspective was that sort of combinatorial innovation really hasn't come online in a big way here. The industry is historically incredibly siloed for for IP reasons but also for trade secret reasons, especially on the process side. Most of these things aren't patentable and they're hard to build defensibility around from an IP perspective, so you really do just have to be very secretive and it creates— it leads to an environment where there just isn't a lot of cross-org sharing, even between companies that shouldn't on the face of it be competitive. What we looked for initially in what sort of prompted us to build Invert versus any of the multitude of other potential software businesses that one could build to service biotech companies was we looked for problems that seemed pretty universal across all of these companies. And so the problem that Invert solves is anyone who, is in the business of growing cells, either to take those cells and sell those themselves as a product, or to have them metabolize something that is a molecule of interest that you're then going to clean up and sell as a product. This could be an industrial product, it could be an ingredient, it could be a material, or it could be a therapeutic. If you're in that business, generally speaking, you're in the business of reducing your cost of goods sold. Reducing your cost of goods sold as a company that grows cells for a living, is basically figuring out how to optimize the parameters of growing those cells and cleaning them up that will lead to sort of an efficiency that is as close to the carbon theoretical sort of maximum flux you can get to. And what we observed early on was that from beer to biologics, that process of going through that design-build-test-learn loop of optimizing your cost of good cell It's really the same thing from a data perspective. And so we sort of early on realized that if we just built for bioprocess, we can sell across a very, very wide slew of the industry. Um, it's still vertical SaaS, you know, it's still a narrow part of the economy, but in terms of the molecules that our customers make, it's incredibly wide, um, relative to a lot more specialized SaaS tools, which also are around.

That makes sense. Just finding a problem set that is universal enough so you're not only servicing like a small subset of customers. And it's also, there's like a lot of growth. You know, there's on one end of the spectrum, there's like Jason from Ginkgo, grow everything as an end state. But it's also the base case, which is like, yeah, we need fermentation to make antibodies and all these drugs and And the big opportunity is there's going to be something like by 10, 20 years from now. I have the data right here. I was just reading about biosimilars. There's going to be something like by 2039, 2029, biosimilar market is going to grow from like $10 billion to $100 billion. So somebody's going to make those.

Yeah. And even that's optimistic or that's sort of an impressive number. But even then, I would point out that I think that the error bars on those things are very, very long, meaning making predictions about technology is just incredibly hard. And I think as usual, we sort of underestimate what we can— we overestimate what we can do in a year, but very much underestimate what we can do in 10 years in this industry. And the bet that we're sort of making here that we thought about making early on was that historically, when a bioproduced equivalent to a petrochemical or an animal agricultural product comes online, they gain market share incredibly fast as soon as they get to cost parity or quality parity, sort of quality-adjusted cost parity with the historical equivalent. And you'd much rather sell to that business the day before they fully take off than the day after. So it feels like we're sort of selling into an S-curve that has already bent, but where the baseline is already everything that gets made in bioreactors, which is, you know, from biologics to beer is a relatively large market that we don't, for the record, sell to actual breweries, at least not yet.

So maybe we can define what Ember's product offering is right now. It seems to be like two main parts, which is managing that workflow, managing that, like, manufacturing workflow, analyzing that data, which is, I think, tougher than a lot of people think, including me. And there's a lot more beyond that. But maybe you can just talk about the key parts of Invert's product.

Yeah. The way that we talk about it internally is that we really want to be the system of record for bioprocessing data. So if you think about bioprocess data, there's a couple of challenges. You have these instruments out in the real world. They capture data. You want that data in a central repository somewhere. Between different vendors of hardware, even two sort of almost identical bits of hardware from two different vendors represent that data in very different ways. So it's a lot of harmonization, a lot of normalization, a lot of trying to map all that different instrument data onto the same data models. We grab data from unit ops all the way from a shake flask that it then in turn at some point inoculates a production reactor all the way down to, to a finished product. And what we do is take all the data, stick it in a nice web app that you can use to manage that data, as you mentioned, analyze it in all the same way that you normally would in Excel or in JMP, but keep it on the same system. In a way that is collaborative both internally to your own team, but also lets you share that data with your contract research organizations, your contract development manufacturing organizations, or any other stakeholders internal or external that you might work with.

What's been the kind of the killer feature of all this? What's been the thing that like users have been just like, oh, this has changed my life? I have my own assumptions, but maybe you could just tell me.

I'd always love to hear yours before, but let me volunteer some first. Generally speaking, as I mentioned earlier, our customers want to increase their yields. That's a sort of a simplified way of looking at it, but they want to reduce their cost of goods sold. To do that, they do some experiments, look at the data, they design a new experiment, they go through that loop as many times as they can until they get better. And what we do is greatly shorten the time that is anything but doing that experiment, meaning the time that it takes to learn from the last set of experiments what you should be doing next gets greatly reduced. We reduce that time quite a bit by automating data ingestion, by automating data analytics, meaning because people want to look at the same data every time, we can auto-generate dashboards. We reduce a lot of time waste in sharing between orgs, between teams internally.— this, you took a sample, that sample was sent to a little chemistry somewhere. Is that a spreadsheet that's returned that has to get merged into some master log, or does that just get uploaded to Invert and automatically is there at the moment that the analysis was done?

I think the analysis part, that's just a big one where a lot of this fermentation data, when I look at it, I have friends who run these companies, I go, show me some data. It looks like stock trading data or something. It looks like these charts and just I have friends that work in quant hedge funds. It's like, oh, you can bring the same capabilities from Wall Street to fermentation. You can probably unlock so many more insights into what kind of fermentation runs, what parameters you should optimize for to get a better yield. Ultimately, you're trying to just, like you say, lower cost or just get more of what you want.

That's exactly right.

And so on that point of analytics, you know, you wrote an incredible blog post about the 6 S's of biomanufacturing. It was like, it was really interesting. I think it kind of stole some ideas from lean manufacturing and expanded them for bioprocessing. Maybe you can talk about what those successes are and how Invert kind of is building software to solve all those problems?

Yeah, listen, a lot of, um, we often find that it's hard to find common language to communicate to our customers, both existing and future ones, uh, about the sort of shared challenges that the industry faces, because, uh, very often they don't talk a lot between them. So we try to come up with some of these sort of simplified language that lets people mirror themselves, um, in those types of problems. And so the 6 S's that we talk about in that particular blog post, though I won't take any credit for having written it myself, just for the record, are the data sources, the sizes, silos, software, security, and standards. And to sort of summarize what we're talking about there is that across the bioprocessing workflows from R&D through to production, and just for the record, in first services, early R&D companies and companies that are in full-scale production and they run their actual production on the bird. There's a lot of disparate data sources that are all meant to be compared but are generated entirely separately from each other. And you could very feasibly, and most companies do, spend an untold amount of human resources collating all this data, trying to make sense of it, making sure there's context in, uh, this sample was taken out of this reactor at this time, and therefore the work— the, the actual amount of fluid in this reactor is lesser than it was a little bit ago. So now we have to adjust all these calculations that, um, that try to figure out what the current flux is and so forth. All that context ideally lives in one system so that you're not continuously trying to figure out what it is is happening. And that's a lot of what we talk about in that in that piece, what parts of context needs to be pulled in? What other software suites have to be connected here so that we can pull in data from those? What types of standards like FAIR and other industry standards or FDA requirements here do we need to live up to? Are we talking time series, as you mentioned earlier, just like stock trading? We're talking long time series here. What types of these are we comparing? Are these apples to apples? Are these oranges to apples? And so forth. That's a lot of what we spend our time thinking about.

I think that makes a lot of sense. That's what makes it a little different than other spaces like genomics, like you were talking about before. It's just there's so many different moving parts to fermentation. Trying to produce an antibody or cell-free meat. It's everything from making sure you're compliant to temperature, pH. And so having a unified software solution to manage all of that. And so for building that software to do all that, how do you think about like standardizing runs? How do you think about, you know, One of the last features you alluded to was making it very easy for you to share your data to a CDMO or another partner. What's been the kind of the benefits or your customers have felt when you have standardized analytics? You can have standard formats. It makes sharing that much easier.

Yeah, there's a number of answers to this question, and I hope that they're all valuable. Any scientist listeners, to the show will recognize the, the time suck of creating reports, either for managers or external stakeholders and so forth. A lot of what we do is just auto-generate those reports. That turns out on its own to be a very, very valuable bit of product. Another such example is the relationship between a, say, a contract manufacturing organization and a sponsor, their customer. We're seeing now, we work with a number of CMOs and CDMOs and their customers, and we're seeing now that whenever a CDMO is on Avert and the customer is on Avert, again, that tech transfer, that sharing of data that describes the process to be executed, gets a hell of a lot faster and shorter and more enjoyable. It sort of takes it from, here's this Google Drive or SharePoint folder with hundreds if not thousands of files and descriptions of those files of what they represent, to something that's a little bit closer to sharing a Google Doc in the, in the sense of share with Josh at, yeah, so forth.

Interesting. So with all those kind of product features you've built out, there's one thing you're an expert at though, it's go-to-market. So that's the one thing you know. And so how did you think about go-to-market for Invert and you know, as you layer on your knowledge learning about biotech, what, what kind of strategy did you find that was successful for Invert to get some early customers? Were there particular markets that were more amenable to software initially than others in terms of sales cycles? And also, were there certain features that were more attractive to sell versus others?

Right, that's a great question, and I think the main point to make here is that we haven't invented this particular wheel in any real sense. You know, we sell to companies that make cells, they grow cells in bioreactors. There aren't that many of those companies, on the order of sort of thousands, right? That means that we can very feasibly just call them all up and ask if they would like something like Invert. In reality, it's a little bit more complex than that because you have to find the right person at these companies, and some of these companies are very large, and some of them might be disqualified as customers for whatever reasons. But practically speaking, that is what we do. We initially, and I will admit to having been a little slow to sort of increase our public image, but we initially were very quiet about what we're building, not because we're afraid that somebody else was going to come along and steal what we're building, but because We wanted to make sure that what we're building was very informed by what prospective customers told us they wanted and would make a difference in their working lives. So the first— our company now is about 2 years old, and I say that really the first year and 3 or 4 months, we're just working very, very closely with a select number of smaller initial customers to really understand what the gaps in their workflows were. And where we can really create a lot of value that either reduces time stock or time wasted or gives them capabilities that they didn't have previously without going out and hiring data scientists or data engineers and so forth.

And so on that kind of initial point of like a build-up, customer discovery and refinement, like how do I, like were there any like transformative customers that maybe helped you go from a demo to a product, or what type of experiences within those periods where maybe you were a little more quiet that set you up for where you are now? One example is just having one customer, but maybe there's something other— something else I'm not considering. But in those early days, you look back, what kind of was a key driver for the growth you have now?

Yeah, we always are in a pickle here because we, for the most part, can't talk about our customers and what they use EverTwo for aforementioned reasons of secrecy in the industry. But I'm sure there's a couple that we can talk about and who we mentioned on our website. We just published a case study with Sci-Fi Foods, who, as you know, I believe, because you are close to that team, grows mammalian cells to make lab-grown meat. They've been a very informative earlier customer to us in understanding how growing mammalian cells is different from growing microbial cells, which is most of the rest of our customer base. And again, it's just a very, very close working relationship there. There really is no alternative to just spending a lot of time talking to customers. And understanding what their needs are. And I'd love to sort of be able to provide some sort of shortcuts here, but I truly just don't think that there are any. There really is nothing to do but to talk to customers and build product.

That makes a lot of sense. How did you not get overwhelmed, even though there's not that many biotech customers versus like enterprise software and other types of industries? How did you not get overwhelmed by the diversity and get pushed and get pulled into a whole different cell types, different therapeutics, non-therapeutics. How did you— did you have—

did you—

when you built Invert, did you feel you had to narrow your focus to a particular subsegment of biotech, or was it generalizable enough where my product can sell into anybody who's growing something? So it doesn't require that level of expertise or customer empathy, right?

Yeah, so I think that's a great question. And first of all, we've done things a certain way. The jury's still out on whether or not that was the best way to do it, right? It's hard to, it's hard to judge the decision by its outcomes. And we've thankfully done very well in the 2 years we've been around, but there's a chance that we could have done a lot better by doing something else, right? So don't, so don't take any of this for gospel. But our initial sort of go-to-market impulse here was that the traditional sort of advice here would be to find one niche, narrow down on it, really win that niche, and then expand from there. I took sort of a counter position to that very early on because whatever niche you narrow into there, I think you're gonna— one of two things can happen to you in our particular field. Um, either you're gonna set yourself into an addressable market that's too small Or you're going to end up selling only to biopharma customers, and it's going to take you a long time till you get to revenue or actually a product in use that lets you get that feedback cycle of usage and product improvement going, which is what you need. So we sort of very, very intentionally initially went after customers that were quote unquote easier to sell to. That means customers in less regulated industries. We also very intentionally have in the first 2 years of our existence made sure to land customers who work tightly with customers who do all sorts of different types of bioprocessing from gas fermentation to lab-grown meat to traditional industrial biotech in microbial environments and so forth. We really wanted to make sure that we were covering all of our bases. And building a product that without any customization works for all those different use cases. So we think a lot about designing for the extremes, designed for highly regulated therapeutic environments and designed for less regulated, say, biopacifiers or something like that environments. And the middle part will really just sort of solve for itself is our belief here.

And that requires some really disciplined product design. I mean, just from knowing a few companies, biotech or SaaS, sometimes you build something that's so bespoke and you kind of break your internal rules to get that customer. But to be able to like be strict about our product does this and it's selling that product, it's pretty hard because, you know, ultimately, like, you may be alluding to a lot of these biotech software companies become consulting businesses and, you know, do that for a few years and it kind of maybe gets— potentially gets stuck So on this point of, do you have any fun stories of, like, installations or customer acquisition? Like, did you visit some cool bioprocessing site? Because these things, if you actually visit a biomanufacturing facility, some of them are pretty cool. There's some ones in Brazil I'd love to visit. Like, these massive steel things, and it's pretty impressive. There's one in Sacramento you could go visit. Any fun stories like that visiting the site that convinced a customer or fun customer stories?

There's certainly a lot of very, very, very cool sites here in this space and they're just marvels of engineering really and they're truly magical. Just to not tell them sort of the most obvious story here, one company that we worked with early on and we actually have paused our relationship to for a little bit, but we'll take our relationship back up with again, is a company based in respectively San Diego and Copenhagen, Denmark called White Labs. White Labs grows yeast for microbreweries or for breweries in general, meaning they grow liquid yeast and sell it to breweries. They grow those cells in these very, very, very cool bioreactors. And for the listeners, I'm using quotation marks here for bioreactors, which are plastic bags. Um, not in— don't think sort of traditional single-use, uh, bioreactors here that are still inside of a steel vessel. These are just hanging off of really what is a glorified coat rack. Um, but they're incredibly sophisticated and, uh, very, very, uh, in a very, very cool way. What they do is Once they're done with that fermentation process, the bag itself will be heat welded at such a point where the majority of the actual product will be at the bottom of it and the majority of the waste product will be just above. And they heat weld it, they cut it off, they take that reactor itself and ship it to the customer. So it's completely aseptic the whole way through. And the actual sort of manufacturing vessel is also the shipping vessel that the customer receives, which I think is really, really cool.

Like, they grow yeast in like an IKEA bag or something.

That's right, there's a lot of, there's a lot of cool labs that do, that use stoma bags for these types of purposes as well, because a stoma bag is, you know, 40 cents or whatever they are, and they're made in the clean room and generally easy to handle.

That kind of shows how easy it is to grow yeast. Grow things. If you can get yeast to make it. Yeast is not that— I've grown yeast. I've made beer. I remember one in college, I was curious. I wanted to use yeast to ferment alcohol, and I did. It was like, oh, that was pretty easy.

People do it in prison every day.

Anybody can grow anything. So that's the power of biomanufacturing long term, is it democratizes production.

Yeah, listen, I'm glad that my monoclonal antibodies don't get produced in quite that way, and that we have slightly tighter regulatory regimes around those types of things. But it does drive home, I think, a very important point that growing cells is growing cells. And I don't mean that in any way as to diminish the sophistication of the work done in this industry. But it is truly astounding just how similar making a very large variety of products can be from a process perspective, which I really think is what we as an opportunity have leaned into.

And so for Invert, you kind of have these customers and they're using your software. How did you go about thinking about the business model long-term evolving? Right now, the core value prop is you help customers grow things more efficiently. But there are also other big problems in biomanufacturing, especially in therapeutics. Compliance. You can read the news every few months, there's some sort of CDMO that gets fined for fungus growing in their facility or something. How do you think about Invert evolving types of customers you serve? I think on your website you allude to CDMOs being part of the equation. And how do you think about that product set feature evolving, especially around compliance, which, which for higher value products like antibodies becomes just really, really important? It could, you know, during COVID you saw this at Catalent, uh, where they had major delays that hurt them during COVID So How do you think Envert expands over the years?

Yeah, I think you answered most of that question in that as biopharma becomes a larger and larger part of our revenue portfolio and the sort of the number of customers that we have, the regulatory offerings really end up taking a lot of weight there. There are already great QMS systems out there, the quality management systems out there. There's a lot of other tooling that goes into this. Most of those companies, I think, do a great job, and all we want to do there is just integrate and make sure that, that we stick to what we're good at, which is helping people optimize their processes, and the QMSs stick to what they're good at, which is helping people make sure they're compliant as they do that. The way that we've been thinking about market expansion here is this, uh, we have sort of moved from initially, initially solving a data management problem to a data analytics problem to now a data modeling problem. Meaning, uh, we— the first 2 years of our journey has been letting people do what they already were able to do but a hell of a lot quicker, to now, uh, giving those same users access to nonlinear, uh, methods of analytics, things like off-the-shelf ML models to optimize yields and do machine or model-based DOE and all these other types of things. We spend a lot of time building product in that direction. And as you also alluded to, a lot of what we spend our time thinking about is making cross-org data sharing as simple as possible. That's a lot of A lot of the time that we spend, it's a lot of what our users love about using Invert, making different types of hardware interoperable and making different ORCs interoperable.

I think especially in fermentation where, you know, there are some anecdotes I have where fermenters have these like physical documents, like notebooks they carry around with the master protocols and, you know, you have the fax thing because it's trade secrets, so they don't want to share everything. And so that's another maybe barrier for very few fermenters. And these fermenters are born in the alcohol industry, so maybe they can't patent their products. But on this point of building new features, maybe the cherry on top is AI, artificial intelligence. Over the last year, all these new releases, they've transformed my life in some ways using these products and figuring out my own use cases. And I think we're all other companies trying to figure out how AI can enhance their product offering. For Invert, have there been particular places where this current instantiation of AI, which is like a transformer model LLM, has kind of sped up either product development or maybe enhance the customer experience?

There are two answers to this. The first answer, which I think is the most important one, is that we have been a machine learning company from the beginning, meaning I think in the very first instantiation of Invert, what we truly wanted to build was if we could just go do bioprocess modeling using machine learning as a service, that's what we would do. And once we started talking to customers about that, we discovered that nobody had their data in a square enough and accessible enough format that that was a feasible offering. So we sort of had to build all this data management tooling beforehand. And so that's been very much part of our DNA initially. Post, post post the whole LLM boom that we've seen the last couple of years, as you can imagine, we have what is sort of at this point table stakes features and functionality around using natural language to prompt certain plots or data analytics. You can do things like ANOVAs and regressions and whatnot using just natural language in Invert, and Copilot has access to all the data that you have access to in your instance of Invert.

Yeah, I think that kind of reinforces the experience of a lot of life science companies. The real bottleneck is getting valuable tokens in AI language, right? Structuring data that's valuable and then doing analytics and modeling on it. But there's also biological data and also text data. Yeah, from my experience, the biological data, that's where you have to structure it yourself. But there's a lot of text data out there, notes and stuff that in my business of venture capital, LLMs have been transformative in. Just to think, we read a patent has been pretty easy for me now. And I would imagine for your customers, just be able to like structure notes, like fermentation, all scientific work has a lot of notes.. And I think you kind of said this, there are easier features to help people collaborate and whatnot. But on this point of like, you know, looking forward, like what really excites you about biotech the next few decades, you know, and where do you see Invert playing a role in like making all these discoveries happen and making these like products, these biological products that go beyond therapeutics into synthetic meat, enzymes to help various climate problems, and so much more.

Yeah, I mean, I think in many ways the answer here is fairly obvious in that it truly does seem like we have— we as a civilization have a real shot at replacing the petrochemical supply chain and a lot of the animal agricultural supply chain with bio-produced alternatives. I'm incredibly excited about the prospects of that whole technology. I think synthetic biology sometimes gets a little bit of a bad rep because, you know, as an industry it's been around for a couple of decades now, and you can't really throw a rock and hit one of the winners, meaning those companies haven't sort of universally won. And I think people draw the wrong conclusion from that in general. I think people generally conclude that symbiosis technology isn't feasible. In reality, what I think they're observing is that there are technologies that favor incumbents a lot more than startups. There's sort of a particular set of problems that are a little too sciencey to be startups and a little too engineeringy to be academic problems or academic groups. And so they lend themselves very well to the large industrial players and If you sort of try to enumerate all the bioprocesses that have been scaled in the last 20 or 30 years, you will see the vast majority of them have happened at the big industrial players. And so I actually think that the industry as a whole have made a hell of a lot more progress than we sort of see at the face of it, just because it all happened internally at these large companies. And I believe we'll continue to see that. And I'm also hopeful that we will, that as confidential innovation comes online, a lot of these the sort of expensive bits of the capital needs in building a biocompany or industrial biotech company gets unbundled from that company itself, that we can begin to see some of the startups really win in a big way in the space as well.

That's a great point in terms of unlocking all the innovation that has already occurred and maybe democratizing access to these tools. How do you think Invert, like, are there specific, other than build a great product, like, do you think there are specific things that, like, other than Invert's, like, software product, what, maybe what can you do, or what does the general ecosystem need to do to make that a reality? What, you know, what other tools need to be out there other than software for biomanufacturing? I'm assuming cheaper DNA synthesis. A few other things, but it's a look at the toolkit to fully democratize access to biotech so every small company has a shot to be big. Yeah, what's top of mind for you other than biomanufacturing?

Yeah, I mean, well, if you say other than biomanufacturing, let's start with the biomanufacturing.

Biomanufacturing, that's obvious, that's obvious, but that's Invert's doing that.

So yeah, I mean, I spent a lot of time thinking about the downsides of this industry growing and some of the other trends that you were just talking about. I think while the prospects of much cheaper DNA synthesis are deeply exciting and are going to make for a much more wealthy world in general, that also comes with some very real downsides. If you can— if there's a DNA printer in every high school, and you can download smallpox off the internet and there's a school shooting every day, it's not entirely clear to me that we're not going to end up in a situation where we'll have pathogenic risks that are much worse than they are today and are engineered in a way that they might or might not be today, but definitely are easier to do. I spend a lot of time thinking about that world. I think we have to cover our bases a lot more than we do there now. A lot of the other stuff that I spend my time thinking about is how to make sure that the startups that are coming online right now have the right financing options available to them. There's a large number of companies that are using equity dollars to finance things that really seem like they should be using debt facilities of different kinds that no one currently can figure out how to underwrite because they're just too bespoke to each company. I think as we begin to see the industry proliferate and grow, hopefully we can unbundle more and more of those layers of the supply chain and the value stack and sort of commoditize them and turn them into assets individually that can be underwritten separately so that each individual company doesn't have to vertically integrate from day zero.

Yeah, I think that's where products like Invert are invaluable because it reduces all of the complexity about making something into just API or a piece of software. And that's the dream. You know, I'm very bad, I'm a very bad bench scientist. And I've always, I always dreamt about while working in the lab, I'm like, I stink at this. Why can't we just automate all this? Why can't I just like program all this?

And so that's kind of the dream of most biologists is to get out of the lab Yeah, I mean, this is however where that sort of naivete of people coming from tech entering biotech does tend to show up, which is, oh wait, why can't we just have AWS for X? And as it turns out, the servers are incredibly fungible and you can use them for all sorts of things. And it turns out that bits and bytes encode in exactly the same way whether you're using to represent X or Y, whereas the real physical world does have a lot more variety to it. And it is really hard to create elastic assets that can act in that sort of scale-up, scale-down type way that we all ideally would want for the biotech world to exist in.

Yeah, I think for that kind of vision of the world, it takes a fresh mind. Even me, Trent, as a scientist, I have attachments. I have things I— just customs I acquired just based on habit, and it's really hard to deprogram yourself and unlearn something. And so I think biotech needs more people like you to come in and take a fresh approach. I think the hard part is learning something really specific about the industry. I think you're underselling yourself, just listening to you and talking to you, just like It's probably a lot more work than you're leading on to. And, you know, be grateful that you found some really awesome people to advise you. I think maybe that's a big lesson is, you know, surround yourself with people who know more than you and you learn from them. And that's the thing to be grateful for is the team you build.

Yeah, I think if there's anything to be really grateful for in my job, which is thankfully a lot to be grateful for, it's, and which was an unintended side effect of my job that I wish, I wish that I had thought of beforehand, but I truly didn't, was, is the fact that my days are filled with talking to people who are so much smarter than I am, both on my own team, we've hired the best people out of the coolest biotech companies in the industry, and our customers who are truly just saving the world, or at least attempting to, and hopefully one day will in many different ways. It's an absolute pleasure and frankly the honor of my life to, and my co-founders as well, to be able to spend the contents of our working days engaging with people who are that brilliant, that ambitious, and that generous with the working hours of their lives.

That's a great way to end it. I'm really excited for what Invert does. We'll talk, we'll do another one of these in a few years to get the updates. But Martin, I really appreciate you taking the time. I think a lot of people are gonna take inspiration from this.

Likewise. Thanks, Josh. Really appreciate it.