Okay, I think we are gonna get started here. All right. Welcome back to the 46th annual TD Cowen Healthcare Conference. I'm one of the tools DX analysts here at TD, Brendan Smith. It is my pleasure today to be joined on stage by the Twist management team. To my left is the CEO and co-founder of Twist, Emily Leproust, and to her left is the CFO of Twist, Adam Laponis. Thank you guys for joining. Thank you. Thank you. Awesome. There's a lot to discuss with Twist these days. I know we had a great breakfast this morning, and I know we've caught up a couple times just since the FQ1 print as well. Maybe to help level set the conversation just a little bit, Twist is, you know, marching towards profitability, Adjusted EBITDA breakeven by Q4 of fiscal year 2026, I should say, still on track. Maybe before we dive into the, you know, AI tailwinds here and the NGS segment, just kind of help us understand how we should think about push and pull between really maximizing revenue growth quarter-to-quarter and that really hitting that Adjusted EBITDA breakeven benchmark and really what's kind of North Star for Twist March 2026. No, thank you. That's a great question. That's, that's our North Star. Let's, let's start there. To level set, we have a semiconductor platform to write DNA, the, the strategy that we have is to bring more volume on the chip. Because when we bring more volume on the chip, the, once the fixed costs have absorbed, the variable costs are very low. Then, then the business model works really well. Over the last 13 quarters now, what we've done is ramp revenue, the, 75%-80% of the net new dollars has been dropping to the gross margin line. That means that we are able to bring the gross margin from the low 30s, which was bad to now more than 50%. We had to do that because the math of getting to Adjusted EBITDA breakeven does not math unless you have at least 50% gross margin. We've had shown great discipline on the Adjusted EBITDA on the OpEx line. The losses have been shrinking. In terms of the level and finding that balance of getting to Adjusted EBITDA with the maximum growth potential, I think we've keep doing what we have been doing, which is leverage our NPI engine to launch new products. Those products, even if it looks like the application space may be very different, at the end of the day, it all brings volume on the chip. It doesn't matter if it's a panel for, you know, shrimp genotyping or genes for drug discovery or cancer testing. You know, it all goes to the same chip. That NPI engine has enabled us to, over the years, increase our SAM from $2 billion in 2020 to $7 billion now, and we think it can get to $12 billion in 2030. That NPI engine enables our sales team to sell more to existing customers and as we always do, also find new customers. That's how we ramp that revenue. In terms of the OpEx, you notice that probably in our Q1 we increased a little bit the OpEx. We see an opportunity to maximize the growth rate at which we can exit in Q4 without putting at risk the timing of the Adjusted EBITDA breakeven. We've made some selective and deliberate investment there to help. You know, at the end of the day, it's innovation, execution, commercial violence. Okay, great. I think that's kind of a nice entry point into maybe the first of your two kind of broad stroke segments here. I know on your FQ1 call, and we've talked about this quite a bit since, you're now kind of increasingly flagging AI as this emerging, increasingly potent source of potential upside over the near term. Also as kind of an indicator of really where the broader space is moving, right? This is kind of one of the first indications that we've seen on the impact of kind of traditional wet lab tool spending as a lot of pharma and biotech continue to invest in a lot of these capabilities. Maybe first help us understand how some of the different customer types that you've called out as using Twist products are leveraging some of these AI capabilities, and ultimately what that means for where you meet them in their supply chain. It's touching to a segment that's very important to us. That's the therapeutic segment. Last year we had $111 million in therapeutics growing more than 25%. That's drug discovery, therapeutics discovery. Years ago we put in place a strategy to grow in that segment. We saw it as a big opportunity for the tools we provide. The strategy we've had was to have a full menu. It doesn't matter if you did in vivo, in vitro, or AI, we would be there to meet you where you are. That was really that really paid off last year when we saw the market shift and in front of our eyes, but in a way we also enabled that catalysis of the market shift where we saw big pharma companies, startup, AI startup, drug discovery companies, and even some of tech companies want to leverage AI. For AI, what happened was the number of sequences that they had to test were so big that either they didn't have a dry lab or they don't have a wet lab at all, or the wet lab did not have the capacity to handle the new volume of sequences. Customers where in the past may have bought DNA from us, or they may have bought proteins and done the work themselves, now they just don't have the capacity to do it, and so they needed someone to provide them data, sell data. Therefore, last year we leaned in and built the automated capability, high throughput of not only making the DNA and the protein, but also delivering data. For those data customers, they want a few things. One of the things they want is they want a full menu. Even if they're going to mix and match, for every order, they may pick some test, and then the next order, other test. If you don't have a full menu, you're not really in business, number one. Number two, they want throughput where they, you know, they want thousands, multiple thousands of sequences being processed at the same time. The next one is related to the previous one, they need speed. If you're not fast, then why bother with AI? The whole point of AI is in vivo, in vitro takes six weeks for a campaign. In AI, you need to be able to do it in two weeks. Then the last part, they want, you know, something that's affordable. So there, again, scale helps with that. That's what our customers want. We've done the pilots, we've done then the building of models for them, where it's a big bolus of data at one time, and they need some time to digest it. Then once it's digested, they come back to turn the crank and just do drug discovery. Then they want to update their model, refine it, the where we provide. It almost seems like this moment was built for Twist or Twist was... Twist, it was built for this moment. It's the moment where the GPS and the iPhone find the Uber app, like the killer app, in a good way. Again, the speed, the scale, the full menu, high quality, and the right price point is driving the business for us. I know that size as a therapeutic company we're we are one of the top companies. It's a very fragmented market. Very few companies are above $50 million of revenue. Now at $111 million last year, we are way past it and we want to accelerate through that. It sounds increasingly like, and we've gotten this question not just in the context of Twist, but really across tools, the entire tool sector, right? As pharma, as biotech is just pouring more money into their own AI capabilities, ultimately, what does that mean for wet lab tools, right? Traditional wet lab tools. We, I mean, we had a great conversation yesterday with heads of R&D from a couple pharma companies who pretty unequivocally said that, you know, absolutely, they will be spending more on wet lab tools to help validate some of these models, right? It sounds like, you know, as a lot of your customers are looking at the data that they have, they're also recognizing the data that they don't have, and they are kind of cycling through a lot of new sequences and a lot of potential new structures, and they need answers to a lot more questions in very fast turnaround time, realistically. I think this morning you spoke, and I would love to kind of hear a little bit more about this. Some of the comparisons that some pharma customers themselves have even run against some kind of DNA providers and where Twist kind of shakes out as key kind of points of differentiation in being able to meet this moment when it comes to kind of the AI capabilities. Yeah. We're very, very pleased that one of the big top five pharma presented data at a conference where they compare five providers of data, and they send the same sequences to five different companies. All five of those companies say that they provide data at high throughput in 20 days or less. When the data was shown, from a price point of view, it really was about the same price. We were the lowest. We were the cheapest, but it was very similar. When you looked at the turnaround time, the time it took from sequence to delivering the file, Twist was 17 days. We're within the advertised 20 days. The other four companies, even though they advertise less than 20 days, the fastest one was 41 days, and the slowest one was 57 days. That's really is the big differentiator that we provide is the speed in that case was more than half of the closest competitors. If you compare, you know, 41 or 57 days compared to an in vivo or in vitro, it's actually not much faster. It's not that great. But, you know, at 17, that was smashing. We'll keep going faster. Again, it's not 17 on one sequence. It's not 17 days on one sequence. It's at scale for hundreds or thousands of sequences. It's not an accident. We're able to do that speed, that scale because we make the DNA the fastest, thanks to the silicon chip. Once you're making the protein, we are expert at extreme automation. We can make, you know, on plates hundreds at the same time. We have automated the characterization, so the machines that characterize. Typically, it's a PhD scientist, they work 9-5, and then the machine doesn't work in the weekends or at night, whereas, you know, we have automated it with an operator and it works 24/7 and the quality is better. We are leveraging the full Twist advantage of extreme automation, digital tracking of all the samples in the fab. No human in the loop kind of to make decisions. It shows in the competitive data. Maybe just to help us kind of contextualize what this ultimately means for revenues really near term. I mean, can you give us a sense of kind of the relative contribution just within kind of FQ1 and ultimately just kinda based on the conversations you're having with some of your customers now, I mean, how big of kind of a TAM tailwind could this be for Twist in 2026 but also kind of in perpetuity? Like, is this a five, 10-year kind of underappreciated longer term growth opportunity? How do we kind of get from here to there from today? You know, it's two things, right? We always tell our sales team, "If you meet someone, a customer, and they don't order, go see more people. And if you meet someone and they order, go see more people." That's number one. Just go see more people. Make sure that you flood the zone. And then number two is once you get the order is you know, focusing on the customer experience. Just absolutely delighting the customers. Give them, you know, no reason to go somewhere else. And then you can add to the offering by, you know, doing API integrations and, you know, some scientific support. There's additional things that you can do to really make the relationship really strong. Yeah, it just, making sure that everybody knows that we have it, and then once you have the order, absolutely delighting the customer. Repeat business, and we are seeing it now. Customers, they just turn the crank. Every batch there is, you know, controls and so they can do the llongitudinal tracking of the quality of the data and, yeah, we think it's gonna go. Yeah, it will continue. The models are still immature in some ways where it's hard, you know. It's easier to model VHH, so small pieces of antibodies. The full antibodies are harder, and then the bispecific, forget it. There's room for even more model building in the future. When we talk to customers, we still talk to more traditional drug discoverer that don't believe necessarily in AI, and those people need service as well. Again, our motto, we're not here to tell people how to do drug discovery. We have to say, "What do you want to do? We'll help you do it." We meet them where they are. Then maybe just one more on this part of the business today. Help us kind of understand, you guys recently announced the licensing deal with Invenra, right? How does this kind of fit squarely into where you are seeing the evolution of this particular business segment moving? And should we expect kind of more deals comparable to that or within that part of the ecosystem moving forward? Invenra, it's a license for bispecific technology. As you probably know, an antibody, you know, is-- looks like a Y, and the two branches bind to the same thing. The two branch are identical. There are formats of antibodies where the two branch are binding to two different targets. In practice, it's really hard to make pure bispecific antibody. Just you have to get the ratio right between the two, the two arms and, you know, there are 100 different formats of bispecific. You get purity issues. It's hard to purify, so it's hard to do in high throughput. Invenra, they developed a technology where it's almost auto-purification, where, you know, you get a pure bispecific easily by, almost by default. You know, so easy even I can do it, right? That means that you can automate easily. We think that that will open the door for building AI or models for bispecific AI discovery. You need an LLM, you need the L, the large in LLM datasets. Now with current bispecific, it's very hard to do, but with the Invenra technology, we think that it will be easy. It's an opportunity for us to open up the market. We used to do bispecific, the old format, we are doing them equally poorly as everybody else. And now we have a chance to really. We're completing the menu of tools. In terms of are we gonna do more, you know, for us, it's a high bar to do something like this. We don't do it a lot. Again, we are very focused on our Adjusted EBITDA breakeven goal. Again, it's not impossible, but it's a very hard, high bar. Yeah. Okay. All right. I wanted to make sure we get good ample time to also unpack NGS part of the business too. I think there's kind of an increasingly loud cohort of Twist bulls who are arguing that if you know, you're long MRD as a concept and as a growth vector within the space, realistically it makes sense to be long Twist. I guess, I mean, help us understand what makes Twist an especially attractive partner within kind of the MRD space. Realistically, you know, what are the important qualities that somebody, either existing players or emerging players have to come to you for if they really wanna kind of own the future within that growth segment? We agree. First of all, we agree with them. In cancer diagnostic in general, at the end of the day, it's all about sensitivity. Can you get highly accurate answers? Whether it's screening or MRD, if the test says you don't have cancer, you really don't have cancer. If the test says you do have cancer, you really do have cancer. That high sensitivity is where we are seeing the pool of the market, especially for MRD, that high sensitivity is going to be key. It's being validated more and more. I think there is some data that shows that with high sensitivity, you can find cancer many months before other approaches. Our view is that to get to that high sensitivity, you just need more probes. That it's a numbers of a large number of probes. We provided an MRD tumor-informed tool set where you get 500 probes for the same price as 16 if you use an IDT type of probe. If you want more than 500, if you want 1,000 to up to 10,000, it just costs a bit more. At that point it becomes a curve of... On the x-axis, it's price. The more you spend, on the y-axis, you get more sensitivity, and then you choose where you want to be on that curve. The majority of the high sensitivity tests out there are leveraging that approach for tumor-informed MRD. Now it's a question of time until those gets validated and then commercialized and then reimbursed. We see a number of customers, their revenue ramp is pretty significant. You know, last year, MRD was more than, you know, was more than 200% growth for us. Still low as a small number. We can see it accelerating and some of the volumes that those customers are forecasting are very healthy and do not scare us at all from a production point of view because we have the silicon chip that scales really well. It's just going to be more volume on the chips. We provide those probes at a very good price to our customers and still are able to get gross margin. It's gonna be a win-win-win. A win for us, a win for our customers, and a win for the patients. Okay. Great. I mean, when we think about trying to model out kind of relative MRD contribution to the NGS business over let's just say kind of three years, how should we kind of I mean, you're talking about ramping up of volumes over this timeframe, like, what are some of the potential kind of inflection points in this ramp, again, over the next thre years as we think about the relative contribution to just your overall revenue mix? Yeah, no, thanks for the question, Brendan. It's an exciting time in MRD. I think that the outcome, as Emily said, is pretty certain that if you want higher sensitivity, being on the Twist platform makes sense. The question is speed and how quickly do you get there. We said in our guidance for this year is one to two points of the overall Twist growth we expect to come from MRD. We expect that ramp more heavily weighted towards the back half of the year. That being said, the bigger question I think on many folks' mind is, you know, we're currently not factored into the major test that's out there. If you look at where everything is, there's 90%+ market share in one potential customer. How does that evolve over time? I think, the biology and the sensitivity will play out, and, we're here to support the business however it is. I think that inflection point of how quickly does sensitivity get adopted is gonna determine how quickly, this continues to ramp. Yeah. Fair enough. I mean. I guess when we kind of think about, you know, growth of the MRD space broadly, and to your point, kind of the relative contribution of existing and kind of, and, you know, emerging players over the next couple of years. Give us a sense of like what you can tell us about the kind of% of COGS for a given MRD assay, and like some of the important considerations as we kinda try to back calculate a little bit, admittedly, you know, back of the envelope kind of thoughts there, but how we should kinda think about relative breakdown of that market for you guys. In the past, for more catalog assays, they are custom, in the sense that each customers have their own, but for a customer, you know, they research biomarkers, and then they finalize a design, and that's the design that they do the validation/verification. That's a catalog just for them. In those cases, we try to get about, you know, 10% of the COGS. For bespoke MRD, it's a little bit different, in the sense of, you have to buy upfront, 12 tests. Every patient, the 12 tests are purchased as a unit at the beginning. That means that the revenue potential is actually significantly higher. You can... You're thinking of, you know, hundreds of dollars, for one patient. Those 12 tests may or may not be used depending on the trajectory of the patient. Of course, we hope that they will be used, but it's typically, you know, every three months for five years plus time 0 and times 3 months, that's how you get to the 12. What that means is that for us, not only the revenue potential per patient is higher and then the dynamics of ordering also can change. In the JPMorgan deck, we for NGS, we showed four different customers, the order pattern, the revenue pattern. If you go back to that in the top right, there's a cancer diagnostic customers. You can see the quarter-to-quarter, the revenue goes up and down and up and down, even though the volume for that customers, you know, when they report publicly is just ramping monotonously. The up and down comes from how they want to manage their inventory, even on purpose to manage the balance sheet or just being surprised with volume. We see ups and downs quarter-to-quarter for catalog test. For bespoke, it's not like that. It's there's a patient, we ship in days, boom, it gets we book revenue. With MRD Express, it will be even faster. It could be 24 hours between receiving the order and booking the revenue. We think that, it will be much more monotonally in a good way, for bespoke MRD. There's not those quarter to quarter swings, based on inventory management decisions. Right. I think kinda related to this, it's a question we get fairly often, and I'm sure you all do too, but when it kinda comes to the question of whole genome sequencing, right? Obviously, you guys have a more recent product that's kind of addressing some of this as well. Maybe kind of speak to how you envision that aspect of the MRD market kind of evolving moving forward, and then how Twist is ultimately thinking to be able to compete irrespective of how some of it ultimately shakes down. I mean, we think whole genome sequencing is gonna be a big part of the MRD testing upfront. I think it makes sense to either use a really, really big panel or to use, you know, whole genome sequencing. For the more routine, you know, every few months test, you have to sequence so deeply that, you know, it's just unrealistic, I think, that it's gonna be done by whole genome sequencing. If you absolutely want to do it by whole genome sequencing, you definitely want to use our Library Prep Kit because it provides the biggest sensitivity, the least bias, at a great workflow and great price. I think for applications like cancer, when you're looking for a needle in a haystack, where the allele frequency of the mutation you try to detect is just so low, I think you're gonna need panels to do it in a way that's cost-effective. Okay. All right. Great. Maybe just in kind of the last minute and change here, you know, zooming out back to kind of bigger picture, right? We, we talked about, you know, the AI tailwinds to the more traditional synbio part of the business. We've talked about some of the growth drivers for NGS now. You know, I think we've kind of unpacked where some of the more near term opportunity for upside to 2026 numbers could come from. You know, help us contextualize where you stand today, versus like what does the Twist five years from now ultimately look like in terms of, you know, the revenue split between your two streams and what you think could realistically kind of surprise between now and then? Yeah. I I mean, we think about it in two ways. One is a SAM extension. We mentioned that right now the SAM is $7 billion, and we think that we can expand it by new product introductions to $12 billion or more. That's number one. The second aspect, axis is, you know, just world domination, is we want to be number one in a world where there's no number two. We tell the sales people all the time, until we have 100.000% of the market, we're not gonna happy. We're gonna absolutely lean in. It's great to have great product, but the commercialization and making sure that it gets tested in all the hands, is a muscle that, you know, we are really good at, and we'll keep practicing those two things. Okay. I think with that, we are at time. Thank you all for paying attention. It's always great to see you guys. Thanks for joining. More to come today. Thank you so much. Thanks, Brandon.