Twist Bioscience Corporation ($TWST)

All right. Good afternoon, everybody. I'm Luke Sergott. I cover life science tools and diagnostics for Barclays. With me, I have Emily Leproust, CEO of Twist; and Adam Laponis, CFO. Long time we've been doing this. So again, appreciate for making it down in Miami, in March, as we were saying, is not too bad.

But I guess from the jump, let's talk about the resegmentation into the SynBio plus the Protein Solutions and how you guys are an AI winner. And this has kind of like been a big theme across the space of why -- because you're not -- there's not a Claude plug-in for you guys? Like how are you guys the AI winner? And walk through the kind of the resegmentation, how that fits in with that.

Yes. Thanks for having us. Thanks for the question. It's definitely one area where we have a lot of momentum. The DNA Synthesis and Protein Solutions grew 27% in Q1 year-over-year. And so our traditional drug discovery is done in vivo/in vitro. People inoculate the mouse with the target and then you extract the DNA. You screen for the novel antibodies or you take a large library of more than 10 billion antibodies, and then you pan them through phage display. And in both cases, you get 10 to 100 antibodies that you have to characterize. And so the outcome of what we used to sell to customers when they give us a target was up to 100 antibodies that were characterized. We are also serving customers who are doing their own discovery in-house, by selling them DNA and/or protein. We still do that. And last year, our therapeutic drug discovery was about $111 million, growing more than 25%. And so we cannot achieve scale in drug discovery because it's a very fragmented market. A few companies get more than $50 million. So that's the traditional before AI. And then when AI came in, what people did was use compute to come up with thousands of sequences of antibodies. So instead of using in vivo/in vitro, the computer spits out thousands of sequencing -- sequences. But then they need the data, they need the characterization. What is the affinity, the functionality, the epitope binding, the developability, thermostability for those thousands of sequences? And so they came to us because we could make the DNA, we could make the protein, we could characterize. And what they really wanted was speed and scale. And really, we think we are the best in the business where, if you give us thousands of sequences, we'll be able to give you data in 15 days. And it's data for the full menu. And so what we are seeing is customers build their model either in one or several shots, so again, thousands of sequences, and then once they have the model, they go into the drug discovery process where they declare a target and, through their model, come out with a bunch of sequences. We do the building and the testing, give them data. And the benefit of using AI versus in vivo and in vitro is in vivo/in vitro is 6 weeks to get to a hit. And with AI, it's 2 weeks. So you just can do 3 rounds of discovery and engineering, while you could do only 1 round with in vivo and in vitro. So it's been a great momentum for us and it's contributed to our growth.

So when a pharma company or a biotech, like you just have like -- or got the Invenra collaboration. But in the past, it was you'd layer out and you'd get these press releases you just partnered with XYZ biotech and we're going to provide the antibody optimization platform with them. How has that moved further downstream or upstream, I guess it is, for the SynBio side? And now that you're doing this as one whole business where you were doing the DNA, the synthesis piece and now you're doing like the protein optimization or the antibody optimization together, like how was -- like what's the next step here for you guys as you're thinking about this workflow?

Yes. And so our next step for us is we want to have a full menu, right? Again, when we compete with people in the drug discovery space, and there's a lot of companies and they all have a niche. There is a company that they only do mouse and there's another that they only do humanized mouse, and another one, they do a need, that kind of phage display, someone else does yeast display. Everybody has their own little niche. And what we want to do is provide the full menu and meet the customers where they are. And so we are very happy to sell data, but there are still customers, they don't want data. They want to buy DNA from us and they want to do the work in their lab. We are very happy with that. But at the same time, for the new AI drug discovery companies that don't even have a wet lab, we also want to be their partner. So our approach is to not tell people how to discover drugs. Our approach is to, again, meet them where they are, provide whatever science they want to do. We are going to have the tools for them and enable them to go faster with the highest quality and the best price so that their budget can go further. And at some point, you want to talk about Invenra, I don't know if it's the right time, but it fits in the strategy of, in that menu, one area where we were not good -- or actually, I should say, we were equally bad with everybody else, was bispecific. Bispecific, obviously, really hard to make. To make an antibody, you need 2 vectors, right, 2 plasmids put together. For bispecific, it's 4. And if you don't have the right ratios for those 4 coming together, when you express, you get something that's not clean and you have to purify to get the one bispecific that you want. And that makes it really hard to process in high-throughput. Invenra, they've come up with a system, a platform where it's almost auto-purification where it doesn't matter what the ratio is, it comes out clean. And so therefore, now we have what we think is the only platform in high-throughput to express and purify bispecific and, therefore, enable the discovery of bispecific molecule in a high-throughput. And so when you marry that with AI, not only can be used by traditional companies that want bispecific, but for people that we want to build a bispecific model through AI, which is very difficult to do now, it will be possible. So it fits that strategy of just adding all the tools that someone may need such that we are the one-stop shop.

How is -- like from the throughput on the bispecific, how is that being -- that was being done by the pharma company themselves. They have to build out that group?

Yes. So bispecific, the concept of bispecific is simple. In practice, there is more than 100 different flavors of bispecific, but they all have the same issues around expressing them at high purity, which they don't. So any flavor out there, you have to express and then there's a heavy step of purification. So at Twist, we are able to make all the flavors, but we don't have an advantage to make them pure. But with Invenra, we do. And then the other thing with the Invenra platform that's very smart is the engineered linker that they have is very small, is very human-like and is in the clinic. And so you don't take immunogenicity risk with the Invenra platform. And so again, it's -- any customers, they say, "No, we want our own bispecific," we're happy to do that. Like anybody else, we won't really have a high-throughput advantage for [ antibody ] approaches.

Yes. And on the protein side, you guys talked about being more choosy in what projects you're taking as you're thinking about this as a more holistic across the SynBio and the protein synthesis space -- or Protein Solutions. How much of that is because of the capacity? And this is kind of what I was talking about before where you take on almost any project from a biotech. And now like some of these projects are starting to scale and you're bumping up into like ability to deliver, which goes against the DNA of your company, is like you want the fastest turnaround time. So the balance there of how you choose these projects, like profitability, things like that, about as you mature into this market?

Yes. And so our approach, I would say our North Star, is always have customers back of the truck, right? Don't worry about capacity, it's our problem, we've got you. And then the second is that we don't want -- frankly, we don't want to subsidize our customers' R&D. We want to be paid a fair value. And so it's a lot easier to do that now that we're established as a high-quality and fast and a great price. So it's much easier now than it used to be 5 or 6 years ago when we were known for being a DNA-centric company, and then we'll tell customers, "Give us a target, we'll give you a drug." It was harder sell then. Now I think we are recognized as a very high-quality, fast providers of products, DNA and protein, as well as services where, end-to-end, you give us a target, we give you a drug, such that now our job is to make sure that we stay ahead of the demand with capacity. They are part of the process where we have ample demand. For instance, on the DNA side, I think we've said that we have capacity for 3 million genes a year. Last quarter we shipped 271,000 genes, plus we used 50,000 genes in-house to make proteins and characterize and sell data. So you can see that we are still far from -- near -- or far from capacity. There are other product levels where we're probably closer to capacity. It's ramping very fast. Data characterization grew more than 200% last year. And so there, we are adding capacity ahead of the demand. But yes, our goal is we never want to say no if it's -- if the price is right.

Yes. I mean that's commercially violent.

That's commercial violence right there.

Speaking of that, so let's shift gears to the NGS side. I guess, the timing issues from the commercial customers switching from translational to the clinical side, like that seems to be behind us. Talk about how you guys are thinking about that ramp from a test perspective, but also just kind of an overall update on the NGS side from the liquid biopsy piece of where you guys are playing, like where you're seeing a lot of the momentum in the business coming from?

You want to start?

Yes, absolutely. Great to be here. And no, I mean, we're really excited what we're seeing in NGS. I mean a lot of the end markets continue to show significant growth, could be liquid biopsy, MRD and then many of the other diagnostics. We also see longer-term opportunities in new markets, ag being one of those markets that we know is ripe for NGS opportunities. And we're just getting started in that. When we look closer in to 2026, what we said is, in Q1, excluding that 1 customer, we were about 18% growth. And overall the business that customer ramps back in, we see a path to being a 20% growth in our NGS business by Q4. And I'm happy to say that, yes, we're seeing the customer order, things are going great and the relationship is quite strong. So we're feeling very good about where things are going.

Okay. And then as you think about -- you guys just came out with the TrueAmp, right, for the library prep. And this is, I guess, more of an existential question where one of the bear arguments against you guys were, one, SynBio market is not as big as we thought it was. And then two, it's going to be, on the NGS piece as you provide enrichment tools for panels and the market's moving towards genomes or panels to exomes, right? And now the market is moving to genomes, it's like you have a terminal value problem as that market shrinks. So I feel like the TrueAmp is there to get into the library prep for whole genome. But talk about where you're -- is that right? Is there a risk to the NGS business as you go from whole exome to whole genome?

Yes. It reminds me of AGBT in 2009 -- 2009 when someone stood up and said -- I won't name -- I won't say who, but said that panels were a flash in the pan. And so we are clearly, 17 years later, debating whether there's life for panel. And absolutely, there is life and we're very bullish about the growth. It is true that in some applications, for instance, rare disease, that rare disease in the U.S. is going to go from exome to genome as the price goes down. But for rare disease outside of the U.S. where the reimbursement is nowhere near what's available in the U.S., actually people will start doing exome. So even in rare disease ex US, there's great growth. And then for cancer, I would argue that the lower price of whole genome is really good for our business because it's enabling more tumor-informed MRD. Because to do tumor-informed, you have to have a whole gene sequencing upfront. You can do exome, but it's probably more powerful to do genome. And so our view is that the lower cost -- our view today might be the same as it was in 2009, which is as you lower the cost of sequencing for whole genome, you're going to enable more applications of panels. And we are seeing that now in AgBio, where AgBio, they're currently using microarrays, so it's a market that's extremely price-sensitive. And now that the price of the whole genome sequencing is there, Twist for sequencing is going to make -- is going to be cheaper than running a microarray for all the AgBio applications. And so in general, I disagree with the premise that the whole genome sequencing is going to be bad for us. At the same time, what we sell is workflows. And what we realize is that the performance of the enzyme is somewhat limiting the use case of what can be done. And so that's why a couple of years ago, we engineered a best-in-class ligase to this day, which lowers the limited detection for rare mutants, so that ligase is great for increasing sensitivity of the assay. And then we next put our focus on the polymerase, which is what we just launched with TrueAmp. But what we find is that people, they like the PCR-free protocol because it's convenient and there's no bias. But if you could do a PCR, it's just easier because you have more material to work with. And so with the TrueAmp, what we did is we provided the same performance as a PCR-free but with the convenience of a PCR. And we did that by engineering a polymerase that has extremely low bias at high or low GC, and an enzyme parameter that was able to go through repeat regions without stalling or skipping. So we think it's the best of both worlds. Again, PCR-free-like performance, but with the PCR convenience. And it's the kind of thing that we are showing that Twist is not only great at making DNA, but it's also great at engineering enzymes that are key to a particular workflow. And so for whole genome application, we now have what we think a terrific library prep that's going to enable us to go into new markets that we were not in, such as, for instance, academia, for NGS.

Okay. And then with the capturing the workflow and becoming that solutions provider, is that the genesis behind MRD Express as well?

Yes. For MRD Express, the genesis was -- I see it's 5 years ago at AGBT, we launched MRD, we saw that tumor-informed MRD was going to be the winner. And way back then, there was a lot of bear thesis against that as well. Now it's clear that tumor-informed is the way to go. But what we're seeing is that the medical need is going to be for high-sensitivity MRD. When you say "No, you don't have cancer," well, it's better that you really don't have cancer. And if there is going to be a recurrence, you want to find it as early as possible. So high sensitivity is going to be key. So that's our MRD. We get 100 -- 500 probes for the price of 16, or we can get thousands of probes. But what we also heard is, "Yes, it's great to get 10,000 probes in 5 days. I mean it's amazing. But you know, it would be better if you could do it in 1 day." And so that was the genesis of MRD Express, is actually a customer request saying those 5 days actually, those can -- it can limit the window. And so that was the customer demand, and we responded. And now we have MRD Express. And so we -- sometimes we are seeding the market with technology, that's what happened with MRD. But in the case of MRD Express, we also listened to what customers need and we rise to the challenge.

On that tumor-informed versus tumor-naive, but like as you think about where you guys are playing within your liquid biopsy customers, in general, there are a few that still have not adopted from your technology. I guess like, what's the -- when you go out to those big players that still haven't adopted, like what's their pushback? What's the reasoning why they won't come to you guys? Because it seems like you're the only ones with this 1-day turnaround and accuracy and cost and everything else.

It's like you're listening to my sales call review every Friday, it's like we are doing well, but why aren't we doing better? Why don't we have 100.0% of the market? There's inertia, right? Victory is certain. Timing, still to be determined. What we know is when we get into a pilot, head-to-head, we win the panels, we win on the workflow, we win on the enzymes, we win on the supply chain, we win on IVDR. And I think right now, nobody is getting fired for choosing Twist. And we're seeing the sentiment channels. They were our customers before, but they were customers that were saying, "I picked someone else before you are in business and I won't switch." And now we are seeing more and more those people calling us saying that they need to switch because the supply chain confidence that they are getting from our competition is not where it needs to be at the scale, the volume they are now. So we are playing to win. We are playing to win it all. And I think we -- it's a question of time.

Got you. In the last 55 seconds, Adam, let's talk about the margins. The gross margin trends here, towards 55% plus, it's been -- it seems like that's -- you guys are building good momentum there. Targeting breakeven -- EBITDA breakeven at the 4Q. Talk about where you think -- and I know you're not going to put numbers around where you think like margin is going to be. But when you think about the incremental margin opportunity built within the business, where it is now and where you think it could go, like walk us through that.

So look, if you look at the journey over the last 3 years, we've seen our margin growth, over 20 points of margin growth. Obviously, some of that is through the continuous process improvements. But the vast majority of that is the fact that for every dollar of revenue growth, whether it be on the DS/PS -- the DNA Synthesis and Protein Solutions side or it be on the NGS side, we're targeting about 75% to 80% of that dropping through to the gross margin line, the revenue growth dropping through. And we're seeing that pretty consistently as we scale up any of our areas of the business. I think that trend will continue. And so what's going to accelerate or decelerate the pace of the margin expansion is how much incremental effort you put on driving out costs versus introducing new products. And so what we've said is now that we're above the 50% gross margin line, there's no going backwards ever. And we're going to continue to see progress as we move forward because we know we can drive that revenue. We know we can do it. But we want to make sure we're also investing in the capabilities. So the pace at which we grow at is to be determined.

Got you. And then when you're thinking about from a guidance perspective, you guys typically would -- you guide conservatively and you tend to beat on the numbers from a quarter or for a year. But at what point do you start leveling out on that margin piece? Because again, you got to take on -- do it on the investments. And then does the change in kind of the segmentation and the ramp of MRD, the ramp of all the NGS stuff, does that set you up for like significant upside much more so than you ever saw in the past? Like typically, you beat by like a few percentage points here and there because they're longer-term contracts or longer-term projects. Is the business starting to change now where you can get a lot more juice, I guess, on the quarter than what we've seen?

I guess you'll have to wait, is the answer.

All I needed to hear. Amazing.

See you in May.

On the Investor Day, any sneak peek there or?

Look forward to it in May.

Okay. All right. I appreciate it. Thanks again.

Thank you so much.

Yes. This has been really good.

Thanks, Luke.