Perfect. All right. Good afternoon, everyone. My name is Robbie Bamberger. I am a senior research analyst here at Baird, covering life sciences and diagnostics. Pleased to be joined by Emily Leproust, who is the Co-Founder and CEO of Twist, along with Adam Laponis, who is Twist's CFO. For those less familiar with Twist, they make DNA to order. They write DNA on a silicon chip rather than in plastic tubes, which lets them synthesize hundreds of thousands of unique sequences at a fraction of the cost. Was able to see this in Wilsonville recently, and it was a very cool setup in their lab there. I guess before we get into specific segments of the company, would you mind just giving a quick snapshot of the company, where we are now after FQ3? What went right in the quarter? Anything you would point out to investors?
Hello, thanks for the invitation and the introduction. Yes, at Twist, we built a technology that miniaturizes the chemistry of making DNA. DNA has four letters, ACGT, and we have a printer that has four color print heads, the A head, the C head, the G, and the T. Then we print on a silicon chip in the square, the size of two iPhones, or the new iPhone Duo. Typically, people can synthesize 96 pieces of DNA. At Twist, because we have miniaturized it, we can make 1 million pieces. So we are a foundry. That means that the strategy for Twist is to bring more volume on the chip. It is kind of like, the concept of a plane.
If you fly an empty plane, it is bad, but once the plane is flying, if you can sell an extra seat, that extra seat is basically almost 100% gross margin. So we have been on a journey to add more and more products, all made from the same chips, that serve lots of applications. If you look at 2021, the product portfolio, compared to 2026, we have added so many products that our SAM grew from $2 billion in 2021 to $7 billion in 2026, and it is going to $13 billion in the future. We are serving a number of markets. We have, of course, a small part in the academic market, and our big focus is around diagnostic tools, enabling especially cancer diagnostic, but also rare disease, as well.
So that is our NGS applications division, as well as our DNA synthesis and protein solution division, where there, we are serving mostly the therapeutics drug discovery. We have had great growth over the last few years, 23% growth in Q4, I think. Q3, sorry. Q3, 23% growth in Q3. We just reported it. Then for Q4, I think our implied growth from the guide was 25%, and more than 20% growth for 2026. As a reminder, our fiscal year ends in September. A big deal for us that we told investors is we are going to be profitable in Q4. So there is 15 days left, in 15 days, we are profitable. We need 30 more days to close the books to prove it, but come hell or high water, our goal is to be adjusted EBITDA break-even in Q4.
That's a great overview. Thank you. Maybe just starting specifically with DSPS, about half the business which makes that synthetic genes to order, revenue grew 39% in FQ3, accelerated from 28% the quarter before. Gene shipped were up 56%, so very fast growth. I guess, what are the major drivers of your growth there? Any particular product line or customer class you think you're taking the most share from there?
Yeah. In DSPS, it starts with D, which is DNA synthesis and protein solutions. That was where we started. We have many flavors of DNA. You can have it as a pool, you can have it as an array, it can be clonal, non-clonal. You don't need to know all the details, but basically, it's like Starbucks. There's a million ways to order your coffee. You go on the website at Twist, and there's many ways to order your DNA. What has happened is, and especially with AI drug discovery, people have been coming to us saying, you know what? I get your DNA, but then I have to make a protein. Can you make the protein for us?
We've been taking the DNA we made, and then we'll apply the automated and high throughput ways to express the DNA into proteins. Again, there are many flavors. Some people, they want a CHO expression, some they want a HEK expression, some they want a cell-free expression. Again, we'll do whatever you want. That has been a fast product group for us. Some customers are saying, actually, I don't want the protein. I just want the data. Can you just analyze this and send us the data? That's what we've been doing, where we take the protein we just made, and then we take the target that the customers want, and we can provide them binding data, expressibility data, developability data, epitope meaning. Whatever data they want, and then we sell data.
What's happening is maybe when we sold the DNA, maybe we'll have gotten $50. When we sold the prep DNA, that would have been $100. When we sold the protein, that would have been $200. Now with data, depending on what they want, it can be $300- $500 of data. As we go up the value chain and serve people further in drug discovery, we are able to extract more value, and that has been a great source of growth for us.
AI-enabled drug discovery has been a huge portion of your potential future growth, and people have been talking about it. Just maybe last quarter, you talked about the shape of that growth, $25 million last year, $50 million in orders 2026, $100 million next year. I guess what gives you confidence to talk about that 2027 order growth a year out, when historically this has been a little bit shorter cycled, more transactional? Maybe just thinking about 2027 AI growth there.
Yeah, maybe I can add some context/clarification.
Yeah.
What we said was in 2025, we had $25 million of orders. At the Investor Day in May, we said that we had visibility that we will be able to have triple-digit percentage growth in 2026. In August, because we have more visibility, we believe that we will be able to do triple percentage growth again in 2027.
Yeah.
That visibility comes from us, of course, talking to our customers, but also studying our cohorts. Our first customers in data, where we sell data, those customers were in the first half of 2025, and we have seen that customer move from a pilot, and then there is a wait time, and then they do a big model building, so that can be multi-million dollars orders. Then there is again a wait time where they are crunching data, and then they come back, and they do routine drug discovery. This gives us, this is a target, the sequence, give me the data. Okay, move on. Target sequence, give me the data. That gets again and again they use us to do AI-driven drug discovery. It is not model building, it is actual discovery. We have seen the movie before from the customers that were the early cohort.
Now, if we get a $100,000 pilot now, we know that in so many months that will turn into millions of dollars of orders. We can add up all of those customers, and the confidence that we have in the number is the confidence that we have in making sure we have the capacity to meet that demand.
Yeah.
Because the worst thing that could happen for us is a customer comes in, says, I have a $2 million data order, and we cannot do it because we have already tied ourself with 50 other customers. We want to always be able to say yes to the next data. You are right that those orders get turned into revenue very quickly.
Yeah.
If you want cell-free data, we turn it in five to seven days. If you want the true data, it's 15-20 days. But 20 days is kind of like the upper limit, so it gets converted quickly. But we do have visibility from having the experience with so many customers over more than 18 months now.
Should we think about that order growth pretty much in line with revenue growth
Yes.
going forward into 2027?
Yes.
Okay.
Yeah.
That's good to know. Thinking about Anthropic, I want to spend a little bit of time on that. Recently, Twist was named an independent evaluator chosen to produce AI-designed proteins for Anthropic at scale. Can you maybe just talk about what Twist is doing for Anthropic there, and then how many designs were made across those 15 targets noted in the press release? Maybe just talk a little bit more about that relationship.
Yeah. We are very careful because we have a very strong NDA with Anthropic, so we have to let them speak. We can comment on what they published. What they published was a pilot, right?
Sure.
It is a taste of what is possible. They picked some targets, and they designed minibinders.
Minibinders are not protein. They designed thousands of minibinders, and they wanted someone to test the binding of those minibinders against those targets. We were able to make the DNA, make the protein, express the protein, and do the SPR data. That blog reports the SPR data. In the SPR data, there is no developability data. Of course, a lot of what we sell is SPR plus developability. The portion that was done with Twist was with clonal DNA expressed to HEK cells with SPR. If you go on our website, you have other choices. If you want, you can do non-clonal DNA expressed with cell-free, and you can use BLI or other data sets. Again, every customer can choose how they want to do their science. We are there to meet them where they are. You are right, what Twist published-
Yeah.
is a small study, but it is very powerful. It shows that it works. Their model was able to pick molecules that bind very strongly to the target. If you wanted to turn that into a drug, you probably will have to at least extend to a VHH. You probably will have to do some developability testing. Hopefully, it is the kind of data that convinces everybody that, yeah, AI-driven drug discovery works. We see it every day. It is better when a customer publishes the data.
Yep. That makes sense. Was that Anthropic work already inside of your expectations for fiscal 2027 orders, or no comment on that?
It is all included.
It was included already as of last guidance?
Yeah.
Therapeutics revenue grew 49%.
Again, this was a small.
Yeah, small portion.
small portion.
Okay. Therapeutics revenue line grew 49% last quarter, and it was down 1% sequentially. Can you just help us understand how lumpy that business is, and then what flips that back to that sequential growth in September quarter?
We had sequential growth in Q3.
Yep.
In Q4, we'll get to 20% growth year-over-year. I don't know if there's anything else you want to-
On the diagnostic side, yes, that's correct. On the therapeutic side.
Therapeutic side. Awesome.
I think was the question. We did see a flat quarter-over-quarter.
Yep
but about 49% year-over-year growth.
That's right.
Obviously, as we talk about the AI drug discovery, the majority of that, almost predominantly all that activity is inside that therapeutics bucket.
Yep.
We are seeing that continue to ramp over time, and a lot of confidence in that. That's great. Would you also mind just talking about the Amazon Bio Discovery relationship and how that's driving therapeutics revenue?
Yeah. We have a deal with Amazon Bio Discovery, where Amazon Bio Discovery, they did the heavy lifting of putting all of the public AI models together in one space. Because if you are a scientist that wants to use AI, there could be a high activation energy to get going. Amazon did the hard work. If you're a researcher, you just go on Amazon, buy some tokens, give them your target, and you're going to be able to leverage, again, all the public AI models to generate some sequences. The compute gets done at Amazon. Then you have to test which one is the one that works the way you want. We are one of the wet lab option to do the synthesis of the DNA, synthesis of the protein expression purification, and then the testing.
For us, we see it as a great way to democratize AI drug discovery and enable new people to get going. For us, it is a great way to get new customers, aggregate volume, so we were made for that kind of work. It is when you are thinking, you hear about the cloud for computers. There is a cloud. Well, what is a cloud? A cloud is a data center, right? The cloud is not in the cloud. It is real brick and mortar that needs real power, real water, real cooling, AC, computer chips inside. Well, we are that cloud for the wet lab of AI. That drug discovery wet lab happens in Portland, Oregon, and in Boston.
Awesome. For pricing, your gene shipped grew 56%, while DSPS revenue grew 39%. So implied revenue per gene has come down a little bit, but how much of that is really expressed genes in the academic promotion versus underlying price? How should we think about ASPs from here or there?
Now you have to be very careful with ASP. Actually, frankly, we do not manage-
Yep.
ASP. Because what is happening is, like you said, there is a lot of gene volume growth, and an oversized portion of that growth is from antibody-
Yep.
work. An antibody is a human gene, an antibody on average is actually small or short. It is about 600 bases, 600 base pairs. So 600 ACGTs. Whereas a typical gene is more like 1,500 bases. We charge per base.
Yeah.
It is a price per base. When the business is growing, the human drug discovery part is growing faster.
Yep.
The proportion of the size of a gene starts to shrink. It looks like the ASP is going down, but actually what is happening is the business is reaping in the human drug discovery. But the rest of the business is going well. The academic market, which is typically non-human, that grew 30% as well. We have launched complex genes. Complex genes tends to be longer, tends to have a higher price per base, $0.12 instead of $0.09. Again, we do not want to be measured on ASP, that is why we actually do not report ASP.
Yeah.
What we want to be measured on is ramping revenue on gross margin dollar expansion, then on profitability.
Yep. Turning over to NGS. Revenue grew 12%, in line with Q2, and you are guiding back to 20% in FQ4. I guess what drove revenue this quarter? Is there a way to think about how much of that re-acceleration is from customer adoption versus existing customers scaling into that commercial production?
No, great question. When you look at our NGS business, the predominant amount of volume and revenue is coming from diagnostics,-
Yep.
oftentimes those are years in the making. You go from a customer in the R&D development phase, to the clinical phase, to the commercial phase, and we see that ramp happening in real time today with a number of our customers. A lot of visibility and a lot of deep relationships there. When we know we see what's coming, we feel really confident in the forecast, and we are seeing those customers continue to ramp. I think we had about 8% sequential growth quarter-on-quarter, and we expect to see that sequential growth continue moving forward.
How durable is that 20% growth as we head into fiscal 2027?
I know I do not want to give away guidance here at this point in the year. But again, we recognize and understand and appreciate that 20% is a marker in terms of growth rates. We are encouraged by the continued momentum of our customers, and we look forward to giving a lot more detail here in the coming weeks.
Maybe just thinking a little bit more about NGS, how should we think about a typical account progression from initial product evaluation, then through validation to commercial production? How long does that journey usually take, or is it not the same across customers?
In NGS, you describe it well. There is an R&D phase where they have an idea for a diagnostic assay. There is an R&D phase that starts with a pilot, where they come to us and to all of our competitors, and they say, those are the regions I want to sequence. Give me a protocol that gets me the data for those regions at, frankly, the lowest price point. Also at a workflow that is as convenient as possible. Because if you are going to automate a workflow to hundreds of thousands of patients a year, you do not want to do it manually. You want it to be nicely automatable. It starts with a pilot. Typically, we win because we have such high quality in our DNA that we lower the cost of sequencing. Then once the pilot is won, it goes into scale-up.
That customer sets up the methods in their lab, on their automation, and we are happy to help them, whichever they choose, Hamilton or PerkinElmer or Tecan. We are going to help them to set up the automation. Then the next step is the validation and verification. That is where the test, is their test actually predicting the disease state that they are trying to do? Can they prove that they can differentiate normal versus disease? That can bring a lot of revenue to us because it can be 10,000 patient, 20,000, 50,000 patient. It can be very big tests. There is a lot of revenue that happens for us during that validation verification. Then the test gets launched commercially. Typically before commercial launch, there is another big bolus of revenue that comes in because they get reagents on the shelf.
And some customers, they choose to order once a month, once every six months, some once a year. Then we wait. Then there's no revenue for a while. Because as they ramp, they are using the test that they have in the regions they have in the fridge. Then once they have used up that region, that's when they come back to reorder. So there's an awkward moment between the validation that has big revenue and the test that has fully ramped, where there's not much revenue from that customer for us. But that's why we don't rely on one customer. We have thousands of customers. Then as to there's a direct relationship, the more tests they sell, the more revenue we get because ultimately, they need to use some Twist reagent for every patient. So that's for assays where every patient gets the same test.
Yep.
For instance, liquid biopsy, rare disease, anything that's tumor-agnostic. Then there are some tests that are tumor-informed, like MRD. And actually, that is a different setup for us because those diagnostic companies, they cannot buy ahead a lot of inventory for tumor-informed because they don't know what the test is going to be. So for there's a more direct relationship. One patient, one test, we get the sequence today. If it's express, we ship tomorrow. If it's not express, we ship in five days, and then we invoice five days later. So there's a direct relationship. As the patient count goes up, our revenue goes up five days later.
Yep. Academic and government, kind of moving to end markets, it was a nice surprise. FQ3 grew 32% versus 3% the prior quarter, driven by U.S. accounts, with several large customers returning. Despite the noise with NIH, what are you guys seeing among this customer set? How much of the ongoing promotion versus durable customer wins is driving that growth?
First of all, we're under-exposed to academia. I think if you are a Twist account manager, and you have to go chase $2,000 in academia or a $2 million deal with an AI company, your mind may be focused on the bigger deals. We are slightly under-exposed. But we offer extremely high value, especially in an environment where funding is shrinking, number one, and two, shrinking is variable. It's hard to know when you're going to get funded. What we offer as a brand is you get high-quality DNA, you get it fast, and you get it best price in the industry. We're lower price but higher quality and fast. It's extremely high value.
Yes, discount or with promotion or with helpful mind share, but I think what won us that nice growth is as some funding got released, people preferentially choose Twist more and more because of the high value that we bring.
Maybe moving on to financials, longer-term guide. You have doubling revenue by 2031, which implies mid-teens compound growth. That 2026 base is now higher than when you guided at your Analyst Day. Does that plan now seem a little conservative based off of your initial 2026 guide? Maybe can you give some high-level understanding between DSPS and NGS through 2031?
I think in general, we are a high-growth company. Once we get to adjusted EBITDA breakeven in Q4, we signal to investor that we'll self-fund the business and reinvest those profit to drive growth. We have multiple opportunities. Today, we talked about AI-led discovery. We talked about diagnostics. We didn't talk about mRNA personalized cancer therapy.
That's another big growth opportunity for us. We didn't talk about animal genotyping. It's not a sexy market, but actually, there's a big growth opportunity for us. All of those applications, they all come from the same chip. We are not over-indexed on any one market, and that's by design. Yeah, I'm glad to hear that it's a conservative plan. We are in it to beat and raise. We've had 14 quarters of revenue growth. We've brought our growth margin from 30% to more than 50%. We signal that we'll get to 60% gross margin, and so we'll keep running the business. But at the end of the day, it's about chip utilization, bringing more volume on the chip. As we get more and more applications that need custom DNA to enable that value, that we're going to have volume growth coming to us for a while.
We have about a minute left. Maybe just capping off, what do you think are the two biggest opportunities for Twist over the next year or so?
The two biggest opportunities.
Oh, three. We'll go with that.
It's AI-led discovery, for sure. It's MRD, and then it's mRNA personalized cancer therapeutics. Those are probably the three big ones. What we like is they have different timescale, and they have different feels. AI-led discovery, we participate in the R&D, the discovery part, with mRNA personalized cancer therapy. There, we participate in the manufacturing. Every patient's who will be manufactured, part of it could be manufactured on the Twist platform. Then with MRD, it's a patient gets a bespoke test. All of those, they require customized DNA at scale.
Awesome. I think that's just about all the time we have, so please join me in thanking Emily and Adam, and the team from Twist Bioscience for coming on. Thank you, guys.
Thanks, Robbie.