Okay. We're going to kick it off. I'm Tycho Peterson from the Life Science team. It's my pleasure to have Illumina. Welcome, Ankur.
Yeah. Thanks, Tycho.
Maybe just set the stage here, just some of the coming out of the third quarter, some of the gives and takes, and how you're thinking about trends heading into next year.
Yeah, sure. As a reminder, very pleased with how Q3 overall panned out for us. As a reminder for everyone, we laid out a framework for Illumina a couple years ago in 2024, in terms of returning the company towards high single-digit growth. We've made tremendous progress towards that, despite different changes that happened in the end markets for us. 2026 is a step there, where we are steadily in the mid-single-digit growth over the quarters, as well as in our guide for Q4. Q3 was the right step in that business as well. Some highlights, again, our placements on the instrumentation remain very strong. We've said our clinical business is where a lot of this upside demand has been coming from, especially in the U.S., as well as a little bit outside the U.S. as well on the clinical markets.
The consumables growth has held steady, in line with both the increasing install base, as well as where it can go going forward. We have made a lot of progress in building up our multiomics as well as our BioInsight portfolio, in line with what we have said in 2027, where we continue to expect to see the increased contribution into revenue growth from that. Lastly, equally importantly, made 25% op margin during the quarter, still guiding to 25%-ish in Q4 as well. So for the second half to be at 25% operating margin is also kind of good milestone in our trajectory of getting to the 500 basis point margin expansion in our core businesses over that period of time. All in all, a great quarter.
You mentioned the strong instrument placements, over 80 X placements. I think one of the things people are trying to understand a little bit more is just the relationship between placements, instrument revenues, how you think about kind of the long-term returns on some of these clinical placements that are obviously going to be used more heavily than research. So talk a little bit about some of the gives and takes on these strong instruments that you have seen.
Yeah. We are very pleased with where the instrument placements have been, especially over the last four quarters. I think we mentioned that we have now had more placements in these quarters, four quarters, than we had during the first year of launch, which is pretty unique. A lot of that has to do with both, the two things. One, just the underlying clinical demand being very strong, but also very importantly, we started laying out a framework for what NovaSeq X could do, and the additional throughput and the value our customers can get out of NovaSeq X over an extended period of time, which is allowing our customers to make a lot of long-term bets on the platform, especially around the clinical validation. In terms of consumables growth, well, that is the formula.
More placements over a period of time gives us even more confidence about the increasing consumables growth, the larger the install base. Yeah. You look at our consumables growth, it is being driven by clinical, it is being driven by X. I think a question that always comes up in that, in the formula that you are probably referring to, is the pull-through dynamics there. What everyone has seen and what we have seen over the period of time is that our clinical customers, when they buy boxes, their pace at which they get to peak pull-through usually takes longer than what the research customers typically do. Our clinical customers run through validation, and additionally, a lot of their pull-through also works as the on-market test or the clinical trials, whatever they are adding the capacity for. As that ramps, we see the pull-through pick up accordingly.
That dynamic is playing out as well, because a lot of our above-average placements have largely been on the clinical side.
I think one of the things you've done a good job is highlighting just U.S. clinical consumable growth has been over 20%.
That's right.
Any reason that you don't view that as sustainable, just given what you're seeing in the market today on the U.S. clinical side?
Look, if you look at the U.S. end markets, largely driven by some of our largest customers who have been very public about what their plans and ambitions are. You look at the drivers of growth in that space, the core therapy selection has held up quite well. Then the MRD and the upcoming MCED are layering in additional growth within that market. If I look at what our customers are saying and what their ambitions are, if they continue to execute to those levels of ambition, then that is good for us, right, in terms of our ability to sustain the growth rate in those markets consistent with where our customers are going. That is the dynamic in that space. We also look at, because we talk about U.S. a lot, and given the amount of focus those companies get, our business is international.
One of the dynamics that we also very closely watch is the evidence generation. The clinical adoption historically has always been led by the U.S., right. It is U.S. where the innovation happens first, the at-risk investments happen first, and over a period of time, as either clinical or on-market evidence gets generated, the adoption in rest of the world starts, usually in Europe and some of the wealthier Asian countries, and then more broadly. So both in Europe and the rest of the world, the MRD as well as early detection is still a tailwind in the future. We have not started seeing much of a tailwind in those outer regions out of those two oncology dynamics yet.
How about on the, just the reimbursement front internationally. I know there has been some pressure in parts of Europe, Germany, France. I guess as you think about some of these growing categories and maybe the scrutiny around reimbursement, does that become a headwind internationally on some of these newer categories too?
Yeah, I think we mentioned about some changes in Germany recently. In the grand scheme of things for our size of business, the impact was relatively small coming from the reimbursement there. The way we think about reimbursement overall dynamics is still in the end about clinical utility. The more the evidence gets stronger around the clinical utility of our tests, their ability to say, "We are able to diagnose the disease sooner," that is the primary factor for managing the reimbursement. For us, the puts and takes are always around the right level of reimbursement as well as right level of adoption, right. Higher the adoption, higher the ability of many countries being able to afford these tests has always been part of our business case of saying, "We want to democratize use of NGS.
We want it to be much more broadly adopted across the entire world. Wherever pricing eventually lands, if the adoption is wider, it is good for our business.
I think clinical conversion on the high throughput side of X, it is approaching 80%, 85%. How do you think about just the tail end, the remaining 15%, 20% to convert those remaining clinical customers that have not transitioned by the end of this year?
Yeah. Just as a reminder, we were at, I believe, 78%, Conor? Yeah, 78% total conversion at the end of Q3. We are getting close to that 80%, 85% range here over the next few quarters. We had laid out a dynamic for our customers to say, after once we get to 80%, 85%, there are roughly two types of customers who are left behind. One who have a very profitable test on market, and they are prioritizing their investments in the new tests rather than porting the old test onto the new technology. So in those cases, we think it is going to be a long tail. Most of these customers are right now focused on winning and enabling new indications, and new markets is where their focus is.
The second set of customers are where you have a test on market, but the volumes are so low that it actually does not make sense to make the investment or transition, because you are still kind of making money on that test. We draw examples. Even until last year, we had customers using our HiSeq instrument, which was 10 years old, one generation behind the NovaSeq 6000. So our thesis is once you get to 80%, 85%, what is remaining behind is a really long tail over a period of time. But we will see. We will always have the right technologies for those customers. If they want to choose to move sooner, we will be there to support that as well.
You talked about the importance of clinical evidence generation. Obviously, there's a lot around MRD, some of these emerging categories. How do you model out a market like MCED?
Yeah.
That's new to more consumer-facing.
Yeah.
Maybe it's emerging faster than people had anticipated. How do you plan for that?
Yeah. So, great question. Yes, some of the companies have moved, are moving quickly. The way the MCED market, at least right now, you're seeing is both the multi-cancer detection as well as targeted early detection. Both of those tests are making significant progress. I am actually very pleased with some of the targeted early detection tests as well. One example I use the most is around lung cancer, where there are multiple tests now that are getting to a higher than 70% sensitivity for stage IA kind of lung cancer, which is phenomenal from an early detection perspective and our collective ability to treat that disease sooner or at least interject sooner. So, we think from our perspective, that's a massive opportunity in terms of number of tests that can be done.
We work closely with our customers across MCED and the early detection to enable the workflow for those tests in a manner that if and wherever the MCED reimbursement lands, where again, the balance is between can you see mass adoption at a certain reimburse price, and at that reimburse price, can our customers still have a viable business, right? We work with entire across the spectrum of customers who are going with really targeted small assays with MCED, and then ones who are going really deep with extensive sequencing as well. Across that range, we work with all of them, but you take a three to four or maybe even a five-year view, and you think about several million new tests getting added into the runway is kind of how we think about it.
Okay. But even in a cash pay market like MCED largely is today, you feel like you've got decent visibility?
Oh, yeah.
Yeah.
Oh, yeah.
How about, I mean, there's been a lot of focus on the AI companies trying to move into this area. Are they emerging as a realistic customer base today?
Yeah. So, excellent discussions is where it is. Let me step back a little bit, right? The AI companies have come in today. We started making a bet on data and associated software capabilities a while ago. When we laid out our strategy in 2024, we did start talking about adding the third leg to the stool from a data and services business. We were more focused towards supporting pharmaceutical customers, both in the discovery and any support on the development side of things. That's where we started investing. We started building up the Billion Cell Atlas on our own ahead of the market as well, and made money on it. The AI companies, two years is a I never thought two years would be a long time. But the AI companies are coming in with a new set of discussion within the same space.
The error bars in the discussions we're having with most of them are really large. Okay. The focus on how much data do you really need for a frontier AI model, to make a lockstep change in the understanding of biology, as well as the offerings that they want to make towards pharma and other customers. What do they need themselves or need to pair up with? The error bars on those discussions are really large right now. I can tell you that when those discussions come to fruition, we will be open about it, right? But we're beginning to see a lot of small company neo AI, we call them neo AI companies, who want to set up a laboratory, who want to generate some data. That demand is visible, right?
There are companies who are either buying instruments or asking us to generate some data for them, or are buying into our datasets for their own examples. For our size of the business, that usually turns out to be small dollars. It doesn't move the meter yet. When you think about our framework for 2027, where we've said 1-2 points of extra growth coming from the new products, we think the BioInsight and the AI is going to be one of the bigger contributors in that 1-2 points.
Why is the build it and they will come model the right one for the Billion Cell Atlas? We have seen that before, Celera, others back in the day. I mean, what makes you think pharma this time will want to kind of pay to access data?
Yeah. We very firmly believe, based on what we have seen, that the AI tools are indeed doing pattern recognition and connection much, much better than anything we had even two years ago or three years ago. When you put that multiple sets of data together, their ability to fill in the gaps and connect back to the historical data is significantly stronger than anything that we have seen over the last several years. When we tie that back to the work that our teams do with our own LLMs, for example, you have looked at SpliceAI2, where we have covered another 10 points of genome coverage from a noncoding or variants of unknown significance perspective. Tapping AI on top of that to say we are beginning to understand what this additional set of variants actually do to the disease pathways is powerful. It is really powerful.
Our view is that if we work with our customers closely and build the datasets properly in the manner that it will help them with their process, which is what we did with the Billion Cell Atlas, to say, "We will invest forward, but we will still make money on it based on all the agreements that we have had with pharma." But we work closely with them to say, "What is the right data structure? How will it be useful? Are you seeing the disease pathways?" And we are adjusting along the way. When we delivered the first 250 million or 300 million and the analysis was done, they came back and asked for, let us say, different types of cell types, et cetera. And we are adjusting along the way, much closely with those customers. We see demand.
We see signals coming out of that data, and we see pharma deeply engaged in all of those works.
Yeah. Pharma historically hasn't been a huge customer base for you. I mean, do you see that changing materially in the next couple years?
I anticipate it to be bigger than before. I anticipate we will strategically work much more closely with pharma than we have in the past. We've supported pharma in many different ways, including through our customers. But one of the two big enablers which is allowing us to get even closer to them, one is of course the AI side, but the second is we now have the breadth of the multiomics portfolio, which allows them to characterize their samples in a unified, structured way at a level of depth that really wasn't possible before. We've just launched spatial a few months ago, and the kind of characterization that the spatial product, along with the whole genome and transcriptome is doing, is that people look at it and say, "I wish I could have been doing this for 10 years," kind of thing.
That ability to work directly with them, ability to improve our products more targeted towards what is really required for drug discovery, is we think will engage much more deeply going forward than we have in the past.
Multiomics, I understand you kind of talked to it at kind of a higher level, right, in terms of the overall impact. I guess as we think about what could have the most impact of these technologies in the next year or two, what are you most excited about?
Yeah. So, I apply a BioInsight lens to all multiomics now.
because we have the choice, which we're actively making, as you've seen us do with the single cell space, is to say, "I'm going to sell the kits, but there are areas, specific disease pathways or specific organs where I may actually lead with setting up the data and then monetize it through the data or associated models as well." So, when you apply that BioInsight lens, the combination of proteomics and spatial looks very, very interesting. Similarly, combination of whole genome plus proteomics looks very interesting. And then methylation plays in specific disease pathways, which also looks kind of interesting. So, it's going to be a playoff, generally a playoff whole genome plus something based on different kind of disease pathways.
I think if we were sitting here a year ago, I would have had 10 questions on Roche. It's been a little bit quieter lately. But I guess—
Yeah
—just on the competitive front, how are you thinking about that? Not just them, obviously Ultima, and Element, and others.
Yeah. As a company, we're always paranoid about competition. I absolutely love that when I came into the company. One thing I didn't anticipate but was positively surprised was how much this team thinks about competition today and how much it thinks about where Illumina needs to be or where the technology needs to be in five years from now, both from a workflow perspective as well as our internal cost basis perspective. We live by that every day. Yes, there are multiple new competitors in the market now. Our focus still remains on two things. Do we have the best workflows? Today, increasingly, the customers are beginning to see what the cost of their entire workflow is, and if they make a single sequencer choice, then what does that mean for the whole workflow?
We are still consistently being able to outprice every possible competition out there. Our customers are willing to pay a meaningful premium because the ecosystem holds well, because of the reliability, the quality, and the consistency that we bring. We've started sharing our roadmaps with some of our largest customers. They also know what to expect in two to three years from now from Illumina, so that when they're making their bets today about the tests that will come to market over the next three years or so, they can rely and understand what could their P&L look like, or what could their workflow, the lab capacity look like in that period of time. We're approaching it from a full workflow perspective. We have a very, very strong pipeline of products. We always, every year, look at a 10-year roadmap.
We have pipeline for next five to six years. We cost model to say what should be cost in our hands be in three or five years from now, if in five years the actual test is going to be, let's say, 300x genome coverage. If that becomes the standard in five years from now, will that still be a viable business for our customers? We prepare for those kind of scenarios, and sometimes even more. That's how we think about competition, about where the market should be in five years from now, and that allows us to stay ahead.
Maybe a silly question, but you've laid out a roadmap to get more accurate over time, Q70.
Correct.
I mean, do you also need deeper coverage as you get more accurate?
It will go both ways, right? As you get more accurate, the deeper coverage allows our customers to manage their portfolio to a smaller number of tests. You can do many more things out of a single test if you can go deeper coverage within an economically viable model, right? So, that future, we think, from an industry perspective, could be very meaningful.
Sounds like you are feeling a little bit better on research. I think you called that out on the 3Q call. Just maybe talk what you saw in 3Q, and then obviously post-midterms, I think people generally are feeling a little bit better on academic.
Yeah. Couple points about what we saw. One, you have seen our growth rates, decline rates, slow down in the research space for the last few quarters. Even if you look at the actual dollar size of the business in the research market, that is kind of picked up in the last couple quarters, sequentially now three quarters in a row. Part of it is sequential. Q1 is always the smallest, and then it kind of ramps into Q3, Q4. But we saw a little bit of a steady ramp in that, which kind of gives you a little bit more confidence to say the trend appears to have changed a little bit. But it did not change enough to call a turn in the market itself. You have seen us being very cautious about calling a turn in the market, and that is where we stand today as well.
A lot of that has to do with when we still talk to our customers, we see the pickup in activity, they are talking about, "Oh, we got funding," but then in the same sentence, people still talk about, "But I don't know what to expect going forward," right? No one is saying Our customer conversations are, "Yes, it is getting better," but the caution hasn't gone out in those conversations for them. Until we see that sustained recovery, until we see our customers making slightly longer-term bets around their own lab capacity, or they see a pathway to start building into larger projects that they have a view to fund over a period of time, it will be harder to call a turn in the market there. But the signs overall, meaning as negative as it was throughout 2025, it doesn't sound as negative.
Part of it is probably people have gotten used to it a little bit, but just the dollar amount seems to have improved sequentially.
How are you feeling maybe just on the portfolio question? I mean, there is so much focus on high-end and X. I mean, as we kind of think benchtop, you also have iSeq. Just talk a little bit about how you think about the fleet beyond NovaSeq and investments maybe you are making there that don't get as much attention externally.
Yeah, yeah. Very exciting overall. We will talk about things when we are getting ready to talk about things. But from a portfolio perspective, we have talked about where we are going with X and how much more SBS can contribute from X going forward. We already have a new product for low throughput in the market, right? For the mid-throughput space, where a lot of that business has either gone to high throughput or towards low throughput as the low throughput capacity improved, we think, especially outside the U.S. where a lot of the testing is decentralized, there is likely a future where we will come out with something on the mid-throughput side. We will talk about it when we are ready. You also have heard us talk about workflow-based approaches.
We think especially that side of the market really demands and will be much better served where the actual workflow experience improves dramatically as well, because these are mid-sized labs. They are not the ones who really want to put together a lot of pieces and stitch together the workflow. We will approach that market from a workflow perspective, so it makes it super easy for a lot of our customers with lower throughput volumes to tackle that. We are investing heavily in that space.
Okay. We got one minute left. Just quickly on China, what is your strategy there? I mean, there is the inclusion on the UEL. Just talk a little bit about how you view China going forward.
Yeah. On China, as people know, everyone knows we were put on UEL, and in Q3, at the end of Q2, we kind of lapped the compares of having been put on UEL in the early part of last year. In Q3, our business in China grew 5%. Strategically, we are assembling more in China locally for that market. We are assembling instruments and distributing them. The consumables business appears to have stabilized over that period where our large install base continues to get served from our consumables. Our focus is to keep investing there. Now, in terms of the actual UEL issue, that is a national level discussion, right? And it will get resolved when it gets resolved. Until then, our strategy is to keep supporting our customers.
Great. We are out of time. We will leave it at that. Thank you.
Thank you.