Oxford Nanopore Technologies plc (ONT) Earnings Call Transcript & Summary

Hi, everyone. Good morning. I'm Tejas Savant. I'm on the Life Science Tools and Diagnostics team here at Morgan Stanley. Before we begin today's session, for important disclosures, please see the Morgan Stanley research disclosure website at morganstanley.com/researchdisclosures. And if you have any questions, do reach out to your sales rep. So it's my pleasure this morning to host Oxford Nanopore. And speaking on behalf of the company, we have Gordon Sanghera, CEO; and Rosemary Sinclair Dokos, Senior VP of Product and Program Management. So first of all, thank you, guys, both for joining us. Much appreciate it.

Maybe, Gordon, just to take things off, many of your Life Science peers have been seeing a challenging macro that's leading to lengthening sales cycles, a constrained funding environment. How would you describe those dynamics today versus at the start of the year when you already had baked in caution into your initial guidance framework? Could you just describe what you're seeing in the marketplace? And how you've navigated through those conditions in the first half of the year?

Sure. So what we are seeing is a continuation of last year, particularly the second half, we started to see this general slowdown. As I've said before, I think it's not a hangover from the pandemic. It's just a new normal. There is just a lot less money available for government-funded programs. And we did see lengthy sales cycles. But also, we've seen even granted funding takes longer to get going as well. Because whilst they secured a grant, we find that they haven't got the ancillary staff or the additional setups they need, particularly in population scale large programs where you need your limbs for tracking and all these things. So we absolutely feel that pain the rest of the market is feeling. Now what we've worked hard on doing is really looking at the unique value proposition we have with the native DNA with the long-read methylation, structural variation, all of those features real-time near the point of origin. And so we've managed to overlay. We've seen underlying growth as we expected in this new normal, slower growth in LSRT, and we've met that expectation. But we've got these additional applications, which we can talk about later, like Plasmidsaurus, which have really allowed us to continue to grow in a market where our competitors are shrinking.

Got it. Fair enough. I want to start with the recent tech update. Maybe -- I think you highlighted Q25, Q26 sort of accuracy at the moment. Can you just walk us through the journey to get your -- the AI/ML aspect, this new chemistry. What helps -- has helped you improve the accuracy on the platform?

Sure. So it has been a journey, one that our teams have worked really, really hard on. And it is -- our system really is a combination -- it's a confluence of everything STEM from new nanopores, new enzymes, how we manufacture and how we create our flow cells, how we bring all of the different buffers that are involved, what impact those have. And of course, on top of that, we lay an extreme -- sort of an extremely sophisticated layer of ML/AI on top of it as well. So all of those groups work tirelessly together and push really, really hard. The biggest jump that we've had over the last 12 months has been a move from LSTM networks to new transformer-based models. So that's the wonderful thing about the open innovation that we have, is that many times we benefit from shifts that happen -- technology shifts that happen outside of Nanopore, so the transformer models are really coming to fruition. And what we've done is taken that technology, applied it to our platform, and we've gone from a Q21, Q22 to Q26. The next steps, of course, are for the teams to accelerate those algorithms. They are still computationally expensive, but we've made tremendous progress between London Calling, which was only in May. Between London Calling in May and NCM, which was last week, the team have almost already doubled the speed in that short space of time to transform the model. So we're are on a really, really good path to get that higher accuracy productized and productionized.

Got it. And how has the customer conversation change. Guys with the availability of the better accuracy nanopores, the customers essentially see Oxford's accuracy issues as being a thing of the past at the moment? And what is still left to get the platform to Q30 simplex accuracy?

Before Rosemary gets into detail, let me give you a sort of thematic high-level strategic answer here. So we are now 10 years since we've launched our first product, which seems insane, but it is 10 years, 14,000 or 13,000 publications in. What we're seeing is a reflection of that maturity and moving into mainstream, our more applied customers or customers who don't do NGS for a living. What I mean by that is they're not research and translation or they're not de novo sequencing people. They're just seeking an answer to a problem they're trying to solve and they're looking for a solution. And that maturity reflects, I would say, and that's really grown in the last couple of years since our Q20 launch in March a couple of years ago. In terms of digging further and mining more from this, there's lots more to come and Rosemary can talk a bit more about the detail around that.

Yes. So the -- we have demonstrated that you can get to Q30 simplex. So we know the path that we need to walk. Some of it is -- again, we are still on sort of A series for NVIDIA. NVIDIA Systems came out with Hopper and Blackwell. There are also other opportunities with Apple have come out M1, M2 then our M3. So Apple really, really quickly sort of accelerating the game. And that's really important for our smaller sequencers like MinION. So we've got those sort of partnerships for small distributed and large and high throughput. The thing that we've been investing in more recently is looking at the downstream analysis path. So a lot of our machine learning experience and expertise has really been focused on pure signal deconvolution. So when the DNA is in the Nanopore, how do you get the most out of that signal. But then there are other tools that detect variants downstream tools that do polishing, alignment, variant detection. And we haven't -- the team haven't really looked at those for the last 2 to 3 years. And so now we're starting to actively work with the community who develop those tools, start tuning them up more expressively for Nanopore, and we're starting to see enormous gains in the ability to detect variants accurately. And so now it's just another layer of accuracy that we can bring and ultimately, the accuracy that biologists really care about. So whether a read is Q26 or Q30, it's a nice thing to have, but being able to detect a SNP absolutely perfectly, that's what really matters to them. We're being able to take an SV or methylation site, and a lot of those tools are outside the base calling and outside the raw read accuracy, and that's what we're starting to invest in now as well.

Got it. And then on Duplex sequencing, guys, can you just provide feedback from users of that chemistry, what applications do you see being most frequently used today? And how are you thinking about the trade-offs between greater accuracy versus throughput and cost?

Yes. Okay. So Duplex. So what we found is that when we launched Duplex, it was phenomenally valuable for everyone doing T2T assembly, really, really great. But however, Duplex was a trade-off in read length and output. And everything assembly is the longer the read, the better. And so what we're finding is that we're getting superior accuracies, assembly and consensus accuracies out of Simplex at Q26 than we were out of Duplex at Q30. So again, it's kind of sort of talking to that space that I just said, sometimes it's not the raw read accuracy that matters, there are other things that are at play. So the thing that we're really -- so our platform and our team are very, very much concentrated on just getting Simplex to work to Q30 and creating a much, much simpler platform. And we believe we've got very, very strong signals that we can do that, which then removes that complication of having to add -- of having to supply to chemistries and having to get customers to decide when to use one or when to use another. So we're very much sort of pushing Simplex to Q30, despite the fact that Q26 is already phenomenally accurate and customers are getting fantastic results. Something that is very, very cool and Nanopore unique around Duplex is the ability to detect hemimethylation. So that's why there's a methylation on one strand of the DNA, but there isn't methylation on the other strands and we've got the few options that we're working with customers who are interested in hemimethylation on how they can get that without Duplex. So we're really pushing Simplex all the way because it's a 10-minute prep, it's really easy and why wouldn't we want to offer our customers an easy and very, very accurate product all in one go.

So I think Rosemary is underplaying the innovation. The remarkable breakthroughs with machine learning/AI, in particular, in the last couple of years, coupled to the hardware signal to noise improvements which are partly from our enhancement in the chemistry membrane, the Nanopore and the motors, all of those have really allowed us to smash the glass ceiling. So Duplex -- we could get so far, but Duplex took us right to the top. And now we've -- it kind of gives a reflection of the pace of innovation of machine learning/AI as well. We've almost launched and retired it in the space over a very short period of time. So Simplex will be more superior. And with T2T, it's all Simplex, but with ultra long, because T2T is about assembly. But there are some really interesting application areas we have for Duplex like the hemimethylation, which our customers are really excited about because that is biologically significant that there's methylation on one strand and not the other. And again, it's is pushing the frontiers on that multiomic strategy and really enabling our customers to really get to the heart of these understanding the previous position to genetic abnormalities beyond the 4bases.

Got it. Okay. So Gordon, actually, that's a nice segue to my next question. You launched that T2T workflow for POPC customers. How does the Nanopore-only workflow improve upon how T2T has been done previously?

So 2022 front cover Nature Methods was long reads. That work was a consortium of many labs, [indiscernible] including Bionano, PacBio and Nanopore. I don't know if there's any short-read in there as well. And so if we do it in cost rather than number of flow cells and kits because I don't know that, something like $30,000 and -- 3 months, broadly speaking. You can now do that in 2 or 3 days with 3 to 4 flow cells. So you're $3,000 Nanopore only. And that is significant. Now just to be very clear, you don't need T2T genomes on 100% of the population. What you need is local sub-ethnic groupings with local reference genomes. Our precise program in Singapore is really fascinating. There are 4 dominant population groups in there. So they will create 4 T2T ground sets of truth. So these are 3 sets. But then they'll use Simplex to rack through the large cohort, but they can always go back to a ground truth set where they've got some complexities or they didn't quite understand all the quality scoring or the confidence levels in a part of the genome. They're not quite there. They can go back and say, "Well, that's not very interesting. Well, that is quite significant and we want to look at this more." So that is an important and significant breakthrough moment as we move away from genome in a bottle and the handful of reference genomes we have, which are European or North American and so really having customized local population scale. And that tenfold drop in price and 1 month to a couple of days is significant in enabling those local reference genomes on T2T.

Got it. Makes sense. Switching to the P2, I think you talked about over 1,350 devices in the field today. You're, I think, now in early access for the P2i, can you just tell us how the P2I fits into your overall product portfolio? And any early feedback from your customers?

Before you do?

Yes. There you go.

That's the P2 without the I. So you have to add the NVIDIA shoebox on the end, but we bought one along because we can.

Yes, thanks. It's a great question. So we've seen this -- we've built this sort of relationship and model already for the last few years with our MinION and our GridION products, where MinION is really designed to go out and enable customers who perhaps don't have an enormous sort of funding pot that they really want to fantastic and exciting signs. And then when they realize the technology is there, it's working well, but they want a more end-to-end, more lockdown controlled solutions, controlled products, the MinION customers are upgrading to GridION. And it's exactly the same with the P2 solar and the P2i. The solar is there as a highly accessible product that customers can buy, they can plug into machines with their own configuration. They can get that up and running. And it really suits sort of the academic market really, really well. When you start getting into the industrial markets, they want to be told exactly what to do and how to do it. They don't want to be figuring out what computer they should buy and how it should work and how they should download software. They don't want any of that sort of that piece to come with it, and that's where the P2i fits really, really beautifully. The feedback from the early access is really, really positive. I think our customers are getting fantastic results. It's an incredibly stable system. We're being really, really sort of careful about how we do rollout of products, making sure everything is really watertight, really tested to make sure that they land and delight our customers and P2i so far is having really, really great reviews.

Got it. And then the MinION Mk1D, guys, that's -- it's a really cool sort of instrument, hooks up to the new iPad Pro. Where do things stand on the development front there? And how are you thinking about that opportunity? Has your thinking evolved at all versus when you sort of first talked about it?

I'd day, yes, it's going well. So we're broadening the access to the MinION itself. One of the things that the MinION Mk1D was really designed for -- over the years, again, our accuracy has gone through new breakthroughs. And we really want to give every single customer, irrespective of all they've done is buy MinION for a couple of grand. We want to give all of them Q20 data. The early MinION didn't have really a temperature control. It's a simpler instrument. The Mk1D has much, much better temperature control, which you need in order to generate the highest quality of data. So it was due a refresh. The MinION itself as a stand-alone device is now 10 years old. It was a USB 3. No one has that anymore on laptops that you buy. So we've got USB-C. We've really sort of just done some work there. And then the other element that we were dealing with was the Mk1C instrument that we had, the sort of the combined MinION. The algorithms that we've developed, the Mk1C was just not keeping up with it anymore. And so we really wanted to ensure that we were going to develop the best integrated solution for customers who wanted MinION. And pairing it with an iPad, particularly as Apple came out with its New Silicon, made complete sense because you get battery, you get connectivity, you get an amazing user interface. So the Mk1D, the instrument on its own is out with early users now. And so that's this one. And the iPad section of it will be out next year.

Got it. Makes sense. Gordon, I want to ask you on the path to a sub-$200 genome on the PromethION obviously. Can you remind us how much the genome cost today on your platform? And how long you anticipate it will take you to get to that $200 or sub-$200 price point? And what gets you there? And also, I guess, on a related note, can you just speak to the applications that can be unlocked as you sort of move down that path?

So I think -- so right now, we can get on genome on a PromethION flow cell. And our list price is $650 but there is volume-based discounting. Particularly in population scale programs, we have model there. But let's assume it's in that $600. It's a good number to think about. We -- in terms of multiple levers, we're not using -- there are 3,000 channels on the array. We're not using all of them. So there's headroom there. The current motors, the translocation speed, the DNA runs at 450 bases per second. We launched a product at 30 bases per second. The ceiling there is 1,000 bases a second. So there's headroom there as well. And then in addition, whilst we haven't yet got 100% of the active channels, we do see something called blocking. Over time, the pause get clogged up irreversibly. So we can't clean them up. So all those different pieces, we are trialing right now. Some enhancements in all of those levers, and we're kind of moving from 100 GB of flow cell to around 150 GB. These are median, not maximum. And we've seen customers -- early access customers get 200 GB, but 150 GB median. So that's 1.5 genomes. So we're making these breakthroughs in these various selection pieces. So getting that to 2 genomes per flow cell is within reach, particularly with the latest release as we roll that out, and we'll see how much more we can improve that in the field over the next 6 to 12 months. It's in early access now. That gets us into the $300-ish mark. 3 genomes gets us into that $200 piece this is a fully annotated genome with no gaps with methylation, with structural variation and all of that. So we will maintain a premium price. If somebody wants to sell $100 or $50 gap genome of PCR, which is really just a SNP chip the human genome. I don't want to trivialize this. It's been a great technology for 20 years, but its time has come and offering fully annotated, high-definition, full-color genomes in that sort of $300-price range, which we're in touching distance of, I think the gap between sequencing by synthesis and long-read Nanopore in content gets close enough that people will more often than not rather than today in population scale programs, we see 10% to 20% long reads. I think those price points will really increase that number. And with that better understanding of the multiomic contributions that we're getting with predisposition to disease and biological insights in genomics.

Got it. I want to switch to direct RNA seq. That was -- I mean, certainly got our attention back when you talked about it first at London Calling. Can you just talk about -- as you think about the applications that, that can unlock with that capability, give us a flavor for what's the customer feedback been? And how do you see that sort of journey and when do you expect to go live there?

I'll just talk at a sort of macro level on the numbers, and Rosemary can do the science detail, as she does it so much better. But if you look at our -- we started to split our revenues up by customer type, a bit more clarity on who does what. So our underlying revenue, 70% is still the research markets, LSRT. But in there, about 9% is clinical, but 9% is also biopharma. And at Capital Markets Day, October, 1 year or 1.5 years ago, that wasn't -- 1 year ago actually. Certainly, a year ago...

Yes.

That wasn't even on our radar. We thought the biopharma thing was a slow burner. The success and awareness of plasmid sequencing which is pretty common now in domestic U.S., 3 Plasmidsaurus really coming in. And that has resulted in a halo effect and a pull-through into biopharma -- pretty much most big biopharma now evaluating Nanopore both in discovery and in potentially biologics manufacturer as well, QA/QC. And RNA directly is -- this is thematically true for other application areas. But with RNA today, it takes multiple platforms, 8 or 9, so $1 million, 5 or 6 people and a month. Direct RNA on Nanopore on a GridION, it transforms it to 2 days. And then that's the sort of step change. It means a discovery phase of 3 months is reduced to 2 months. And you get a second bite of the cherry when you do the QA/QC when you've made it for identity and integrity. I've probably done Rose a bit on the answer. So I've just gone into auto pilot there, but we've got lots more of innovation to bring. We're only at 200 bases a second. We think we've got great accuracy gains. I don't know if you want to talk a little bit more about the R&D programs on direct RNA?

Yes, there's many. And I think -- so direct RNA, one of the pieces that everyone is incredibly excited about is this biomanufacturing QC for RNA vaccines. So our RNA vaccines are an enormous new opportunity given kind of the outcome of the pandemic. There's a whole other sort of lifeblood for direct RNA, which is people doing transcriptomic research. And that is a very significant proportion of the market today. A lot of people doing bulk RNA and single-cell RNA experiments on platforms -- on short-read platforms. And what they -- what you get out of short-read platforms and RNA is effectively accounting application. So if you're reading RNA on those platforms, how many times do you see a particular gene being expressed versus another gene being expressed. So much like the DNA conversation that we have with our customers, we have exactly the same on RNA. They're missing isoforms. They're missing an enormous amount of biology, when all they do is counting. And also, there are lots of very novel RNA modification that no one understands the function. So what we have on the research end of RNA -- of direct RNA is an enormous amount of discovery for the scientific community to go out, and that's going to drive an enormous amount of publications, which is one of the sort of the key drivers to make sure we're continuously providing a technology that drives new discoveries that creates novel applications and drives demand for our platform. But in terms of something that is quite concrete and quite well defined, the RNA vaccine and the RNA QC is a very, very immediate and fantastic commercial opportunity for the platform, too. So it's nice when you get 2 out -- 2 for 1.

And similar sort of question on protein sequencing, guys. There was an article in, I think it was GenomeWeb yesterday that talked about a bunch of papers by a company called portal biotech around nanopore-based protein sequencing. Can you just give us an update on where your protein sequencing efforts stand today? What are the next steps? And is this a medium-term opportunity for you? Or do you think of it more of a moonshot category situation?

It's definitely short to medium. I'm going to put that out there. We have kept quiet because we do -- we'll talk about it when it's ready. That is one of Clive's golden rules. Red lines that never get crossed, and we will adhere to that. But some -- we've been funding Jeff Nivala who published his work recently that reflects where academia is really making good steps, and we will be talking about protein sequencing. It's not a moonshot anymore.

Got it. Fair enough. I want to switch to the customers, Gordon, a little bit, come at it more from the angle of the clinical customer base, the applied and industrial customers and then the biopharma customers as well. What area do you think you've made the greatest progress in over the last 12 months or so? And then we'll get deeper into each bucket?

I'll do clinical and you can do biopharma. Yes?

Sure.

So basically, on the clinical side, we continue to consolidate our position in 2 key areas. One is rapid insights in acute and clinical care. So we have talked about in the infectious disease arena, looking at pathogens that are drug-resistant, antimicrobial resistant, in ICUs and we are now up to 350 patients in ICU at Guy's and St Thomas', where 45% were on a broad spectrum antibiotic that was ineffective, which can be terminal. So getting from sample-to-answer in 4 hours rather than 4 days in traditional path lab workflow in critical care, that continues to be very important. We're starting to roll that out into other hospital trusts. We've done 60 or 70 samples in one of the children's hospitals, its name now evades me, in London, and they've had similar outcomes. So it's very exciting to see it being replicated elsewhere. On the -- one other acute and critical care one, which I think is really interesting and really shows the benefit of methylation is in real-time interoperative brain tumor classification through methylation. So methylation tumor classification where the surgeon is doing open brain surgery. The section is provided to the lab. In 45 minutes, they can classify the tumor. Then the surgeon -- it guides a surgeon both on surgery and on progressive treatment as well. And whilst that is a workflow that was developed by a handful of academics, Leiden, Nottingham, Germany, I'm going to forget somebody. But there's a group of them, they've kind of created a workflow, and they're using it in screening for information only right now. It's not a fully fledged diagnostic, but it really shows a use case of rapid insights in acute and clinical care that can be transformative to medicine. We are looking at how we can start to turn that into an end-to-end RUO-LDT Clearwave workflow. So -- and in addition, on the clinical side, we have our bioMérieux partnership, where we continue to make good progress on multidrug resistance TB as a reflex test. That will be launched in Q4. We're working with them on isolate sequencing as well. And then Asuragen, who run by Bio-Techne, they're looking at hard to characterize genes in carrier screening. So that test will go out into early access in Q4 as well. So it's all bubbling along and coming together. What's moving really quickly in the last 12 months since Capital Markets Day is a whole biopharma thing and there's a range of application areas.

Yes. So we kind of call it around synthetic biology because synthetic biology is not just actually a critical tool for biomanufacturing and pharma development, but is a critical tool generally in academia and it's really sort of a springboard of creating sort of simple, but powerful tools for things like plasmid sequencing, moving that into viral vector, moving it then into identity testing, contamination testing, RNA vaccine testing. It resonates very, very strongly with biomanufacturing QC, it then begins to resonate with the development groups in those companies who actually find it amazing that they could accelerate their development work, not just their manufacturing -- their manufacturing sign-off process, but also has an enormous sort of reverberation in the academic environment because you can't go to any science lab in any university without coming across people who are cloning into bugs and expressing proteins. It's just a fundamental molecular biology technique. And that's really exciting because it really places things like MinION's straight and first front of mind to any molecular biologist, and it becomes a basic tool for them to be able to do -- for them to do their research and then they go and they explore from there.

Got it. Fair enough. Gordon, on the clinical piece, any color you can share on the turbot or early access for the -- I guess you got ElysION. Is that how you pronounce?

Yes. We name all our internal programs after fish. So don't ask me why. I don't know. Anyway, so ElysION is our Tecan collaboration, sample to answer, not just in the clinical applications, in the biopharma applications as well. They wander in and say, we need that. So again, this is this maturity of customer, reflecting the maturity of the platform, where they just want sample-to-answer. They've got a question, they want the solution and they want to be able to drop their sample or 48 or 96 samples in the titer plate, hit go, and the other side, the answer comes out. That is incredibly high in demand in our list of early access potential customers for ElysION, 2-line GridION in the biopharma space. But the same is true in the clinical space. And over time, as we get that awareness, the next level of adoptees will be ag and food, but they will come later. They're just not the early access types. So the end-to-end workflow sample-to-answer, walk away hands-free, tertiary analysis, all of that is critical in getting rapid uptake in these markets. And hence, we have spent significant time creating a late-stage development team and really pushing our Q-Line formats. Q-Line's done, P24, P2 and the MinION iPad will all come in the next 6 to 12 months as Q-Line products because that's what we need for these applied market. It's an exciting area, really exciting. The sort of 2 pillars of clinical and biopharma are really driving that.

Got it. I'm going to switch to the academic side of things or the research end market, Gordon. A bunch of questions around just government research funding being pressured, both here in the U.S. as well as in Europe. Obviously, in October, you've got the big funding cut for all of us. What are you seeing in your conversations with that customer base?

It is challenging, and there's no doubt about it, right? For every all of us being cut, there's a precise emanating somewhere else. When you look at precise as a flagship for Asia Pac, we hosted a dinner with the leading lights and population scale program across: Thailand, Korea, Vietnam, Singapore were leading it. We did this at the ambassador's house in Singapore, and we had the great and good there. Dame Sue Hill and Genomics England talked about their experiences. So whilst there is some real difficulties with all of us right now, there are hotspots around the world. In Germany, there's -- not in Germany, I should say, in Europe, there's a large population scale program. And so there are hotspots. One has to be agile and nimble to be able to pick up these opportunities. And I will reemphasize T2T is a game changer in opening door for us to have those conversations and have an edge for the long-read winning of those contracts.

Got it. And what -- where is your head at in terms of this push towards multiomics, where it almost feels like there's a view out there that perhaps like sequencing is not enough, and we need to do -- focus more, allocate more budget for the other omics, right? So is there a crowding out effect now? Obviously, Nanopore and long-read offer differentiated value versus short-read, but then you've also got proteomics and single cell and all of those other sort of omics that are making inroads. So how are you thinking about that in terms of the context of the budget?

You're right. There is only a finite amount of cash, and it has to be spread across multiple areas. I think what you will find, and you know what? The sort of academic circles, they swing like pendulums and it's all very trendy to this. It's ultra-proteomic right now, and we've seen this before. I could go back 10 years and talk about microRNA that was going to solve the world's problems. The pendulum will swing into a steady state. But what we are demonstrating and will continue to show is that it's all of the above: structural variation, methylation, copy-number variation and the SNPs and SMBs, and that will drive adoption of Nanopore sequencing. What comes next is proteomics, which is why we're investing R&D and innovation into proteomics. And so the 2 can coexist. And I think money comes from slightly different buckets in terms of funding. And the fact that we can provide all of the multiomics in one sequencing run today, that additional cost and complexity is significant. The fact that the incumbent market leader is now saying -- talking about the fifth base, is right. We think that's great. I think that's some acknowledgment that single-point mutations are not enough anymore. We are well poised and positioned. And with our improvements in accuracy and throughput, therefore, costs coming down, we think we're poised to make a big play there.

Got it. Fair enough. And then on that note, and a bunch of questions on the competitive landscape as well. But before we get there, a bunch of sequencing companies are now focused on single cell even on spatial or systems biology for some of the incumbents. What are your plans there? And how do you view those adjacencies?

Yes, sure. So we've been working with the single-cell workflow for a couple of years now. Again, we offer that additional view of the transcript. Something that customers really enjoy about pairing Oxford Nanopore with someone like 10x, for example, is that our workflow is really simple. So it's a single tube experiment, the sample practice under 3 hours and then they're straight on to sequencing. And the sequencing modality has been really flexible. There are a lot of people right now who send -- who send all of their samples away into a core. They have to wait for weeks and weeks and until that core has had amassed enough samples to switch on a large sequencing run. Whereas with our P2 format, people can be preparing single-cell libraries in their lab, and then running it immediately on a P2 and getting their results straight away. And then it gives them that competitive edge, right? Everyone wants to be first to publish. So our technology is really, really compatible with other single-cell technologies and we work closely with those companies to make sure that the lab workflow, the informatics workflow is really simple and easy to run.

Got it. Switching to the competition, Gordon. One of the concerns, I guess, we hear from investors often is with the launch of the Revio, there's been a significant increase in available long-read capacity on the market. Do you see that as being an issue. I mean not just for PacBio, but also from the point of view of seeding the market in terms of Oxford Nanopore's platforms?

Unfortunately, PacBio and Oxford Nanopore have very low market share. I don't think we've saturated the long-read market by any length. I think there's a lot of capacity there. We have been focused on our own improvements and enhancements, not what Revio is doing. We have one significant and large portions of population scale programs. Our numbers speak for themselves, our half 1 results, and we are still on target to meet our underlying 20% to 30% growth trajectory. That's not what's happening at PacBio. But we are comfortable with what we set out to do this year, the innovation we're putting in and the roadmap we're on. And I think there is -- I would flip it the other way around. As we bring down price points and further enhancements in not just the quality, but the understanding and linkage of multiomics to disease, we will take more share if you just use population scale as a proxy for the rest of the market. There will be growing interest and applications in that space. And plasma sequencing is a very small tip of the iceberg of what can be achieved. We are now retiring Sanger systems out of nowhere in 12 months. We are taking that market rapidly. So there are -- we know there are workflows out there, unique to Nanopore and long-read that we are now looking very closely up. We have significantly boosted the leadership from Nanopore. And I'm not ashamed to say we have got some fantastic Illumina, ex-Illumina, senior commercial people who are really driving our strategy around content. And you'll see more plasmid like application areas evolve, which are Nanopore specific, which will allow us to take market share and push the long-read value proposition into mainstream markets.

Got it. So switching to Illumina, actually, and you mentioned the 5-based genomes. So I'll start there. Not a whole lot of detail to go by yet, but how does -- how do you see the availability of that for free impacting sort of the value prop for Nanopore. Obviously, there's limitations of the short-read genome would still apply. But the fact that it's free, does that mean that they can corner a share of the market there?

Well, I'm not sure what it is yet. I mean they still have to PCR. So there must be somehow speculation on my part. Just add that disclaimer. They must somehow be PCRing and retaining methylation motif. So there's going to be amplification errors we don't have. We actually read the native DNA. If you asked 100 biologists, would they actually like to read the actual thing rather than a copied marker? But it's already in our run. We already account for 98% of the methylome, and we're making -- we're running some projects and programs to really push that value proposition. And as people publish more long-read genomes with full methylation, we think that will be significant. And we'll see how they stack up to that. So we don't know yet. I know I can't give you any detail because -- until we get an idea of what they're doing. But I think it's significant and important that they're recognizing multiomics. The CEO of Illumina said multiomics is really important to quality of the genome. And so they are trying to differentiate themselves. And we see that as a step in our direction, which we welcome.

Got it. And then what about the comprehensive whole genome sequencing launch with the onflow cell library prep and they're committing to, I think, pseudogenes and large structural variants, phasing, et cetera, how does that sort of impact the value proposition for Nanopore?

Too early to say. There's definitely a move in the population scale programs to be heavily automated sample-to-answer push button workflows. We are acutely aware of all of these. And as Rosemary alluded to earlier, the downstream processing and structural variation. I mean this is an interesting one where we have the more mature pipelines. I've said that before. We're usually playing catch-up with short-read, but we already have structural variation, copy-number variation. We have the long-read pipelines in place. And we absolutely, if our customers are listening to this, we understand there's more work to be done, but we are in a leadership position on that, and we will continue to build and consolidate on that. We have 1 or 2 customers now using long-read sequencing in routine clinical settings. So they have clinically verified workflows as well. So they've got to build all of that. They're are doing it in silico again. If you ask somebody would they like to have the actual read? Or would they like it stitched up in silico, they would always pick the actual read. So I think overall, the market leader doing these things raises awareness of these parts of the multiomic play, which I think benefits us because we believe we have a more accurate and more reliable and a better quality offering in that space.

Got it. What about just Illumina's new pricing philosophy, Gordon. We've talked about shifting away from price per GB to a workflow-based pricing model. First of all, just curious as to your thoughts on whether that's a viable approach from customers resonating with its standpoint, so to speak? And then how do you see it changing the dynamic for the long-read side of the market?

So let me just first -- they must have been listening to my presentations. For clinical and applied markets, we've always said it will be a per test. So an applied market customer wants a cartridge that can do this many tests [Technical Difficulty] on a regulated box. That's how the diagnostic industry, that's how biopharma work. So we've always had in mind there is a complete distinctive transition of customer types for test pricing. Now starting to price per workflows and then the interim version of that. So I do -- I think they're right, and we completely think about it in that way. Workflows can be charged at different price points, and I think that's where the market will go. And I think that's a reflection of who wants to do sequencing. 20 years after the first genome was done, you now have got more interest -- especially with the lower growth in the pure-play life science research tools market, everybody is having to look at who else would do genomics if they could. So I think charging for workflows and end-to-end solutions is something that we welcome because we believe with our multiomic outputs, we can really provide a very comprehensive and very informative workflows, and we think there's a premium to pay for those. So we concur in that sort of movement in the market.

Got it. Switching to BGI, end of August, they launched or, I guess, they unveiled CycloneSEQ. I believe it's going to be available later this year. First of all, how do you think sort of that Nanopore platform differs from what you guys are doing, particularly in terms of accuracy and workflow. And then if they do end up commercializing the system presumably with a focus on China. How are you thinking about sort of the competitive dynamics in China evolving for Nanopore?

So we don't really talk much about our IP portfolio, but we have an evergreen IP portfolio. What I mean by that is, it is now 20 years since we licensed the original technology from Oxford University, and most of those patents, the original ones have fallen away or expired. But we went out in 2007, and we set about putting together an academic partnership, which at that time was the 5 big godfathers: [ Harvard ], Santa Cruz, California, Oxford, UMAS, Texas A&M. Today, we have 32 license agreements with academics, mostly Europe, North America. Perfect embodiment of that evergreen activity is the current Nanopore, was discovered in '15-'16. We filed IP jointly with the partnership with Belgium lab. So that pattern runs for 20 years from 2018, and that's in the current systems. Now take that and the fact that we've globally filed patents, including China, because we felt that this would be a step change breakthrough in sensing. We wanted to have a broad and deep IP portfolio. So as and when we see the products available, one can understand how they function. We will protect our company and our products, and we will use our IP portfolio.

Got it. Fair enough. The poster with the AGBT sponsors just came out, Gordon, I think it was yesterday or day before, and Roche is up there as the silver sponsor this year. So had to ask you like, what's the latest cuddle on genia, Stratos, is that something that we see happening in the next 6 months?

Their Capital Markets Day '25 or beyond or their technology day, that's, I mean, I know as much as you do on that. We do know that they take a base, expand it 10x and send it through their Nanopore to do something around clinical and probably a sample to answer box. I think maybe something like that's coming, but again, that's just guessing and joining dots.

Got it. Fair enough. And then on China, one of the question -- well, 2 questions, actually. So PacBio is talking of bringing a benchtop instruments to the market later this year, benchtop long-read and they expect that to be a bit of a sort of a driver of growth for them in the region, specifically in China. How are you thinking about sort of that market opportunity, not just from the point of BGI and potentially sort of PacBio's benchtop, but also more broadly with some of the chatter around the upcoming stimulus, right? There's been sort of divergent commentary. Illumina has talked about not expecting too much NGS benefit from stimulus just yet. But we've heard sort of contradictory views from a couple of other people as well. So where are you shaking out on stimulus and then the PacBio/the dynamics for that for you guys?

So on the stimulus front, as I say, I think the post pandemic funding for research is not a hangover. It's a new normal. So we are also cautious about that. But we do believe, and there is a lot of talk around biosecurity and pandemic preparedness. We believe we are uniquely positioned to have distributed networks, nodes out there for early detection. And there is a biosecurity stimulus, and that's something that's happening in Europe as well. So we think there may be some, and again, our focus is that's a hotspot biosecurity. That's where we can focus our unique value proposition with the small, fast portable systems. With regard to PacBio, did you say that they're launching a desktop this year?

Desktop potentially. I mean, they haven't put a sort of fine point on when exactly, but there's some chance we see it by year-end.

To stimulate China is puzzling to me because we find China as an inversion of the rest of the market. And it's a top-down approach. Our biggest customers that tend to be 2 types. They're either certified service providers. So there is a big market for a long-read and that's done through service providers who are some of our biggest customers. And then there's the sort of -- and maybe this is where the desktop thing might be interesting [indiscernible], there's a lot of panels that they want to run in China. Now whilst we see good growth in China around that, it is challenging since the sort of macroeconomic global geopolitical thing. So we remain cautious, and we don't build too optimistically in the numbers we bake in because there is uncertainty there. And that doesn't look like it's going away anytime soon even.

Got it. And speaking of the macroeconomic, geopolitical thing, Gordon, where are you guys in terms of the workaround that you guys were focused on for that NVIDIA GPU and then some of the trade restrictions?

Yes, it's done. It's complete. So we -- our systems, we try and keep them modular so that you can upgrade either sequencing or compute as required. But yes, for that market, we've created a combined product. And yes, it's shipping and in territory.

And then just a couple of quick numbers ones Gordon, only easy ones. Charlie ensured that I only loft softballs at you are the numbers. So as we think about '25, I know you don't want to guide just yet, but what are some of the pushes and pulls there? And help us sort of contextualize '25 in terms of your path to the medium-term targets that you've laid out and reiterated. I think it's north of the low 60s gross margin, top line growth of over 30% as well and EBITDA breakeven by '27?

Let's do the midterm margin first because that's the one I know most about. So we've said 62% in medium term, and we feel confident about that because there are multiple levers for that. There's automation, there's continuing improvements in our understanding of that is mostly driven on PromethION. The MinION chip is already there. We have some ambitions to reuse the PromethION chip. MinION is now reused and that both speaks to not throwing away hardware into landfill, so we can reuse the electronic component. So all of those different levers leave us confident about where we will be in midterm margin guidance. With growth, we have demonstrated in the first half of this year and the guidance we've given for the remaining part of the year, that we have the -- we've got the -- we understand where we can really go into hotspots like single cell, like the biopharma thing that's happening now, like some of the unique value propositions in clinical and population scale programs and really pick out and continue to build growth. And this is a tough year. We know that from Illumina and PacBio's numbers. And so as we look into the next couple of years, the foundations we're building in clinical, applied industrial and biopharma give us confidence that this has been a foundational year for that. We will be able to build on those, and you'll see that coming through in the next couple of years. So we remain very confident about hitting our greater than 30% year-on-year growth to hit the EBITDA and the breakeven in '27 and '28 respectively.

Got it. Perfect. Just in closing, guys, at London Calling, Clive talked about a number of projects that you guys are revisiting, the full length chromosomes in a single-read, the voltage chip for rapid whole genome. And I think this is also the SmidgION ASIC as well. Obviously, these are still in development stage, but which of these excites you the most? And maybe just -- how are you thinking about ASHG this year? Not specifically as it relates to these projects, but even some of the earlier stuff we talked about on protein sequencing, direct RNA seq, et cetera?

I think ASHG, the big news flow will be TCT by customers [Technical Difficulty], and that's the right place further, and our higher throughput flow cell, which will also help customers to get that price point down from $600, maybe down as low as $400 what we get in the hands of our customers. In terms of long term, I still believe on the platform going. I'm an electrochemist. So the voltage chip gets my vote every time. The capability and the road map to get us to a 1-hour $10 genome is significantly important. There's a lot of hurdles technically. Notwithstanding, we need quantum compute to handle that much data coming off a chip like that. But we remain confident as we were 20 years ago when we had a hypothetical chip and Nanopore sequencing and what we delivered, we will see that. But that is very much in research right now, but I'm excited about that.

Got it. Fair enough. This was a great chat guys. So I appreciate you spending the last 60 minutes or so with us and hope to talk to you both soon.

Thank you.

Thank you.