Great! Kick things off here on the life sciences side, on the 2023 conference. Happy to be hosting NanoString, a company that's been here, presenting for, for several years at the conference. Today, we have Joe Beechem with us. He's the Chief Scientific Officer and SVP of R&D, and then Doug Farrell, who's also VP of Investor Relations. Joe, thanks. Joe and Doug, thanks for coming today. I'm John Sourbeer, the US Life Sciences Analyst at UBS. You know, Joe, maybe just start things off with the CosMx launch. You know, backlog and placements there have been pretty impressive. Can you just talk about, you know, when you talk to customers, you know, where do you see the CosMx winning versus some of the other platforms out there with some of the recent launches? Yeah, I mean, CosMx been really successful. We, from the beginning, when we designed this instrument, we wanted to be able to do. Yeah, we wanted to do spatial whole transcriptomes. That's the way the instrument was designed. We've always been the plex leader, and that's really one of the key aspects of how our instrument's been so successful, is that, yeah, leading in plex was a big part of what our platform is about. You know, the other aspect that's really key is that we've developed technology for bio-archived biobank samples, say, ever since nCounter was started, 20 years ago, believe it or not. We're best in class at working with archival, hospital-grade FFPE samples. I think that combination of highest plex, but you can still do that data on clinical samples that a hospital may have, that combination really resonates with customers. Yeah, that's great. You know, there's a lot, I think, there we, we can dig into. Maybe just on, like, a even higher level, you know, what, what does spatial give you that bulk does not? You know, spatial is a real revolution. I think, I think, you know, a lot of us have been in the business a while. I remember I've been doing this for 37 years now, and we saw next-gen sequencing come along, and that was a real revolution, and I participated in that one. You saw single-cell RNA-seq come along, which was great. But when I was exposed to single-cell RNA-seq, I realized there's a missing element. The missing element is: who wants a parts list of how Mother Nature put you together? You don't want a parts list. You want to know how all the parts fit together to function, and spatial provides that extra level of information. No one really wants a parts list. Everybody wants to know how you work, and spatial technologies allow you to put those two things together and say, "Okay, this is how your organs work, or your brain works," or, "This is why you get this particular disease." A parts list doesn't tell you that answer. No, that, that's great. You know, maybe even down that, that path, or even broader path, you know, what kind of applications could spatial be used for that maybe are even being used with, let's say, a single cell today? I guess, how do you see spatial coming in and maybe, you know, changing the entire market? Yeah, once you get to a particular level of plex, you know, last year I introduced and started showing all the data at 6,000 plex imaging, people came up and started saying 6K every day or whatever. I mean, they really love that plex. When you get to that level of plex, you realize, why am I doing a dissociative single cell experiment with my tissue? Then you're throwing away all the spatial information, and you're trying to get it back with a high plex single-cell RNA-seq. When you're doing 6K at spatially resolved, you really gotta scratch your head, "Why am I doing a dissociative, taking this tissue with all the beautiful architecture that's there and dissociating it?" The fields that really get revolutionized by that, you know, you start out in cancer immunology. Cancer, and cancer is all about immunology now, it's immuno-oncology, but also gets you into infectious disease, autoimmune disease, inflammatory bowel disease, and all those key immunological-based diseases. A real upcoming revolution that we're seeing now is that people want to understand the human brain and neurodegenerative disease, dementia. There's a big push in the funding area on the making the brain atlas, the human brain atlas, not the little, tiny mice brain atlases. We're seeing a huge takeoff on what we're doing in the human brain. We're making some special tools for the human brain, because when you're trying to understand how the human brain works in neurosciences, you gotta have both protein and RNA when you do this high plex spatial measurement, and we're seeing a lot of development. Cancer immunology, infectious disease, and neurosciences, those are all really taking off. Yeah, that's great. You know, you kind of touched on it some there with the human brain. This kind of leads to my next question, but with the 6,000-plex, maybe you could just walk. How would that alter study design, you know, maybe even by extension of throughput or consumables? How do you think that, that plays through there? 'Cause it's really amazing when you start looking at the raw numbers. You know, every time I push a button on the CosMx Spatial Molecular Imager, yeah, seven days later, I'm gonna have 2 million single cells done and 2 billion, sometimes 5 billion, 6 billion total transcripts. That's just a mind-boggling number. I'm just pushing go, and I'm not doing anything else, and I end up with multiple millions of single cells. Every time I push a button, I'll generate about 10% of the entire database that's the Human Cell Atlas, which is a single-cell thing that's been ongoing for almost a, almost a decade. That changes things. When you're at that level, I'm not doing single cells, well, one at a time anymore, I'm doing millions of them at a time, and in a hands-off way. I'm not doing anything but just push and go on that. I think that's when you get up to 6K, it changes that experimental design. Yeah, that's a really interesting, quite a bit of data out there. Then I guess, you know, when you think in terms of balancing plex and throughput, you know, where do you think this settles in, in the long term? I mean, I guess there's, there's some natural limitations there on throughput with such a high level of plex. It's a really good, because I've been doing this and been talking to the outside world about this, because the beauty of when you go up in plex with a barcoding technology, like what's in the Spatial Molecular Imager, plex goes up exponentially and the turnaround time goes up linearly. That's an important thing to remember. When I went from 1,000 plex to 6,000 plex, 6x in genome coverage, it only goes up in turnaround times 1.68. That's how many cycles it takes to complete that thing. I mean, I was literally... We were live at AGBT. I was giving a talk there, and one of the customers was right there. I'm talking about 6K. Somebody asked me about throughput, and I said, I just looked at my customer, this Miranda Orr, this brain researcher. I said, "You could have 6x the genomic content, but you got to wait 1.68 times more. Which do you do?" She said, "Oh, it's a no-brainer. I'll take the 6x the content." You find that's what I find in general. People will take the content and then find imaginative ways of getting multiple samples on a flow cell. Our instrument runs with 4 flow cells, 12 square centimeters of real estate. What you find is that people take the higher plex and then get imaginative of how many samples they put onto their flow cells. Things like, tissue microarrays, you can just do large cohorts of those. Anyway, in general, people take the plex and pay to get the geometric increase and will pay the linear, linear increase in turnaround time, almost invariably. I guess, John, I'd add to that, too. It's not unlike the early days of next-gen sequencing, where to do big projects, you had multiple instruments, right? If you want to do tons of samples, you'll need more than a single instrument to do that, and I think spatial is going to be similar. For big projects, it's easy to scale up that throughput by adding capacity. Yeah. That's great. You know, I guess. I still work on throughput, obviously. That's an important thing. In 2024, we'll be increasing the throughput by 2x from the pure chemistry perspective. It's really not the dominant term right now. Well, you basically just took my, my next question there. Oh. In addition to maybe throughput next year, you know, CosMx just launched, but how should we think about maybe other areas for future development there on technologies on the platform? Yeah. For us, you know, we're actually sort of what I call, you know, realizing what the original design of CosMx was all about, and that is spatial molecular imaging at the whole transcriptome level. That's just an unprecedented thing. I think a lot of people thought it was never going to go there. It absolutely is going to go there. Come to AGBT in 2024, and if you want to see what an image of a whole transcriptome looks like, you'll be able to see that, not just from me, but from multiple, multiple customers out there, too. That's one of the areas they're really pushing. You know, the other one that's important is this multi-omic, the fact that protein and RNA, you're in particular disease types, especially in the brain. When you're doing dementia in the brain, you're trying to understand how the brain works. We basically get messed up in our brain by a whole bunch of protein tangles and all this stuff that happens in our brains as we get older. You have to image, high-plex image at the protein level to know exactly where those disease areas are. Then you need the unlimited plex on the RNA because we actually don't understand how the brain works. We don't understand all the complications associated with that, so you have to just measure everything. Yeah, that's really where we're putting our emphasis on high plex protein, unlimited, and right now, we're at the highest in protein. Also, we'll do up to a spatially resolved 120 plex protein. Then at the same time, we'll be doing whole transcriptome RNA on exactly the same 5-micron biobank sample. One sample, all that information coming out. If you actually look at it at an information and data level, it's like, it's the data density is very similar to a NovaSeq 6000. You're doing 5, 6 billion transcripts, which is a NovaSeq 6000 run, but now you're doing them in space, and that's a big difference. Yeah, maybe, John, to add to that, too. The roadmap is not unlike with genomics, right? We launched with a 100 plex RNA. Year two was 2,000 plex RNA. Year three, we went to whole transcriptome at 20,000 plex. The, the chemistry scales very uniquely. That's great, and look forward to seeing that presentation next year in Florida at AGBT. You know, I guess plex seems to be a lot of focus, but another area just that comes up in conversation is, you know, accuracy with the CosMx. You know, what has been the feedback there? You know, how do you think the sensitivity, specificity compares to some of the other platforms? Yes, it's great. It's something that not everybody talks about. Sensitivity is pretty much an even game. I, I don't have any worries about the sensitivity term. People talk about the errors in, in molecular imaging, and they miss the big one. The dominant error is everywhere, is associated with segmentation. The dominant error in single spatial molecular imaging, is putting the RNAs into the right bag, and that involves the cell segmentation. That's the dominant term. You can talk about all these other errors, and they are so far down decimal points on the error. I'm not exactly sure why some of these groups are spending so much time talking about it, because they're really irrelevant. The one that matters most is how well do you draw the plasma membrane, the nuclear membrane, and the cell body? That's why we've made the emphasis on segmentation, because that's the dominant error term. Yeah, I just... It's worth remembering, if you guys remember, talk about it. Ask people about how their segmentation errors, that's the dominant term, 'cause those things are still, and probably the best segmentation, which is probably what we do now, you still get segmentation errors are on the order of 1%-5%. That's as good as anybody in the entire world is doing. Then somebody talks about some kind of false discovery rate number or whatever, that's so far down in the noise, it's really not relevant. Segmentation, and that's why we are, I think, best in class. We are best in class at segmentation. Sensitivity, we have the sensitivity. Segmentation is the thing that people should be paying attention to. That's still a work in progress. The whole world's sort of working on that still. No, that, that's really helpful color there, too. You know, another area with the CosMx is, you know, the combination of both RNA and, and protein. I guess, how is that helpful in for your users of your platform? Yeah, it's just really absolutely essential. When you're getting into trying to understand disease and disease states, how to treat diseases, as you all know, they get treated by drugs that actually act at the protein level. It's just, you know, when you're really trying to understand disease, healthy and diseased tissue, the protein layer of information is just crucial. One of the things the protein layer of information gives you is that you see tissue architecture at a level that RNA just doesn't do. I love all the little 6 billion little RNA dots, but when I want to know exactly where the epithelial layers are, where tumors are coming in to, and the immune system invading the tumors, nothing that highplex protein gives you that image better than anything, by far. If I want to understand the mechanism, that's what's happening when those different cells come together, that's when I need the highplex RNA. That's why these things work together so synergistically. The protein will give you an image of the tissue that's unparalleled. It'll also image a tissue and say, "This is where a disease state is." That's what happens in the human brain. We image these things called phospho-taus that tell you exactly where a tangle is in your brain, 'cause your RNA is not gonna really tell you that. Then the RNA tells you all the downstream effects of having, you know, the anomalous protein where it shouldn't be. That's why, from the very beginning, we knew that they were both essential, and that's why they were always designed that way. Same slide, both unlimited plex on the RNA and up to hundreds on the protein on the same slide. You know, that's, you know, that's, that's really helpful to frame, you know, how you see the users using it. You know, it's early days instrument out there, but maybe even currently or just, you know, longer term, how do you see that mix of consumables across maybe both protein and RNA, and how that might, might evolve from here? Yeah, Doug will help me out on this one. The, you know, the R, you know, the R, you know, just, no, it's just the way life science has been. You know, the RNA, it comes from... Sequencing really pushed the RNA technology and the DNA technology. RNA is still king. There's no doubt on pure discovery and all those things. I think as we find spatial technologies mature, the proteins are gonna be more and more important. I really think it's gonna be lead with the highplex RNA, because that's where the world of single-cell RNA-seq is all coming over. They basically are sort of addicted to RNA, which is great because there's a lot of information in it. I think the protein is sort of, is where it's where you go once you get the RNA data, and you say, "Okay, now I need that protein layer of information." I see those things working that way. Yeah, maybe just from the kind of total addressable market, about 70% of the TAM is RNA and about 30% protein. That's the current, current mix that we see. Obviously, there's applications like pathology and things where that may be different, but overall, about 70-30, with RNA being the biggest piece. Thanks, Doug, and maybe this is a, a good place to ask, you know, with the litigation with 10x on some of the RNA, just any changes in the customer conversations there, how that might have changed your, your backlog, your sales funnel? Any- Yeah. Anything you can provide on that? Yeah, I think initially, it slowed the sales cycle. There was some additional handholding and things that had to be done, but I think at the end of the day, customers realized that, you know, the best technology to advance science is, is what usually wins. You know, we're, we're confident that CosMx provides that to people. I would say there it was noisy. You know, back in May, there was definitely some, you know, kind of friction in the sales cycle that wasn't there previously, but that has calmed down. I think at the end of the day, customers realized that they want to pick the technology that's gonna allow them to advance the science most rapidly. Appreciate that. I guess, you know, maybe, Joe, back to you as well, just, you know, higher level here on CosMx, maybe to, to wrap up, and you start talking about some of the other products, maybe the GeoMx. You know, what, what type of customers are you seeing the most demand, I guess, on, on CosMx? You know, where, where is the new, more near-term opportunity, you know, on those new NanoString customers? Are they coming from CROs, pharma, any, any academic, any details you can kind of share on that? The one thing you find is that these are new customers coming into NanoString. NanoString's been around for 20 years, I think, what was the stat, Doug? It's like 80% or 90% were new customers. 80%, yeah. 80% of the CosMx orders last quarter were from people that had never done business with us. These are largely single-cell customers. Yeah that see the importance of being able to get that single-cell resolution within the architecture of the tissue. Yeah, that's, that's really where I, I see also huge demand is this is the whole world of, of single-cell RNA-seq customers are all realizing, you know, spatial is this is the next frontier, if you don't have a spatial component, you're really not almost in the game anymore. That's it. The fields that come in are really interesting. You know, the, the major impact, just like on most of the life science technologies you develop, the key impact really is cancer immunology, immuno-oncology, that's the dominant one. Very rapidly, you see these other areas of immunology come in, you know, the infectious disease, and the autoimmune diseases and organ transplantation and inflammatory bowel diseases, they all came in. Then, like I said, the one that's just running crazy in an exciting fashion is the study of neuroscience and the human brain specifically. Just to give you one little tidbit, you know, the Society for Neuroscience meeting comes up in November. It's sort of like the biggest neuroscience meeting ever throughout the year. I always thought we had a big presentation. At AGBT, we're always one of the biggest ones at AGBT. This year, neuroscience is gonna beat it. We've got 25 accepted abstracts coming up for the November Society for Neuroscience. A couple of invited talks. It'll be our biggest meeting presentation. It went from sort of in the noise for what we do to being the biggest presentations we've ever done at a meeting. It's really all linked to all trying to understand the human brain and what you can do with some high-resolution imaging. Oh, interesting. We'll have to look forward to see some of those abstracts. Doug, this maybe is a question for you, but Joe, feel free to jump in. Just on, on the current manufacturing capacity, I guess, you know, when can we expect the... you know, given the CosMx order strength, maybe can you quantify some of the lead times expected in the year and, you know, has the 10x litigation impacted those manufacturing plans? Yeah. Any many incremental details you can give us on that? Yeah, litigation has no impact on the manufacturing scaleup. As you recall, we had a, a massive kind of backlog coming into Q4 of last year when we shipped the first systems. We continue to scale that up, we're not at kind of steady state manufacturing yet. We will be in the second half of the year. I think for a customer who orders a CosMx today, they probably get that in 2024, which means that, you know, orders in Q3 and Q4 will carry as backlog into 2024, which again, gives us good visibility on revenue for next year as well. Appreciate that. Then I guess just when you look at CosMx and maybe pull-through, I know you haven't really quantified the range, but I guess, are you thinking this could be similar range to the GeoMx, maybe that $70,000-$75,000 on the pull-through? How should we think about that? Yeah, I think that, that's our current best guess. Now, the reality is, when you look at our, our pull-through last quarter was a little bit above that. It was 85,000. Almost all that is GeoMx pull-through, 'cause people are just getting their CosMx, getting trained up. It, it's really too early to try to say what the long-term pull-through is. Obviously, we've seen in single-cell pull-through rates that were, you know, deep into six figures. I, I, I think over time we'll get a better sense. I think it'll be kind of late next year before we know what long-term pull-through is. Right now, they're stocking orders, and how fast they go through that depends on whether they have the experiments designed, whether they have the samples in-house already. Even at, you know, 70-75, that's well above what the nCounter system is doing, right? nCounter does kind of $45,000 in consumable revenue per instrument per year. 70-75 would still be, you know, the kind of highest throughput that we've seen with any platforms we've sold. Great. Then you mentioned the GeoMx. You know, just maybe just an update there on the GeoMx, you know, are, are you starting to are you placing these within discovery labs, pathology labs? I guess, you know, kind of what's been the traction, customer feedback there, year to date? You know, I think the thing that always amazes me in GeoMx, you just look at the publications that are coming out, the, you know, amazing publication record on GeoMx. GeoMx total publications last year in 2023 was over 100 publications, Nature Medicine, a whole bunch of high-quality publications. We're already past 100 publications this year, where we're at right now already. You know, we had four Nature papers just this last month that are GeoMx. What that tells you is that people with GeoMx are doing mature science, science that has major impact that's beyond just a, a niche type of journal. These are going into Nature, like four in a month is a record for NanoString, maybe a record for almost almost any company out there. Life science probably doesn't have their platform with four in Nature in one month. GeoMx is really digging its teeth into the major problems, and especially in the area of in medicine and cancer immunology, infectious disease. I, I'm really happy with with how that goes. When you got high-profile publications like that, and people read those papers and want to be able to do that same thing, they want their Nature paper, they want their Cell paper, Science paper, and, you know, GeoMx gives them a real avenue for that. Yeah, I mean, to add to that, John, I think from the standpoint of the market applications, GeoMx is definitely much more translational in nature, where when you move to CosMx, you're getting into more of the kind of basic discovery stuff, as opposed to, you know, moving down that clinical spectrum, which is where GeoMx has been. It's probably interesting, not a lot of people will realize this, you know, we have three major platforms, the CosMx, the GeoMx, and the nCounter System, but actually all three of those systems are designed with exactly the same hybridization probes. The exact same hybridization engine works for all those platforms. You could actually take some discovery you found in CosMx that nobody's ever found in the world before, you could run a big spatial clinical trial with it in GeoMx, and then you could make a diagnostic assay out of it with an nCounter assay. I would never have to change the design of any of those probes to do all three of those things. I think it really gives the, you know, research scientists out there, they can write their grants to say: I can discover, and I can take it all the way to the clinic, and I don't even have to change my probes. And I think that's a really unique thing for NanoString actually, to be able to do that. You know, we have FDA-cleared assays that went out on our nCounter technology. We've been- we know how to do those things, so kind of a neat A to Z sort of between our platforms. Yeah, nCounter was cleared by FDA back in 2013. For a decade, this chemistry has been out there in clinical applications. To Joe's point, we're excited about the clinical applications. I think we're in the early discovery phase now. Our experience with nCounter was, it was kind of three to four years of discovery and translational work before people found clinical signatures to pursue. Probably the same is true for spatial. I don't think it's a catalyst in the near term, but I think as you get out, you know, call it three years or something, you're gonna start to see people either developing lab-developed tests or actually FDA-cleared products. That's great, too, on just the overview on the ecosystem and the, and the probes there. You know, I guess just to the genomics, you know, any color on how demand has been there, and have you seen any bundling drive further demand with CosMx, GeoMx as you look to, to this year? We did a lot of bundling with CosMx and GeoMx last year. Honestly, I, I think there's been so much of a focus on imaging that, you know, the, the spotlight was kind of taken off GeoMx. Long term, we think that's gonna be still a product that has long legs. You'll recall last year, at one point, there was probably eight different companies promoting imagers, and I think it just the noise level kind of drowned out, the profiler market. We see, you know, for folks who, who purchased those systems, as Joe said, the publications were up over, I think, about 300 in total right now. People who buy the systems put them to work, we expect to, to see that continue. From the standpoint of the kind of complementary nature of the systems, people can use GeoMx now to do whole transcriptome, identify the genes that they wanna pursue with imagers. We're still only doing 1,000 RNAs at a time with CosMx, so if you don't know what you're looking for, you need a tool to screen and kind of whittle down that list to be able to pursue imaging studies. That, that's a good, you know, kind of complementarity of those two systems. Thanks. Yeah, I guess even Joe and Doug touched on it, but it sounds like you think that there's some legs left in this profile. I guess just maybe, and the GeoMx, how do you think of the GeoMx in the life cycle of what the instrument here and, you know, any areas where you think there, there's, you know, next developments we should keep an eye on for future innovations on the platform? Yeah. We're not done developing new technologies and new products for the GeoMx. They don't let me talk about any of those things back yet, but it's not too far away. We got things, still things that you can do. Especially keeping in mind, you know, the GeoMx strong point is really going into archived biobank samples from hospitals, and then, then they are patients that are treated, a standard care treatment, and they're really discovering these biomarkers and trying to understand how do you predict a response in this patient population versus another? That's really where GeoMx shines, and we continue to develop products that continue to enhance that capability of GeoMx. You'll still see them. There's new things are still coming. Yeah. The throughput as well. I mean, GeoMx can run a huge number of samples relative to the imager. If you wanna do retrospective clinical trial work, GeoMx is still gonna be the right platform. All of the imagers, you know, the, the kind of joke in the industry is it's a race amongst turtles. They all have slow throughput because these are many chemistry cycles they run. GeoMx kind of separates itself with the throughput as well. That's great color there. Then, I guess, you know, moving on to the next platform, be the nCounter. You know, any updates on the nCounter business? You know, where are you seeing demand there? This segment actually had pretty good performance, strong performance here in 2Q. Any additional color you can kind of provide on that? I think, you know, I think the... You know, 'cause when I came in in 2012, I came here to do nCounter. It was a unique technology, goes up to 800 plex, all digital, non-amplified barcode quantization of RNAs. This is a really unique capability, and as a translational scientist, I think, you know, we have over... I think it's about 7,700 publications on nCounter now. It's almost hard to track them. I got one person tracks them all. It's, it's really a difficult thing just to be able to track all the, all the publications coming out. I think the fact that you can just get this quantitative up to 800 plex gene expression, and you get the answer the next day. You got biobank samples you wanna go through. You get the entire 800 plex answer the very next day. You don't go through a sequencing core or something like that. That's really resonating with people. They, they want their answers, and, you know, you send those things out to an NGS core, you may wait four-six weeks before you get that answer back. There's a lot of places don't wanna wait four-six weeks to get that core sequencing answer back, and they know they can just take the nCounter, put their biobank sample in, push go, they'll get their full answer at 800 plex the next day. That's, I think that's a real value statement that stays. You know, another area that there was some discussion on at AGBT was AtoMx. Just, you know, any customer feedback there? Any updates on how that has been developing? Yeah. You know, AtoMx is our cloud-based solution to handle this big data. It is. You know, what you find is that the naive customer may not actually understand what they're getting into when you're doing big data like spatial biology is. You're talking about tera bases every week or so. You've got to be thinking a little bit more long term. That's why, you know, we committed all the effort required to generate a spatial, spatial informatics platform. That is what AtoMx is. It's a tremendous amount of infrastructure you have to do for that. When you talk with customers, especially customers that are new coming into this field, especially the translational customers, they do not know how to handle that kind of big data, especially the translational ones. We basically just took that, so they don't have to worry about it. You, you don't have to worry about what a data blob structure is. Some translational scientists shouldn't have to know about that. We built all that stuff in AtoMx, and it really does resonate well. When I'm out there giving talks-... I look at the people that are coming into the spatial biology field, and I'll say, and I just tell them, "You know, with AtoMx, you do not have to generate all that infrastructure at your university or whatever, to do all those particular things, or your own particular labs." And, and the other goal of AtoMx is not just infrastructure, but we wanted to have the person that actually understood the experiment themselves, be able to do about 80% to 90% of the analysis and get the information back. They don't have to toss it over to some bioinformatician that they don't know. That bioinformatician may not know which end of a kidney is up, but that kidney specialist does. AtoMx allows that person, that is an expert, to be able to look at the data themselves and get about 80%-90% of the information without having to toss it over a wall, wait for some bioinformaticist to give them that information back. Of all the things that are customer pleasing, that's probably the one that resonates the most, is that they look at me and it's like, "Oh, man, I can get my answers and I can do it myself." Anyway, that I'm super happy about, you know, how that's gonna progress. It is hard slogging, there's no doubt about it. Nobody has built a spatial informatics platform in the cloud, but we have, and, yeah, now you see the customer usage going way up, and, I think it's a real customer pleaser, especially for the new customers that are coming into the field that may be not big data specialists, which is actually a lot of them. I guess one other addition, too, is the, we think about these large global collaborations, right? We've all tried to, you know, send somebody 10 pictures or something. It's like, oh, the file's too big. Here, when you have folks trying to collaborate around the world, you're able to send somebody a link to the database. They're able to access that. You know, it, it makes, you know, folks globally be able to get in and work with the data without having to, you know, mail hard drives back and forth or whatever you would have done in the past. Yeah, that's a really important point, Doug, 'cause you're doing these kind of studies, you're writing a paper. Nobody writes a paper by themselves anymore, or a single lab even, it's really unusual. You know, you got 30 terabytes of data, Yeah, you don't mail that around. The beauty of AtoMx is that you literally just send your collaborators halfway around the world a link, and they're in the middle of your 30 terabyte data set, and they have all the same visualization tools that they can share across the world. Yeah, it's a really important point. Not to mention, when your paper is accepted, the editors now say, "How do I make that data public?" It's like, whoa, that's a headache. With AtoMx, you can actually basically push a button and say, "Okay, now that data that's in that paper can be part of the public domain community. That's really great over there. I guess, so it sounds like it's more whether you can facilitate through AtoMx than, you know, the infrastructure savings. I guess, have, have you quantified, though, at all, what the impact could be on that from, you know, an infrastructure standpoint on customers on? I mean, it's a, it's a huge money saver for folks to build their own on-premises infrastructure to do this. There's a massive amount of initial investment and then maintenance as well. In our view of this, we think of it as kind of the iCloud at Apple. It's not meant to be a massive profit center. It's meant to make life easier for customers. The AtoMx revenue will be captured in our service line. We think for most customers, it'll be roughly 10% of what they spend on consumables. If you're spending $75,000 a year on pull through, you probably have $7,500 worth of cloud expense. Amazon is our partner there. That's how we think about it, is kind of building this ecosystem around the platforms that make it easier for customers to use it. Doug, maybe just a couple for you here. You know, any thought just on the general environment, just with academic budgets, maybe in the U.S. and you, any trends there you could provide color on? Yeah, I think that, you know, from the spatial side of the business is about 70% academic revenue still. Industrial adoption is lagging behind the academic folks. I mean, NIH is still $50 billion-ish, you know, whether NIH is flat, our guidance this year is for midpoint of our revenue is 40% growth. Obviously, NIH is not growing at 40%. We think that, you know, for customers, they're trying to add new capacity. Rather than go out and buy their 10th sequencer or whatever it might be, they're trying to put new capabilities in place, and spatial certainly at the top of that list. We've also seen you've got Howard Hughes and CZ and all these funding bodies that are making meaningful contributions to drive these big projects as well. We, we don't think access to resources is gonna be a challenge. The one part of the market that has been tight, as everybody's aware, is the smaller biotechs, where access to capital has tightened up. You know, for us, that is a relatively small part of the business overall, and that tightening really started more than a year ago. When you think about comps going forward, that, I think, is already reflected in the, in the kind of market dynamics. Thanks for that color on biotech. I guess maybe the last macro one would be China. I think that's small, though, maybe 5% of revenues for you. Yeah. Any update there on that? Yeah, I think China, it, it's a, you know, glass half full story for us. It's a relatively small part of our business. We have just started to put direct personnel on the ground. We've used distributors in the past. I think China can be a source of upside for us going forward. I know it's been choppy for some companies with big exposure there now, you know, fortunately, it's not a huge part of the business for us. If you look at the overall CosMx orders, like 30% of those are in Asia Pacific, which is really astounding. That's not typically where you see huge adoption that early on. It's actually of our CosMx order book. It's bigger in Asia than it is in Europe. I can't think of another platform where I saw that early on. You know, maybe just to clarify, a point on just the backlog on CosMx earlier, just, you know, when you look at, you know, the backlog there at 2Q and just placements, you know, net, you know, was there a net increase in backlog, or I guess, how, how do we think about that? Backlog came down a little bit because we've started to install a lot more systems. We're still carrying $35 million worth of backlogs. When you, you think about that backlog and our revenue to date, that gets you to our guidance for the full year. That, I think, gives people confidence that, you know, spatial is on track to deliver what we expected this year. Again, I think for CosMx orders from here forward, probably almost all of those go into backlog in 2024, because I, I don't think we'll be able to deliver those in, you know, by the end of the year. By early next year, we'd like to get to a point in time where somebody can get the system within the same quarter they order it. If somebody buys an nCounter, they get it the quarter that they order it, the next quarter, they're up and running. Spatial, there's still a lag. There'll be a lag of a couple quarters still to get people in there. The other thing we've noticed with spatial is, where nCounter has been around 15 years, everybody understands gene expression very well. Spatial is kind of a new frontier. They've got to figure out their experimental design, get tissue, so we have seen it's, you know, it can take upwards of a year for somebody to ramp up their volumes. That was true for GeoMx, and, you know, we don't expect CosMx is gonna be a lot different. It'll be a kind of a slower build curve on the consumables, but long term, we expect it to be kind of a two-thirds consumable, one-third instrument business, and we have the inverse of that now, right? You saw that kind of weigh on gross margins 'cause we have an instrument mix that's heavier than it's ever been in the company's history, but that'll start to kind of flip around 2024 into 2025 as consumables build up. Just maybe one financial question, just one, you know, I guess, top line imbalance I've seen. You just reported 2Q, but can you just remind us cash position and what the outlook is on that for the year? We had about $120 million at the end of the quarter, so still kind of a good, good cash runway there. In the second half of the year, we're gonna see revenue ramp significantly. We've got kind of a working capital dynamic that flips around for us. Through the first half of the year, we built a ton of inventory of instruments for which we pay our contract manufacturers right away. We wait 60 days for our customers to pay us. All that investment in inventory and AR in H1 is gonna turn around and come back to us as kind of a good guy in working capital in the second half of the year. We still expect we can get the cash flow break even with the resources we have on the balance sheet today. Great. You know, Joe, back, back to you. You know, maybe a big, quick, big picture question here, you know, spatial market has seen some pretty high growth, you know, greater than 40%. I guess, do you think this is sustainable, you know, near term, long term? How, how do you think about just broader picture on the, on the spatial market? It's a fun one because I feel like I've been through a couple of these runs of different technologies. next-gen was a run like this, and it stayed. It had staying power. I believe spatial's got exactly the same kind of staying power that next-gen did, because it's just gonna be a fundamental piece of information that everyone's gonna need to have. Every organism on this planet is built in a spatial dimensions of X, Y, and Z. The fact is that all of our textbooks and life sciences, all of our understandings of things are basically... don't have XYZ coordinates on them. They basically are, most of that information came out of low plex technologies or bulk technologies. You know, this feels a lot like what next-gen felt like early on, and I think it has the same type of staying power. I, you know, I do think the single-cell technologies don't have that kind of staying power, and I think you're almost seeing that now. Those things are bending over because, yeah, again, you got to know how the pieces, the parts fit together to function. That's what everybody wants to know, and that's why I sort of started working on this so early, because I, I knew eventually people would wanna know how it all fits together. I think next-gen and spatial are here to stay. I think, then I think single-cell is probably gonna be something almost... yeah, it's hard to know exactly, but I don't think it's got the same staying power. Maybe just add one, one kind of closing point, too. When we think about the total addressable market, you know, there's roughly 7,000 labs out there that can own both the CosMx as well as the GeoMx, so that's kind of 14,000 instruments out there. We currently have an installed base of about 450 systems, so we're, you know, certainly in the 1st inning, it's kind of low to mid single digits, market penetration. Most of these cycles, if you look at, again, whether it's next-gen or single cell, these are decade-long product cycles. They're things that play out over a long horizon. So we're, we're super excited about the growth in spatial, not just in the near term, but, you know, for, for many years to come. That's great. It sounds like long horizon there, too. You said, you know, decade or so to play out. Is that the way- Yeah you're thinking about it? Absolutely. I think they're gonna be here. It's just like, you know, think about microscopes in life science. We are every lab, and you can't have a lab without a microscope. In the end, these images are just microscopes. They are just really crazy microscopes with this amazing amount of plex built into them. I mean, microscopes are here to stay. These images are gonna be here to stay. I think, you know, that, yeah, this will be the way imaging gets done from here on out in life sciences. Great. You know, maybe in the last minute here, just, you know, last one, and we even, I guess, digging in a little bit further, you know, just on clinical spatial adoption, you know, where do you think we are today? You know, where do you think we could be in five years, and where do you see the most near-term opportunities there? Yeah, I mean, I think, you know, if you think about medicine in general and pathology, you know, every time you take every tumor that gets taken out of a biopsy, gets imaged as a H&E or whatever. It's a fundamental part of the medical fabric of the globe, right? Doesn't matter if you've been India, Russia, it's all the same thing. Now we have this whole layer of molecular information, all of that molecular information used to just be in molecular diagnostics. You got tumor mutational burden, assays, whole exomes, whole genomes, or whatever. That's the world of molecular diagnostics. You get this crazy combination where you basically have, you know, immunohistochemistry, which really isn't a, historically, a molecular thing, but it's the basis of how medicine gets practiced. Then you got molecular diagnostics, which is this advanced understanding of how, what's happening in your body, and with high plex spatial, you're basically taking those two worlds, and you're just cramming them together. They are together for the very first time. You got the power of molecular diagnostics, but with the fundamental understanding of the spatial context at the same time. I, I think it's a real- you can see how those two things fit, and they come together, and I, you know, I believe, and, the whole reason I got into this kind of business in the first place, is that you want to have an impact on the world, and you want to have an impact on the health of the world. I think that's why a lot of people are in life sciences and especially tool makers like I am, and this is gonna be the thing that does it. We're gonna bring those two things together. It will really transform our understanding of health and disease. There's no doubt in my mind. Now, what assays finally turn into clinical assays, the form, I think in some respects, it's a, it's a bit irrelevant. I mean, you know, you can say, okay, it's gonna take the same path like NGS did. 'Cause at NGS, the early days of NGS, say, "Okay, you're just gonna discover with NGS, you're gonna make yourself a little qPCR assay, and that's the diagnostic." Then what you find is that, okay, well, I need a panel that's a little bigger than the, than the qPCR things can do. All of a sudden, tumor mutational burden came around, then whole exome. You know, I think that it's gonna follow the same type of progression. They're gonna marry, you're gonna get clinical applications, and then the clinical application answers are gonna get more and more sophisticated, and as they do that, just in NGS, your plex ramps up to accomplish whatever is needed. That's the long-term vision for that, for that overall process. I'm super hopeful it will change our understanding of health and disease big time, just like NGS did. Spatial's gonna do a whole another level on top of that. Great. Well, that's I think, a great place to end this. With that, we're out of time. Joe and Doug, thank you for joining us today. Great, thanks. Thanks, everyone. I appreciate it. Hey, thanks, John. Thanks. Fun conversation.