Tempus AI, Inc. ($TEM)

We're going to get started. Great. Thank you. Welcome, everyone. Welcome to our inaugural Investor Day. Great to have you here in person, and thanks to those who are joining us virtually today. Before we begin, just a quick reminder, today's presentation will include forward-looking statements, which are subject to risks and uncertainties that are described in our SEC filings. We have a number of our speakers today that are outlined here on the slide and are joining me here on the stage that you'll be hearing from throughout our sections. And here's a quick look at the agenda. We'll be starting looking at our diagnostics business, and then we will turn it over to data and applications. And then finally, take a moment to walk through the financial outlook. Each section will be followed by Q&A. We'll take questions both from the audience live and virtually. And with that, I'll pass it to Eric.

Thank you. Welcome, everybody. We're going to go kind of quick, but we'll have a Q&A section throughout. So hopefully, we'll be able to get to all your questions. We also had a few -- we have a few releases out this morning, so just giving people a heads up. So one is already, I think, hitting the other ones coming short me. 10 years ago, we started attempts to solve a single problem, which is could you use AI to essentially unlock precision medicine. See if my clicker works. No clicker. Technical difficulties on clicker. The laptop isn't connected. So we need to... Just want it's like you feel like here at your house where you're just kind of moving around the room. Given that this is not our space, it's the building space, it's hard to complain. It were nice enough to just use it. We disconnect the laptop. Okay. Hold on, I'm now clicking, which is exciting. Okay. Given our clicking capabilities are back, I'll start over. So 10 years ago, we started attempt to solve a single problem. Could we use the artificial intelligence to unlock precision medicine. In order to do that, you basically need 2 things. You need vast amounts of proprietary data to build models to bring the benefits of [ AIDA ], health care and you need a distribution system to take those insights and deliver them to the hands of physicians and patients. Tempus is unique in that it has both. We have vast amounts of data. We have vast compute and modeling capabilities and a vast distribution system to essentially target the main use cases within health care, which is can you match patients to the right drug, the right trial, the right therapy. But in order to do this, you have to build a sustainable operating system. And that, I think, is one of the first places that makes Tempus truly unique is that we have built this connected ecosystem that allows us to essentially generate vast amounts of data from the clinical workflow, turn that data into insights and then essentially feed those insights back in the U.S. health care system, which then makes people want to connect with us more, and therefore, feed us more data, generate more insights, deliver more applications. And this whole ecosystem is now sustainable, which is probably most exciting parts about it. It's large. It operates at scale and it's sustainable, meaning we don't have to invest billions of dollars consistently in making it work. It works every day. This integrated and sustainable system is essentially spanning vast amounts of health care data, morphologic and molecular data. So this is a molecular data that you would generate from sequencing patients or producing molecular data across a wide range of data. It's DNA data, RNA data, methalomic data, imaging data, from CT scans and MRIs. It's digital pathology slides, it's clinical data in structured and unstructured format. And all this data comes together in real time to essentially create this -- I think, I lost my mic, this network of bag system, something will always not work, and we'll just kind of bounce to hit it to. The ecosystem we built now exists at scale. So we have -- we're connected to about 65% or 2/3 of all academic medical centers in the U.S. The vast majority of oncologists in the U.S. are now connected to Tempus in some way, shape or form ordering our tests. We run a large volume of tests, and so we generate lots of that molecular data at the beginning of that flywheel. We're connected to more than 5,000 institutions throughout the United States, which means we're connected to a significant percentage of the U.S. health care system. There's maybe 8,000 hospitals or so. We generate enormous amounts of data and then have built a world-class team to make sense of all this data. So in addition to being sustainable and running this operating system, we are sustainable at scale. The output of all this activity, where we're essentially trying to generate data as part of the clinical practice, turn that data into something useful generate an insight and put in the hands of everybody who needs it, whether those are physicians, patients or researchers. This has now produced an enormous amount of data. Over 500 petabytes of data for people who have been watching the growth of our data base over time. It's really quite extraordinary. I mean, not long ago, we were 50 petabytes of data. Here we are some 5-plus years later at 500 petabytes of data. It spans over 45 million patients. There's over 9 million images in that data set, which is just one of the largest digital imagery data sets that we know of in the world in terms of digitized pathology slides and radiology records also connected to clinical outcome response, including a large volume of samples we've sequenced over 4.5 million. And then the very bottom of this data set or this funnel is over 400,000 of these really, really rich multimodal records. These are records where we have typically DNA and RNA and clinical data and outcome data response data, adverse event data, we have imaging data all the totality of what you would need to basically interrogate real-world data and figure out all the insights we're going to talk about in a minute on the biopharma side. In order to make this business sustainable, we've divided it into essentially 2 parts. We built a diagnostic business and then a data and applications business. So I want to first start with the Diagnostics business. What makes our Diagnostic business unique is the comprehensive nature. If you're going to generate vast amounts of molecular data and then be in the business of connecting that molecular data to clinical data so you can contextualize it, the question you'd ask us why? Why am I going through all this effort to contextualize molecular reports? And the answer is that if you just run sequencing and generate like here's a patient that have a mutation and I'm going to hand that to somebody at best, you're basically in the business of targeted therapies or targeted medicine. You're not really in the business of precision medicine because you know nothing about that patient. And so the journey we set out 10 years ago was could we essentially make diagnostics smart. Could we help contextualize them and basically wrap technology or AI around them to help physicians make really high-quality decisions and help researchers do much more efficient research. On the diagnostics side, that begins at not just generating an insight in a comprehensive manner to get an answer or result, but then bring it all the way through. I know some -- I've learned something, how do I connect it to clinical data so I can essentially figure out like if I find a mutation, I don't want to recommend a therapy that a patient just took in a prior line and failed, recommending that again would be pointless. I don't want to recommend a clinical trial that, that patient is not eligible for because the patient happens to be a smoker. And one of the exclusion criteria of this trial, I would recommend you can't be a smoker. So by connecting rich molecular data or any kind of laboratory test results or diagnostic data to clinical data, you can contextualize it to go from kind of answer to insight. So I found something interesting through EHR connections I'm now going to contextualize that, so I can make a more intelligent decision, which is really powerful for clinical care, but that same data, that same vast amounts of multimodal data where you have an insight connected outcome and response is also what's needed for research. So all the research that people are trying to do to figure out how to make sequencing more useful for cancer patients. Our technologies are empowering that at scale. And then once you're generating lots of molecular data and you're contextualizing it and you're powering a bunch of research, you end up with this last mile, which is how do I put that in as many hands as possible we can. And so the investments we've made in connectivity and AI essentially allow us and will allow us in the future to distribute these insights at scale. And our goal is not just to distribute them to all cancer patients, but to all patients in the United States. Tempus is successful over time, we'll be connected to every hospital in the United States, almost every hospital across all major disease areas. And every time there's a diagnostic insight, we will be in the middle of that trying to figure out how to take that diagnostic insight, contextualize it, wrap a whole bunch of insights around it that only people like us have because of the nature of the data we have and then deliver those to every clinician in real time so that patients are always on the right therapeutic path. Begins with having -- we started in cancer, you have to pick a place to start. So we started in cancer. We started trying to make molecular testing in cancer as intelligent as we could or compressional profiling as intelligent as we could. And we started -- we made a decision early on that if we were going to be in the business of contextualizing tests, we want it to be as comprehensive as we could be. We didn't want to just give somebody part of an answer part of the time wanted to give them the answer all the time. And so in cancer, that begins -- if you look at the compendium it starts with who's at risk of getting cancer, it then translates into who is just being diagnosed with cancer and how do I essentially put them on the right therapeutic path. And that kind of bifurcates into solid tumor profiling or liquid biopsy because not all patients have enough tissue to be a sequence. So you need both liquid solution or blood solution and a tissue solution. And then post treatment, how do I monitor these patients? How do I look for when their disease might be coming back or see if the therapies I'm giving are sustainable. And so Tempus operates across the entire spectrum. We are strong in hereditary profiling, strong in therapy selection, both solid tumor and liquid biopsy and then -- and strong in all the ancillary tests that come along with that and then strong in MRD and monitoring. So we'll cover all that shortly. Here's just a quick snapshot of the comprehensive nature of that portfolio. We have a series of FDA-approved assays. We just added to our portfolio this morning. We had a tumor only FDA approved historically, and we now have tumor only approved, which is -- which for us is quite significant in that it expands the amount of FDA-approved tests we can offer to essentially 100% of our DNA portfolio. And given that we have a DLT pricing, that's quite powerful. So that was a big approval for us. And then we have a series of other -- as Jim will talk about a little bit. We have a series of LDT tests. And those cover RNA, liquid biopsy, other areas, all ex some things of that nature. Several of those are also going down this kind of FDA regulatory approval path. We have a series of pharmacogenomic assays we offer things that have become super powerful these days, whether it's DPYD or UGT1A1, we also pharmacogenomic profiling of patients with neurological issues such as major depressive disorder, bipolar disorder. We have a whole bunch of algos that sit on top of these diagnostics, we'll talk about in a second. On a series of tests that typically ordered alongside these, whether it's immunochemistry stains or other tests of that nature. And then we have, obviously, a fairly large and growing portfolio in rare disease and cardiology and obviously, hereditary profiling. So a significant body of assays. And essentially, when you -- to understand Tempus' strategy, it begins with what is the diagnostic that's going to be most commonly ordered in this disease area and how do I either offer that diagnostic or partner with somebody that's offering that diagnostic, both of which are perfectly fine solutions. Generate that diagnostic data, begin to generate or consume that diagnostic data at scale in real time across a large percentage of the U.S. market and then begin to collect clinical data that's connected to that diagnostic, so I can figure out like what's happening? What drugs are patients going on? Are they responding? Are they not and essentially create a self-learning system to make that diagnostic better and to contextualize it, to make it personalized so that when a physician orders that diagnostic, it actually helps them figure out what to do next. I'm going to bring up Mike to talk a little bit about some of these tests in greater detail.

Great. Thank you, Eric. So just starting off at kind of the top -- one of our kind of core workhorse assays on the solid tissue side is our xT assay. This is a 648 gene assay that combines both the molecular insights that are coming out of the genomic testing as well as how Eric mentioned contextualizes that with structured clinical information. We've had the tumor normal FDA approval for some time. Obviously, as of this morning or as of last night, I should say, our tumor-only FDA approval came through. So we're very excited to be able to extend now this series of FDA approvals beyond just those 2 normal patients. Because as many of you may know, there's many cases where we're able to capture that blood. In other cases, we may not for a number of reasons. And so now this will allow us to service all of those patients that are coming in through the clinic. On the RNA side, there are many cases where the RNA signatures that we're able to identify extend beyond what we're able to find via the DNA findings. So for example, with fusions and rearrangements we're able to now find many of these patients who might otherwise not have a therapy that's found by a DNA. They are able to be found when we're looking at the RNA signatures. So this is an area where we're seeing a preponderance of orders kind of coming in both inclusive of DNA and of RNA. And I'll talk about next a little bit some of the studies that we've done that show kind of that incremental benefit that we've been able to see as a result of adding RNA onto that DNA finding. So we have published a number of cases where we're looking at the incremental benefit. So what are we able to find when we're looking at patients who received both the DNA result as well as with an RNA result. And there's a significant portion around 21% of patients that do find these driver mutations, they're able to now find FDA-approved targeted therapies that go beyond what would have been possible had this patient only received a DNA result. And so not surprisingly, we're seeing a large portion of our patient population include these RNA xR orders in addition to our xT orders. So moving from the solid tissue side of the portfolio to our liquid biopsy. So in these cases where we're not able to obtain tissue or in many cases where clinicians are looking for additional insights beyond what they're able to find by just looking at that tissue by itself. Because of the shedding characteristics of these tumors, liquid biopsy allows for an incremental set of findings in a similar way to how RNA is additive to the xT result our xF portfolio, which is inclusive of 2 assays. So we have 2 flavors of liquid biopsy. The first is 105 gene and the second is a 523 gene assay that allows us to look at and again, find these shedded tumors and the DNA associated with other mutations that are again, looking at those 2 different flavors. This is a faster test from a turnaround time perspective. So in some cases, not only when these patients don't have tissue to be able to be sequenced, they are also optimizing for turnaround time. And so in this case, we're in that 6- to 7-day range, which has been helpful for many clinicians in many patient cases. Importantly, what we're seeing is that these clinicians are also looking at this liquid biopsy and the evolution of tumor biology over time. And so this translates to having multiple xF tasks that are ordered to understand what are the resistance mutations that might be happening in a given patient case. So for example, with EGFR in lung cancer cases, a resistance mutation may be able to be found during the course of therapy. And so we're seeing an increasing cases, clinicians looking at what are the multiple time points that might be appropriate to order a liquid biopsy test. And this is our mechanism for being able to help identify that. Similarly, when clinicians are ordering our solid tissue test, they are, in many cases, adding the xF liquid biopsy test, as I mentioned, and this does provide this incremental benefit. We've done some publications on this that find around 9% additional actionable variants are found when liquid is added into that solid tissue case. So clinicians are, in many cases, including this as part of their standard routine course of care. And then as we do look at what the -- those kind of incremental actionable findings are similar to a number of the studies that we've done on the RNA side and then looking at those resistance mutations, we've isolated that looking at specific subtypes and what are the benefits of xT and xF in different cases. And so this is data that represents what are we finding incrementally beneficial in cases like lung and in breast and in prostate and in CRC. And there are these kind of incremental additive benefits. And so again, additional evidence as to why clinicians are increasingly looking at both the solid tissue profile as well as a liquid biopsy profile.

So just for a second to kind of provide some context. So when we -- when Tempus is about 10.5 years old [indiscernible]. And when we started, Foundation Medicine was the leader by far in [ solo ] profiling and had a liquid offering in [indiscernible] is obviously better and liquid. And at the time, people thought it would be very hard for us to kind of catch up and make progress. And if you look at the progress we've made over the last -- in the labs, maybe 8 or 9 years old, over the last, let's say, 8 plus years where we become #1 or #2 in both of those spaces and have unit growth that is kind of a best-in-class. It really speaks to not only the comprehensive nature of these assays and how kind of good they are, we are. The gold standard, I think we're in that top-tier gold standards in both solid tumor profiling and liquid biopsy, but also the technology we wrap around these tests. We'll show you in a few minutes, our main ordering system, which is called Hub. But in addition to that, we don't just run these tests in isolation. We connect all these tests to each other, and we connect all these tests to clinical data and outcome data. And so what ends up happening is the ecosystem, if you're a doctor ordering our tests, it just gets smarter and smarter and smarter, which has led to such a high new physician acquisition rate and such a high physician retention rate. We talked a little bit in the last quarter about the algos or algorithmic diagnostics that sit on top of these tests. And we have a variety in market today. We have algorithms that predict homologous recombination deficiency, algorithms that predict cyto tumor origin, algorithms that predict a whole bunch of other clinically relevant topics. But also one of the newest algos we've deployed is called our immune profile score, which Ezra's going to talk about in a second. These algos now have an attachment rate of greater than 40%, which essentially means a doctor is choosing to bolt on one of these algorithms, 40% to the time they -- more than 4% of the time, they order with temp which means they find enough value to say, I want that additional insight. Ezra, you want to talk little bit about the IPS?

Yes. Thanks, Eric. And we set out -- I'm Ezra Cohen, I'm the CMO of Oncology, and I'm a medical oncologist by background. We -- if we can advance the slide. We set out about 2.5 years ago to solve a problem with the data that Eric was talking about. Because we have the clinical longitudinal data associated with both DNA and RNA, we were able to answer a fundamental question on oncology, which patients benefit from immunotherapy. Not in just 1 cancer, but across all solid tumors. And that's exactly what we did with IPS. IPS is a quantitative score. Here, you see an example of a patient who scored high. IPS high predicts a benefit to immunotherapy, IPS low predicts a patient that will not benefit from immunotherapy. It goes beyond the traditional biomarkers that we already have and it gives providers that extra insight to make the decisions on whether they're going to use immunotherapy for their patient or whether they should go to a different modality. It's provided a tremendous amount of value, and the feedback has been incredibly positive. But that's a perfect example of what we can do with what we have built at Tempus, as Eric described. Thanks, Eric.

And the benefit of these algorithms that they're essentially providing insight on top of something else that might be previously known. So there are historic markers for IO response like tumor mutational burden or TMB. Unfortunately, it's just wrong too often. It's a -- these are kind of course scores that might be right, might be wrong. And if you have vast amounts of data, you can kind of refine the score. And in this case, we've unlocked roughly 20% of patients who you wouldn't think would respond to immunotherapy that will, and 20% of people you think should respond to immunotherapy that won't. And that journey, which we'll talk about a bit more we get into the foundation model in a second is what we expect to happen across all biomarkers and all diagnostics, not just in cancer for the disease areas. It generates some kind of diagnostic insight, laboratory test result. You connect it to rich outcome response data over time. You track what's really going on, you refine the diagnostic, it becomes smarter, collect data, track, refine, so on and so forth. Before we leave therapy selection or comprehensive genomic profile and go to MRD, I want to talk a little bit about the growth drivers of CGP. We believe that the unit growth rates we're experiencing, which are pretty extraordinary are sustainable for a long period of time for a few reasons. One, CGP is still not saturated. There have been a significant number of reports count recently that estimates the percentage of doctors ordering these tests in the kind of 40% to 50% range. There's a significant number of folks that still don't order these tests even though they should in areas where NCCN guidelines call for these kind of tests being ordering. That's not a small percentage of cases. And so we suspect the whole market will grow. And since we're growing faster than the market, that will accrue to our benefit. The second is that we're the beneficiary of 40-plus years of research where you're essentially tying a biomarker to some kind of therapeutic benefit. It started in cancer. It started with [indiscernible] and it started with all the genomic work we've done in that area. And so if you've looked at the trend, we've been sequencing patients earlier and earlier in their diagnosis. And so we suspect over time, the vast majority of patients even in the Stage 2 and Stage 1, depending on the disease area are going to be profiled. So that's another benefit. The third benefit is that there's been a significant migration to more comprehensive profiling and therapy selection. We've broadly published on the benefits of doing solid tumor profiling and liquid biopsy profiling as have many of our competitors. So there's now a large volume of work around the benefits of doing both. There's an equal body of work about the benefits of doing DNA and RNA, which is kind of having an explosive moment in terms of therapeutic relevance. And so we suspect that trend will continue. So you have a bunch of docs that are ordering that will you have stage expansion into earlier stage and you have this trend being more comprehensive ordering more tests. So we suspect the top companies in therapy selection or CGP will continue to do well. Well, we suspect we'll do better than that group or most of that group in large part because of the technology investments we've made, which integrate our platform, contextualize it, make it really smart. And so doctors have been flocking to our platform over the last 5-plus years, and we don't see that trend slowing down. With that, I want to hit MRD and monitoring. Let me provide a quick overview, and then I'll bring up Kate. When we decided -- just in each area, we have to think strategically, we got into therapy selection. That's where we began. We began in solid tumor profiling, had to earn the right to go to liquid. That was a conscious decision we made. We then began doing hereditary profiling and realize that the best way for us to win in that space long term and really to tap into the bigger part of the market that's currently untapped today was to acquire Ambry, who was the gold standard of that assay. And we'll talk a bit about that in a second. With MRD, it was a bit trickier. It was an emerging space. It was a new space. And so we had to make a series of decisions. One is, what did we think was going to win long term was it going to be a tumor-informed profiling or tumor naive. So -- and that is still, I think, up for debate. And the next question is, how is the science going to evolve in terms of these assays and their ability to detect cancer earlier and earlier. And the only thing we knew is that unlike other spaces, this space was exploding in large part because of Natera very rapidly, meaning what took Foundation Medicine, let's say, 10 years to do from a commercial perspective, Natera was doing in like a year or 2. So this space was growing very, very quickly, getting to scale, generating lots of money and yet the space hadn't yet really kind of landed in the place that's going to land long term. So our approach was to kind of derisk that by having multiple irons in the fire. And so we made a decision to partner with whom we thought was best-in-class on the tumor and form side, Personalis had a whole genome assay at the time and to make investments developing our own tumor naive portfolio, knowing that those investments would be significant over time. So I want to cover -- I want to say one thing that I'll bring Kate. We have a distinct advantage in this area. One is we generate enormous amounts of data. So that data -- the same data that's allowed us to build best-in-class assays in solid tumor profiling and liquid will allow us to build best-in-class assays in MRD, we just -- we've got enormous data points and can learn and refine over time. The other is that were connected to a vast number of oncologists in the United States, and so we're a partner of choice for most people that have emerging technologies. With that, bring up over to Kate.

Thanks, Eric. As Eric mentioned, our main goal is to be able to offer solutions across the spectrum for all patients at all time points in their journey. And there are advantages, as he mentioned, to both tumor-informed and tumor naive. And we like to think that we have the best of both worlds. So first of all, we've partnered with Personalis for their tumor-informed assay, which we mentioned we really feel is best-in-class in terms of sensitivity levels that it can get to, and that's because of the technology that they use. So Personalis and xM NeXT leverages whole genome sequencing from tumor tissue and then actually looks at up to 1,800 different variants kind of leveraging that whole genome information to be able to sort of personalize them and follow and track that patient's variants that are contributing to the disease. So this really results in ultrasensitivity and allows you across breast, lung, IO monitoring and multiple indications to get a really sensitive result way ahead of when you would pick up disease on other modalities like imaging. Personalis has been in the game for a little while. They've been investing heavily alongside us in different clinical studies to be able to show the validity and utility of this assay. And of course, with MRD, one of the challenges we all have in the field is that you really need to tie this to outcomes. And that means that you have to follow these patients for a long time to be able to understand how they're going to ultimately do in the clinic and with different treatment modalities or with different clinical decision-making, how will that translate into survival and other types of clinical benefit. As great as tumor-informed assays are and the fact that they can get to this really low ultrasensitivity, which is very useful. There's also the challenge that not all patients have tissue. Certainly, some indications like lung or breast or certain time points along a patient journey, there really isn't the advantage to be able to leverage that tissue. And so if you want to cover all solutions for all patients, you also need a tumor naive assay or a liquid approach. And we have developed that technology here at Tempest, we've launched a first assay in CRC, where we're able to now just from a liquid test to be able to monitor for sensitivity and pick up early stages of tumors, very similar to the tumor-informed version, and we anticipate to continue to improve on this technology. This is a technology where, as we've been talking about the power of the Tempus data and the fact that we have so many different types of data coming into our ecosystem, this is one space where that data will help you continue to improve when you don't have that solid tumor tissue to lean on. We are also in the journey of continuing to generate a lot of clinical evidence across many different tumor types. So we started in CRC, but we are looking at lung and breast, in [ PINK], in head and neck and all of the indications where a patient luckily may need to have an ultrasensitive assay or an assay to monitor for therapy response. This is good news for patients at large because it means that therapies are able to achieve really deep responses. And now we actually need these diagnostic tools to be able to understand still within patients who've had great responses who might ultimately relapse, why are they relapsing and what therapy do they need next. So over the next few years, we'll continue to expand and roll out these studies and then be able to follow these patients, validate the assays and improve the technology.

So before we get to hereditary, I want to cover a few things. So on the tumor inform side, by partnering with Personalis, we have a best-in-class assay in market. Obviously, they're expanding the number of indications where they're getting coverage. And so we will be kind of unlocking volume over time as the unit economics continue to improve on their side. They started like a year ago, had 0 approved. They now have 3, and there's more coming. It also, I think, speaks on the tumor inform side to the benefits of our platform. We went public, we told people that at heart, we're a tech company, and we didn't expect to run every single diagnostic in the world across every disease area. And so we would partner with people and essentially open up our platform much the same way Apple Assay platform where you have apps that they make money off of. If you look at the unit economics of MRD on tumor informed, it speaks to that. We essentially generate the kind of profit at the present moment from that test that we would generate if we ran it at scale on our own. So we're a bit agnostic as to whether -- from a profit perspective, we're agnostic as to whether or not we partner with somebody and generate the EBITDA we would love to generate or whether we run the test and eventually generate that. So we happen to have -- we're fortunate in that regard. On the tumor-naive side, we launched our first version of this assay, and it was performing pretty well in terms of its stratification of patients. But the market is moving so quickly that the limits of detection in PPM you have to hit is just much lower. And so we were out there with an assay at, I don't know, 500 or 1,000 ppm and the market was at 100. And so we began working on the next generation of our tumor-naive assay, about a year ago, I think. And we are already hitting -- getting close to some of the levels we put on this slide. And so we intend to migrate the entire platform to this new version. At the same time, we'll replace our CRC assay, we're also focused on the fact that given how fast things are moving, if we go one indication at a time, by the time we get to the third or fourth indication, the market will move again. So we have to kind of skip a bunch of that and really go from redoing CRC to focusing on pan cancer. So that's our strategy in naive. Given that it represents 2% or 3% of our volume, it just isn't material today, but eventually, hopefully, will be. On that note, let me hit hereditary for a second. We're going to proceed with Tom, and then we'll come back. It's all yours.

Thanks, Eric. Yes. So with the acquisition and now the integration of Ambry Genetics into Tempus Diagnostics, Tempus picked up a pretty significant footprint in hereditary cancer testing and an emerging and growing footprint in rare disease. And so I'll talk briefly about both of those. Those of you are familiar with the marketplace. There's 3 large buckets of testing orders. The biggest being the genetic counselor space, they order roughly 50% to 55% of all the genetic testing for hereditary cancer testing goes through the genetic counselors. We have about 1,700 active ordering genetic counselors. The other buckets are medical oncology and OB/GYN. The opportunity in territory cancer testing is quite immense for a few different reasons. One is there's an established NCCN guidelines. There's a robust reimbursement model and every commercial payer as well as Medicare covers this testing for patients who'd be criteria. The big untapped opportunity for us is in the unaffected patient population. There are literally over 70 million people in the United States who meet NCCN criteria for hereditary cancer testing, and they're just going undetected. Roughly 1.5 million to 2 million tests per year is what's happening right now. So there's a huge opportunity, and Tempus is very uniquely set up to capture the unaffected patient population. And actually about 2 years ago, which tells me our strategies are working, we've finally surpassed where we're doing more unaffected patients than we are affected patients. So things are moving in the right direction, but there's a huge opportunity here. We have an automated high-risk platform called Care. There's roughly about 250 sites and growing across the United States right now. And what Care does with our partners, we proactively engage with patients. We extract information to assess the risk profile based on NCCN criteria. We do pretest education for those patients, and then we flagged them for our clinicians, then when they walk into the clinic for their next visit, they get informed again about the testing. We get the testing done. Obviously, we run the testing in our lab. When the results are ready, we send those back out we can also send those directly to the patient if they desire and do the post-test counseling. So this automated platform allows us to access that unaffected patient population, which we see as really prime for significant growth. This year, we're adding enhancements to care. I won't go to all of them, but on the front end, we're going to be seamlessly integrating into EMRs starting with Epic that happens this summer. And then on the back end, we have what I call like a safety net wrapped around our patients. patients going through the health systems create lots of care gaps just because of the workflow and the complexity of it. We're launching a product in the summertime for starting with breast cancer and then moving to other cancers that have guidelines for germline testing. Just to help with the scope of how big the opportunity is. One of our pilot sites looked at breast cancer patients alone for the past 60 days that are active care throughout that health system, identified 5,000 patients, breast cancer patients that should have had germline testing that didn't get it done. The Care platform will be able to wrap a safety net or umbrella around that, make sure we identify those patients for our clinicians and then send them back a fly to get tested. So a huge opportunity for growth for us there. From a testing product portfolio, we have roughly 100 tests. I won't go into all those tests, the most popular or CancerNext pan-cancer test. Cancer [ desk ] expanded and the BRCAplus. The advantage for BRCAplus mostly for breast surgeons, from the time we receive a result to the time we get to report in their hands is roughly 3 to 5 days. So that allows them to order a test for a patient and schedule them for surgery within a week, which is a very big advantage for them. On CancerNext and CancerNext-Expanded. These are the most popular tests that we have. I won't go into all the details here, but I do want to highlight the RNAInsight. RNA is in addition to these tests that we launched in 2019. RNA provides us data that other laboratories don't generate. And so for deep intronic mutations, splice site variants at a high level, we basically identify mutations other labs don't and we can classify mutations other labs can't. We did a publication in early 2024, looking at roughly 40-plus thousand RNA patients. It increased our diagnostic yield by almost 9%. So this is the first time in [indiscernible] cancer testing and then like a decade where you can actually prove that you have a better test on the marketplace. We crossed over -- well over doing well over 1 million RNA patients to date, we'll probably do another 400-plus thousand this year. With regards to rare disease, [ AmREIT ] actually been in rare disease for quite some time. We launched the first commercially available exome test way back in 2011. So our product portfolio currently, we have [indiscernible], we have Exome. In the fall of 2024, we launched a test called ExomeReveal, where we took our RNA expertise and added it to Exome. We saw roughly a 20% increase in diagnostic yield over standard exomes, also a nice bump in volume when we launched that test. And then in this summer, we're about to launch our first clinical whole genome sequencing test. From a market perspective and diagnostic yield perspective, roughly 80% of these diseases. There's about 7,000 rare diseases out there. It impacts roughly 8% to 10% of the United States, and about 80% of those are genetic, 50% of those were children. These folks go through this diagnostic odyssey. It takes typically 5 to 7 years to identify the genetic disorder for these children. Unfortunately, with the advanced technology, diagnostic yield is increasing. So ExomeReveal again added more diagnostic yield to our test, and we're about to launch whole genome sequencing, which we expect to see another 5% pickup. So instead of 1/3 of these patients being identified, it will be a little over 40%. And then the last thing I want to touch on because this is also very unique to Tempus Diagnostics or Ambry. We have a program called Patient for Life. And so if you get tested with Ambry, our scientists are constantly reviewing the literature and looking for a new genetic disease connections. When we find these, we reanalyze our patient data and then we contact the physician, provide them with an updated reclassified report for that patient and then have our genetic science liaisons to work with them to answer any questions that they have. This impacts roughly one in 20 of the patients, so about 5% of our total patient population, which is a pretty significant increase in diagnostic yield and identification for our patients, and this is very unique to [indiscernible]. Thank you.

So just really quickly, if you'll notice like one of the common themes here is multiple tests, generating multiple amounts of data connected to other forms of data like data about this patient over time for life and it kind of yields over time a really a more intelligent platform that can grow. And so that's just our playbook area by area. [indiscernible]. So really, I just want to cover one thing. We're going to tally give you a quick demo of Hub for a second. We haven't demoed our 2 main systems, Hub is our main system that physicians use and Lens is the main system that researchers use on the biopharma side. We'll demo those today for a second. But all this comes together, right, in this giant connected ecosystem. Like every part of this company is working on this platform that essentially generates rich diagnostic data, typically molecular data, connected to other form different data modalities, again, phenotypic data, morphologic data, some kind of text or image or [indiscernible], whatever puts it all into this big giant environment where we run compute, generate an insight, put the insight back into the test put the insight back in the hands of a physician. And so this is all happening at scale. And because we're connected to over 5,000 hospitals, where we have data connections and BAAs and legal agreements and IT connections and all these really complicated things, we're in a unique position to pull data out, generate an insight by augmenting it and then putting it back in. So with that, let's talk a little bit about our main platform. The platform essentially, you're going to see this both today in Hub and in Lens. Hub is essentially an ordering tool that we use for physicians to connect with. You can go to it directly on your iPhone or iPad. It's integrated in with most major EHRs in some way, shape or form. And then sitting inside Hub is this brain, which we call Tempus One or One. And this is essentially all the benefit of the agents we have built live inside One. We have built a ton of agents. I don't thousands of agents that essentially take disparate miserable siloed multimodal data and make sense of it and the challenge with large language models, regardless of what model you're using, Quad Google's models through Gemini, ChatGPT, these models were not trained on health care data. They were trained on Internet data. So they just don't work with perfectly with digitized pathology slides or DICOM files from CT scans or rich molecular data, if you dump in trillions of As and Bs and Ts and Cs, you don't get much out of a multimodal, but a large language model. So sitting inside these products is our Brain 1. And with that, Laura is going to give you a quick demo.

So Laura Elster, Chief Commercial Officer at Tempest. I'm going to walk you through the demo of Hub. Great. So I am a provider. I've logged into the platform. This is where they can order tests or order kits and interact with the results. So I'm going to go down to a fictitious patient, [ Christina Collins ]. And you see that [ Christina Collins ] here, her provider ordered comprehensive testing. So you see our tissue test, our liquid test, we have DNA, RNA, a handful of IHCs as well as some of our algorithms. So the liquid biopsy on day 5 came back and it surfaced a [ PIK3CA ] mutation as well as a low blood tumor mutational burden. There weren't any driver mutations found on liquid. The tissue test came back and actually confirmed no driver mutations, but there was a high tumor mutational burden as well as a positive PD-L1 score. And so this is an opportunity where a provider may want to chat with Tempus One, as you've heard and asked something like how common, how common is it? to have a low blood TMB, for example, and a high tissue TMB. And so they can ask a question like that. And in seconds, Tempus One will surface a result. And what you'll find is we're -- including the citations. So here, I might have missed type there. So you can see we can include the citations here. as to where the -- we're getting the information from. And then if you go to next then, at this point now that we have this the provider needs to decide what to do for the patient. And so they're going to think about having a -- putting them on a combo chemotherapy and immunotherapy. So if you look at the RNA results here, we can see the patient has -- finds a rare NRG1 fusion. And so this is something that often wouldn't be surfaced on DNA alone. In fact, we've copublished that 40% of these rare fusions are found through RNA. So at this point, the patient starts on the combo chemo immunotherapy and chemotherapy. This patient also had the immune profile score, which you heard the team talk about and the result was an IPS low in this case. So now the provider says, okay, we have a low IPS. We have that rare NRG1 fusion, both are associated with poor outcomes so that chemo immunotherapy combination. And so because of that, the provider might think to order a MRD testing to now monitor how the patient is doing. So you see here, we've got this time line. We show all the results over time. And in this case, for Christina, the ctDNA, so the circulating tumor DNA in the blood starts to rise mildly. And then over time, at the follow-up time point, we start to see significant elevation.

So just to provide some context, sorry. So you're a demo environment because we have in the production environment or we -- someone shows up with like [indiscernible]. But essentially, what you can see here is that we have all this technology wrapped around the report itself. So we have the ability to take the clinical data we're connecting and have the system automatically generate a summary of that patient or a summary of their current therapeutic regimen or whatever is going on. We -- off to decide each one of these things in the main production system, you could click on these things and actually go into the raw note and read the note itself because we pulled the note out, we've identified it as a notice there. And then the system is just consistently as you make decisions that maybe you want to go from putting a patient on a combination of immunotherapy and chemotherapy to maybe looking at a target on the RNA side, it's keeping track of all that. So at the end of the day, you can -- it's going to get these cases that are going to be very complex and being able to summarize what happened, being able to ask questions becomes pretty powerful.

Yes. So in this case, they're tracking the MRD and becomes elevated, so that's when the provider might go order imaging. And in this case, then the CT confirmed progression. And so this is an example, again, where a provider might want to type in a question. like tell me if I'm now thinking about zenocutuzumab, they might want to ask questions like what are the adverse events associated with zenocutuzumab. Given that this patient has -- from that clinical time line that you saw on the right, this patient is having significant weight loss and questioning is that [indiscernible], is that related to the therapy? And so you see it quickly surfaced here now information with kind of linking out and sending you to more information. So it is a bit of a deep dive into a particular case, but I think the beauty is the way that all these results are accessible, we have summaries and it sort of anticipates the information that providers may need and leave them together and pulls it using the technology and the testing together.

Cool. So just at a really high level, this is also this technology just doesn't really exist other places. So other people have ordering systems where you can track and order and get a result. But here, it's all brought together, allowing you to contextualize it. And the challenge with cancer cases, and I think most of the things that kill us, heart attacks, stroke, cancer at latent stages. They're complex, the comorbidities, the amount of things you have to consider just complicated. And so as an oncologist or a physician, you think you're treating one disease, but pretty quickly, you're treating another or another complication. And here all that information is accessible. In the case of Tempus, you can order a whole variety of tests to begin with, you can track patients, but you can also communicate with the outside world and figure out like hey, if I put this patient on this drug, what's the most notable adverse event? And what does that mean? And how should I think about it? Maybe I'm going to bring somebody back sooner or dose them differently or be careful. And this is the reality of treating patients, right? You're treating a patient, but all of a sudden, they have major weight loss. And so you can't put them on the next dose of chemo or give them the next IO because they're just not doing well. And so being able to get ahead of that is super powerful. On that note, we're going to jump into the foundation model for a second. These systems, we began building software to kind of integrate all of this, that agent One that sits inside the system that allows you to basically bring all these different disparate health care data sets into one place is the same technology we use for our large-scale foundation model. Just for people that don't know, I'll provide a quick overview. About a year ago, we made the decision to build a large-scale foundation model in partnership with AstraZeneca and [indiscernible] providing the majority of the funding they may invested about $200 million to build this model. And then we began building it. The model is quite significant. It sits inside a cluster of about [ 1,080 ] fairly large compute cluster. We loaded in an enormous amount of the identified data across all these major modalities and we've begun generating insights from this data. We had to do significant pre-training significant compute as in run this cluster for like 90 days at 100% capacity or thereabouts, and then do a post training. We announced this morning some of the first insights from that model. And at a high level, what's amazing about this model is you're taking enormous amounts of multimodal data like BAM files at scale. Clinical data, we have billions of notes at scale, things of that nature and then asking it to predict what's happening and it's performing in many instances as well as super small, highly tuned models, which Kate will cover.

Yes. So as Eric mentioned, we're really excited now to kind of enter this next era. You've heard this morning already about the algorithms that we develop and put on top of our diagnostic tests. And part of the future promise here is that we can start to do that at a really broad amazing scale by using Foundation models. And so one example that we've been working on is just starting to take this multimodal data that you've heard about all morning, the clinical notes, the EHR data, the molecular data that we have, both DNA, RNA and images. And to combine that and instead of developing algorithms, you heard about IPS and some of the amazing tools that are already out there today, those were developed by traditional computational data science teams really coming through massive data and coming up with those algorithms and then validating them. Future state is that we're going to have models that will be able to surface those insights very rapidly in a more automated fashion, and then we'll be able to validate them quickly. We've started by just looking at very traditional biomarkers. We know that we already have good biomarkers like EGFR, [indiscernible], we can name a whole laundry list. So we asked the question, could we go even further and contextualize those patients. So in the IPS example, when you use biomarkers like PD-L1 or TMB, which are very well established, we can add insights on top of that and be able to separate patients who will do well or not do well beyond the standard biomarkers. Our model is now able to do that in addition across many different clinically relevant biomarker. So here, we're just talking about EGFR. This is meant as an illustrative example. You can think about any other clinically relevant biomarker and the model now being able to say what patients would do well or not well on sort of standard of care therapies. And so we'll be able to then also look at other biomarkers. So we looked at things that were already well known about EGFR patients. So P53 other co-mutations, other comorbidities. And the model is actually able to pick up beyond those standard biomarkers, other signs that a patient may or may not respond to standard of care therapies. So you can imagine a future where any biomarker, any test you just saw Hub and Laura walked you through that. This type of information could be layered on top of that for future state for physicians to really be able to get a more global and deep understanding of the patients that they're looking at.

Yes. So really quickly, and then I'm going to bring Ezra back up. So this is our strategy, comprehensive tests add on a bunch of alrimic insights that make those tests better, which we are doing today, which is driving our unit growth rate to be so high. And then the kind of next level of that would be run large foundation model at scale, have the system instead of generating one insight every 6 months or a year, generate an insight a week. That's just that people didn't know is the patient going to respond to EGFR inhibitor, are they going to respond to an ALK inhibitor, does the [indiscernible] fusion matter? Are they going to respond to immunotherapy, what adverse event is most likely, how long are they going to be in this therapeutic, whatever it is, generate those insights at scale analytically and clinically validate them, we have a machine to do that, put them into the report. And over time, you just become like it's hard not to get those insights because over here, you're ordering a test putting your patient on a drug, not knowing whether they're likely to respond or not and over here, you can, and we suspect that's the future. Whether -- we're the only company that can offer this or other people offer it. I don't know, but I'm 100% convinced that old world of targeted therapy will die and this new world of precision medicine will show up. On that note, we want to talk to you a little bit about how these algorithms are also affecting our clinical workflow. We have another product called Paige Tempus preview, which is essentially leveraging our digital pathology library to generate a whole bunch of insights. Some of those insights are making calls early. Some of them are making calls when a doctor can't get that information. So I'll pass it off to Ezra.

Thanks. Thanks, Eric. And so here, we have -- as Eric was saying, 2 examples of how we've leveraged Path AI, this Path AI into the diagnostics. The first I'll talk to you about is Tempus preview. And there are certain situations where rapidity of the results is critically important because the therapy depends on that result. And if you choose the wrong therapy, the patient could be harmed. The first example is MSI high. We know that these patients have a very high response rate to immunotherapy. And not only that, -- many of those patients will be on that immunotherapy for years and potentially cured. That's a result the provider wants to know right away because you don't want to put that patient on chemotherapy, you want to put them on immunotherapy. The same is true of EGFR mutations, especially in non-small cell lung cancer. Here is a situation where if you put this patient on immunotherapy, they actually do worse. So you want that result right away. The third example that I show you here is FGFR alterations in -- across several cancers, especially cholangiocarcinoma. Again, the rapidity of that result is critical. And so how will we address that problem? We've addressed it through Dig path. Here, through an H&E slide, the algorithm can be highly predictive of the presence of that alteration, whether it's MSI high, EGFR mutations or FGFR alterations. Giving the provider that quick response, this patient may have or likely has with a high degree of certainty and EGFR mutation. And while the provider is waiting for the confirmation through NGS, they can now select the appropriate therapy and get ahead of it rather than select the wrong therapy and potentially harm that patient. On the other end, it can be incredibly frustrating to providers and to patients to get a QNS result. There are situations where we just don't have enough tissue or the NGS testing for whatever reason fails. That happens in about 7% of the time. And they're really is very few methodologies that can get us below that 7% threshold. So we decided that we would address this problem in a different way. And again, bring in the capability of Dig path. Here, this is called Page Predict. And I show you this is a real-world example. Obviously, the patient's name is different, where we can use the Ditch path to tell the provider that there is a high degree of certainty that this patient has a specific alteration. In this case, it was a patient with cholangiocarcinoma, a highly deadly cancer, and that tumor contained an FGFR2 fusion. Those fusions are highly responsive to specific inhibitors. The result for the NGS came back QNS, just not enough tissue, but with the Dig path, we were able to inform the provider that there was indeed an FGFR2 fusion. That provider got a confirmatory test, and that patient was put on the right therapy with a high degree of benefit versus chemotherapy that was unlikely to work. And again, 2 examples of how we can bring in the multiple capacities and capabilities that we have to provide the best insight to that clinician to help that -- to help get the patient on the right therapy at the right time. Thanks again, Eric.

Okay. On that note, I'm going to turn over to Jim to talk a little bit about the financials and diagnostics.

Thanks, Eric. So I think that gives you a good overview of kind of what is the big driver in terms of our volumes, specifically on the oncology side. We've obviously experienced strong sustained growth over the last several years. In Q1, we had 28% volume growth in oncology, which was building out a very strong kind of accelerating growth rate throughout 2025. We have favorable ASP tailwinds that have led to kind of the revenue growth, and we'll hit that in a slide in a second. And then obviously, with the addition of Ambry and some of their outsized growth, given some of the share gains that led to additionally outsized growth in 2025. Here's just an overview of kind of trends in clinical oncology of volumes and ASP over time. I think the big takeaway here is that we had very strong growth in Q1, about 28%. As we look at April and May, that growth has continued in terms of the orders that are coming in, so they're tracking at a very similar pace. Again, highlighting the durability of the growth from a volume perspective. And then again, we've seen ASPs tick up over time. We've talked previously about this path to kind of achieving $500 of incremental with the announcement this morning of xT [ TO ] CDx being approved, that allows us to capture that $200 at the beginning of 2027. So we're on track xF is sitting with the FDA currently that was submitted earlier this year. So it won't impact ASPs in 2026. But as we get into 2027, that should be accretive as well. And then there's also commercial coverage over time continues to tick up. that's not kind of a flip of a switch, but we will chip away and see improvements over time there as well. On the regulatory side, we saw growth rates moderate in Q1 which was anticipated given some of the large kind of share gains that they had back in Q1 and Q2 of last year. We would anticipate similar growth rates in Q2. But as we get in the back half of the year, we'll see that acceleration of growth in the hereditary business again, as we've done kind of lapping some of those share gains. And lastly, we've talked previously in this 25% growth rate over the next 3 years. that would put us at about $1.9 billion, just to give you kind of the size and scale of the diagnostics business. We provided a rough breakdown of where that's coming from primarily in clinical oncology. But hereditary obviously moderating back to kind of the mid-teens that we had talked about previously. And within that 3-year period, there's going to be periods where ASP may outpace and you may be growing faster than 25%. But what we really want the takeaway to be is that this business is really set up for durable long-term growth. And on the right is kind of the list of initiatives, both near term, which are all being executed on today, but then also kind of the longer-term growth drivers that will allow us to move beyond that for 3-year period. So with that, I think we're going to do some Q&A. I know that they are going to grab some mics to walk around.

Yes Dave Westenberg from Piper Sandler. I wanted to talk about the trajectory model. So the AI model predicts how patients will do over time, significantly outpacing the models at predicting survival. You've demonstrated this across 3 cohorts of famous studies. I believe they are all in lung cancer. Two questions. First, of all 3 of these done were done using historical patient data. What's the plan for using this test on data outside the institution, a hospital or registry that Tempus didn't generate to prove this works in real-world samples. Secondly, and more importantly, at what point does a pharma partner move? This as an interesting research tool to actually make a go or no-go decision on multimillion dollar trials.

You want to start with first and I'll cover the second?

Yes. Yes. We kind of -- we're going to talk more about this actually in the data section in life science, but that's okay. You jumped in Yes. So first of all, just in terms of how this model compares to other more traditional methods that you're mentioning and reading from the press release. So this model, we have many different models. We'll talk more about that, but the model that we're highlighting there is what we call a patient trajectory model. It's actually able to look at patients over time, which is one of the benefits that we think we're really excited about. One of the things we were highlighting is that it's very good at predicting outcomes. That's sort of for that particular model, it's able to look at outcomes, in particular, survival. And so you can start to do interesting things like take cohorts of patients that are very well known and understood from more traditional methods and ask how well the model can uncover other prognostic factors that might change survival. And so that is a use case that can be used. You're right, for clinical trials, and we'll talk a little bit more about how pharma partners can think about that. But it's also a space where we can then start to validate against sort of well-known and understood trials that have that outcome data and then be able to move into new spaces and ask new questions of new cohorts. So that's the way we're thinking about it. In terms of validation, you asked how would we validate this with other sites or other institutions. That's absolutely part of the plan. So we have a vast and broad network of providers and institutions that we work with and a vast network of pharma partners and life science partners who have their own data, both retrospective and perspective. And so the future state will be taking some of the models, asking questions and then working on validation in a lot of those different spaces. So good to up for what's going to come later.

And we'll talk about it in a minute. But at a super high level, we generated inside these foundation models. One main large foundation model, lots of micro models. You generate the insight and then you essentially have to figure out if that insight works across data that you have that you didn't use to train the model. We're fortunate that we generate so much data that we have a huge bolus of data we've used to train the models and a large bolus of data that sits off to the side. So we can make a prediction using that data and then see if it holds up in other data that we have. And then once that's true, you know you have something that's working, then you go seek to basically go to third-party sets and validate it. Diagnostic insights will essentially, for the near term, live on our diagnostic tests because that's where they live. And life science insights or biopharma insights are already being used by the people that have access to these models. There are half a dozen biotech and pharma companies today to have access to one or more of these models that are using those insights to interrogate their R&D portfolio, design more intelligent Phase II, so on and so forth. So they're already being used. And we'll talk about that insight. I don't know you want to call people on it, [indiscernible].

Kyle Mikson from Canaccord. Great stuff. I want to ask a multipart question. So first, on the LRP, I guess. I know you've been talking about the 25% for a while, but the Street's at like 20% revenue CAGR for Diagnostics. So maybe just talk about what we're missing in terms of ASP volumes, MRD and AI and rents up? And then secondly, maybe just talk about the diagnostics M&A kind of strategy. You haven't done one since Ambry, but you're having Personalis 15% or so. Maybe just talk about like how you're viewing dilution versus growth in that segment in terms of M&A?

We can both cover it. I mean, I'll start. I mean I don't follow the stream. Some of the models I've looked at essentially are they have like high growth rate in '26. There was a high growth in '24, low growth rate in '26, '27. Then they had high growth rate in '25, low growth '27, '28. Now they have high growth rate 26, low growth rate '28, '29. So they're just nonsensical. I mean they're essentially saying you're growing really fast, but one day you won't. And is there any logic behind that? I have no idea if what that logic is. We look out at our portfolio. If we're going tell the world, we think we're going to grow 25%. We don't want to look stupid. So we have to believe that like we really think that's going to happen. So there is no street model that I would look at and be like, oh, they know something we don't know, we have more information. And at the present moment, we believe we're going to grow at 25% roughly. Some of it will be -- our volumes are pretty healthy right now. We're growing in the low 20s, so that's awesome. Some of it's going to be ASP lift. We just the one big -- you had kind of 2 big drivers of that. Roughly half of that gain showed up this morning when we got FDA approval for [ TO ]. That's $200 of lift across a massive number of tests. I don't know the number, but it's probably $750 million of gain. So that's a real number. XF when that's approved is another big piece. So part of it's ASP lift, part of it's volume lift. And then you're launching new tests and other things are happening. There's always pluses and minuses. Everything doesn't grow equally up into the right. So some tests will overperform, some will underperform. But as Jim mentioned, the benefit of having this kind of a comprehensive portfolio is we're big enough to absorb that and still deliver that 25% per year. In terms of MRD, we chose personnel because we thought they had a great test. We invested in the company, obviously, as of this morning, that's been a great investment. I don't know how much money we made, but it's a lot of money. And at the end of the day, they've been a great partner, and we're executing that strategy. In terms of whether or not at some point, we'll look to consolidate those companies. That's obviously not for this meeting here, and I don't have a good answer for that anyway. But I will say this, our strategy of not needing to run every diagnostic is the right strategy. like I know it sounds a little crazy to diagnostic investors, but I promise you, Amazon doesn't make every single product. So -- and nor does Apple make every single app. There's going to have to be technology platforms that take the U.S. health care system, which is very complex and translates it to physician care and patient care across the board and the companies that do that can't do every single thing themselves. They'll have to find ways to partner with third parties. So we've always been focused on not just running tests and well but figuring out ways to make money partnering. And as I mentioned earlier, our net unit economics today are as good with personnel as they would be if we own the company. So if we -- one day own the company, all you get is revenue gain. -- you get no net income gain.

Great. Casey Woodring from JPMorgan, and thanks for hosting us today. Can you talk a little bit more about the attach rate of xF and the tissue test. What's the current attach rate there? What percentage of these cases are reimbursed for both tests. You mentioned that I think 9% of patients had unique actionable alterations that were found in xF that weren't observed in xT. So just wondering what a payer would say to that. Is that something that would preclude them from paying for both tests? Just how should we think about that?

Yes. So in terms of the attachment rate, it hasn't changed over time. I think we've published previously with around 25%. It largely has stayed intact over the years in terms of the number of physicians that are ordering it. And reimbursement depends on, obviously, the payer we're fortunate from a reimbursement standpoint that we're in this period where we're seeing expansion of reimbursement. And so there will be some tests that don't get reimbursed, but we're still going to see a net add to the overall reimbursement since we're not at parity with our peers. So we think we're well positioned to continue to win in this space. It's another thing that arms doctors with additional information that is incredibly useful, and that's why we offer.

Yes, it's also worth noting, when we went public 2 years ago, people were like, "Oh, you're running DNA and RNA, and it's not going to get paid for it like it felt kind of like you guys are on the edge". If we were on the edge 2 years going to be public, we're now in the middle of the bus like maybe getting toward the back of the bus. I mean you've got reimbursement rates from our competitors that are 2x ours. You've got these portfolios being rolled out where we've got competitors that are like, click this button and you have 12 tests. They just will keep showing up forever. So like we are not cutting edge in terms of like, hey, order a bunch of stuff and is it going to get paid for. Like there are people that are way further ahead of us that are driving all kinds of unit growth rate by being aggressive. We view ourselves as not being aggressive. We view ourselves as being comprehensive, but not aggressive and that's where we want to be. We want to give -- we want physicians to be able to like logistically make a decision and order things in an administratively intelligent way, but we never want them ordering test they do want and we never want to bundle in 5 tests like a year when they don't really want that. So I think to the extent we were on one end, we're not.

If I can make a comment, Eric. You don't know who that 9% is a priority. That's the other thing to keep in mind when you're thinking about reimbursing these tests. As a provider, I don't know who falls into that 9%. So I have to order both in order to get that answer. The same is true with RNA. I don't know who's going to fall into the that's only going to be surfaced by RNA. So I have to order both. So logically, it makes sense to reimburse. And the trend, I think, is -- I just want -- certainly on the RNA side, that bus cannot be, that training has left the station. We could argue whether or not there will be rules over time about how often you can order MRD test how often they're going to get paid for, how often treatment response modern is going to get paid for. I do think that space is going to be -- as I mentioned earlier on, I don't think fully like all the puzzle pieces are in the right spot. So I don't know where that's going to land. But the RNA trains left the station. It just matters way too often. We're in the middle [indiscernible]. Our competitors been a bunch. You come back 5 years from now and there will be dozens and dozens and dozens of RNA-based therapies that you will need our express data for. We'll do one more -- we have more Q&A, so we'll do it in rounds.

Kallum Titchmarsh with Morgan Stanley. Maybe just on the oncology business and the volume growth you're seeing today, could you just break out that growth between the existing account penetration versus kind of new adds? And then I guess, have you seen any examples of physicians switching to Temps' tests as a result of the technology infrastructure behind it like Hub and Paige Predict?

We'll start with Mike. Any thoughts on the new versus existing?

Yes. I mean we've held pretty steady in terms of addition of new clinicians and new ordering systems. We monitor this really, really closely. So we look at what is the kind of reorder rates over time and then how do we actually add net new physicians that have never order with Tempus or physicians that had previously gone sale has had an order with us within a 12-month period and they've kind of come back in that. Number continues to, at the very least, hold study, and we've actually seen some modest growth in terms of that new ordering position base. So there's really kind of 2 vectors that we're seeing this growth come from. One is in this kind of addition of new customers and two is in deepening relationships with our existing customer base.

I would just add one quick thing. There's only 14,000 oncologists in the U.S. And so when we -- Eric talked about, we think about 50% of oncology as actually order that. So it's always a combination of continuing to have the physicians identify more patients that should receive this type of testing and then tapping into that untapped market of folks that aren't ordering at all. In terms of some of the new products like Paige Preview or Tempest preview, I think we just are deploying these things now. We acquired Paige, I think maybe 6 months ago or 9 months ago or something. And so it took us a while to get these data sets aggregated and bring some of the benefits of those products into our platform. This QNS thing is not small. I mean it's just -- there's no way to solve that problem. We have that problem. All of our competitors have that problem. You just occasionally don't have enough tissue. occasionally, alumina sequencing process just doesn't yield the results you want. So being able to make predictions so no patient is left behind is pretty powerful. On the other side of that, there's just a certain number of diseases where doctors want answers in 1 or 2 days. And we will be the first people that can offer that at scale. And so that's also powerful. So none of that stuff is currently showing up in our unit growth, and I suspect it shows up over the next, I don't know, 3, 6, 9 months, 12 months.

Let's jump to the data business, and then we will come back to Q&A, if you don't mind. Okay. So to be sustainable, 2 main businesses, a diagnostic business and a data business. Our data business is the one that I think for a lot of diagnostic investors is like unfamiliar. So we're going to spend some time trying to walk through like how it works and why it's growing and so successful. First of all, if the question is whether or not data and AI and technology are going to permeate drug discovery and development and health care, the answer is like 100% yes, it can't not happen. Every industry whoever has said it's not going to happen has been washed away by technology. And I just give you one example, like you can go back to 1960, 1970, when people were like trading on the New York Stock Exchange like orange futures and would have bet you their life that this could not be replaced by technology and it's all been replaced. So like that's just the unstoppable nature of technology and health care will be a beneficiary of that, but it is coming. In our case, we have spent the last 10 years really building the piping to generate a healthy and sustainable data business. That piping is all about how do you pull data out of the U.S. health care system at scale. How do you combine it with something else that makes it super interesting molecular data, how do you produce an insight? And then how do you package up that insight. If you think of it, we've got 2 end customers. We have to package up an insight for a doctor. Your patient is not going to respond to this particular EGFR inhibitor, you should know that. [indiscernible] it up for biopharma company is far more complex. So this is -- it's not -- it's typically not a simple answer. They're designing trials. It could be novel discovery. There's a lot going on. So you have to give them the tools to interrogate this data. And in that regard, we stand on. If people are wondering when we started the IPO process, 2 years where we went public and people -- I think if I would have asked 9 out of 10 investors, they would have never thought our data business to be this big and they all would have thought we'd have met massive competition. Both of those are -- have proven the opposite. Our data business is big and growing, and we have almost no competition. And it's really the technology and tools we that wrap around this data, the connected platform, the analytic capabilities, the ability to deliver data at scale, interrogate the scale that's unique. And this data comes from many sources. You can't just be a sequencer that generates DNA and RNA data and be like, I'm in the data business. We have data coming from our care gap products, our clinical trial matching products, our real-time clinical trial matching products our AI tools and technology, our diagnostic business, our radiology products, our pathology products, our cardiology, we have many, many ways to get data, which allows the data set to be real, contemporaneous. And it's also -- it's part of this network effect that fuels our diagnostics business is also helping our data business. The more data we collect, the more insights we generate, the smarter our platforms get, the more people want that data, they're making -- they're licensing it, which allows us to invest in those tools and it becomes this really positive virtuous cycle. This is used across the entire competitive of decision making. What I think a lot of people understand is like, why are people licensing your data? And Ryan is going to get into some use cases and the tools around it in a second. But it's really the entire R&D, the discovery and development life cycle. Do I have the right target going after the right indication, do I need a biomarker, how do I design my trial? These kind of questions aren't worth $1 million or $5 million to a big pharmaceutical company or big biotech, they're worth hundreds of millions. You get it wrong, you've got $1 billion failure. You get it right, you've got a $10 billion franchise. Our data and our modeling tools are used really across every aspect, from early-stage R&D through clinical development, now into commercialization, we just touch our products touch really anywhere you could use data or AI or modeling to help make more intelligent decisions where they're certainly not in oncology at scale and will be the other disease areas over time. Ryan, do you want to jump in?

Yes. So this is a quick snapshot of the platform. And like Eric was mentioning, we've been licensing data to our biopharma customers for several years now, and we have been in the business of licensing multimodal records for longer than anyone in the industry. This is -- when we say multimodal, we mean combining DNA, RNA, treatments, outcomes, images so that you can really understand what's going on with these particular patient populations in the real world. And so this is the kind of a snapshot of what it looks like, but I actually rather just show it to you so you can see what it looks like because to many in our space, analyzing multimodal data is not easy. And so acquiring the data is one feat, but organizing it under a common data model, being able to make it useful for people that are coders or noncoters is essential to get -- to turn data into insight. And so this is a platform that sort of is the front door of our data set. And so our customers, we build these data sets for them and they can interrogate that data through this system. We've now actually embedded AI into every step of the workflow. And we'll actually -- I'm giving you an early preview that we're announcing in a few days here about sort of the new launch of Lens with these AI tools. And so for every step in your journey as a user of the system, we have Tempus One as sort of as a copilot or coscientist to really help you on your journey to generate insight. And so one thing that I can do is I can start to build cohorts from simple natural language. This is an essential step in order to make sure that the cohort is sort of built fit for purpose. And so while this is building, you can start to see that we have various other data sets and projects that I, as a user, have already created. So things around particular either tempest data sets, public data sets things of that nature are all in this space so that I can start to compare and contrast different cohorts over time. And so what you can see is that we can now get deeper into things like not just a particular biomarker that's known today like KRAS, but I'm building a cohort for lung cancer, adenocarcinoma. For those patients that were treated with first-line therapy that have a KRAS mutation and then quickly identify we have 3,000 patients already in the system. I can now either save this query, I can adjust this filter, and I can quickly start to refine this cohort over time. And so what you're seeing now on this left-hand side is that my copilot help me build this cohort, but I can take the steering wheel and actually refine this further with these filters on this left-hand side. Each filter is essentially an inclusion or exclusion criteria that our pharma customers are thinking about, right? So they start with a population of interest. And then they're trying to understand for the particular population that I'm going after for my trial, is there patients in the real world that I need to better understand, again, with these kind of this combination of DNA, RNA treatments and outcomes. And really the essential step is that you have to be operating at scale to get to the bottom of the funnel that is significantly powered thousands of patients. So having millions of patients at the top of the funnel is essential in order to really get to something of real interest. This system allows our users to be able to visualize these different modalities of information and interrogate this information in much more granular ways. And so for that 3,000 patient cohort, I can start to see basic things like, okay, what are the demographics of these patients, the distribution of age, things of that nature. But I also maybe want to understand co-mutations like Kate and Ezra were mentioning as well, and you can start to see that not just in terms of prevalence, but you can also start to run feasibility on treatments were these patients given of the 3,000 because I may want to select some of those prior treatments as part of my inclusion criteria. Now one of the things that our customers are doing all the time, and the question that we got earlier on was how are our customers using this for trial design decisions. And one of the most important aspects is to make sure you have your patient selection strategy correct, which means am I going after the right patient population with my drug or not, right? And so one of the things you can quickly do in this system by just -- even by clicking a few steps here. I can start to compare cohorts over time. And so I can start to look at other data sets that I may have added, I can start to look at that -- those changes and start to build these kind of graphs on the fly. I can go a bit deeper as well, and I can look at even things around the things that we only Tempus can provide, something like co-expression analysis as well. And so I can start to dig a little bit deeper, and we allow our users to get to this type of insight in -- literally in a second or in days. And so looking at things like beyond EGFR and KRAS that we know of today, but looking at novel biomarkers that are coming in the future, like MTAP deletions is essential for our drug developers. And so this is sort of a detailed breakdown of that sort of data set that I've already built that's looking at not just the KRAS mutations that were treated in first line, but looking at something specific as MTAP biomarkers and what that is doing to the various kind of behaviors in the real world. I can look at the distribution of gene expression. I can look at sort of the correlation or the pairwise expressions between not just one marker, but 2 markers at a time. So looking at MET versus EGFR, MET versus KRAS, MET versus MTAP. These are the kind of the various iterations of questions that our users are going through, and they can get that insight all within this tool. Now this is useful for people that -- even if you're not a coder, but we also have connected our data to a computational platform like our and Jupiter notebooks to be able to go even further for those that actually want to code, right, and actually have that capability. And so again, we've embedded Lens or Copilot here from lens to be able to help people ask certain questions around how do I create an onco plot, looking at the most top 10 most frequently altered genes. I can quickly do that, figure out sort of look at what tools that I need to call but I can also start to see the actual code that was written. And I can pull that up here. It starts to run the actual sort of analysis in my environment. And again, we're giving the user the control here. The code is populated, but they can actually edit this code no different than their -- what they would do in our environment. But the most important thing is that I can quickly get to insight I can refine this, and I can have my files generated instantaneously. And so these types of outputs are the things that really drive our business today. And so maybe we flip back to the slides. Being able to have that go through those iterations very quickly get to these types of outputs is really the first step in a multistep journey for our drug developers. So if we can switch back to the slides. I can kind of then cover some of the other aspects. All right. So I already walked you through sort of the kind of the step 3 steps query, building data sets through natural language, using agentic workflows to be able to analyze this data in much more granular ways. And one of the things that we're really excited about is connecting this rich multimodal data to the compute environments that we use to actually train our foundation models, but actually connecting compute infrastructure to help our customers build and fine-tune models as well. And so one of the things that we do is not just helping understand different co-mutations and helping early development also are helping clinical development in late-stage development for those high-risk, high-reward types of decisions like a Phase III global trial but even growing bigger part -- growing part of our business that is addressing what is happening in the real world is really helping those commercial and medical affairs teams around better understanding of things around clinical care gaps and things beyond that. But I want to spend most of the time maybe addressing that question head on around like what are our customers doing and what are they getting out of this type of unique data set. So I'll walk you through 3 examples. The first one is a global biopharma company that really was interested in advancing their immunotherapy franchise. But they really need to think through, could they uncover additional novel biomarkers in a particular patient population. And so here, for this particular project, we were able to assemble a data set that really didn't exist in the world. A 5,000-patient data set where we had biopsy samples and DNA and RNA segmenting pretreatment and posttreatment. That type of data set will allow us to figure out what these particular immunotherapy treatments are doing to tumors that ultimately are leading to different outcomes in the real world. So again, this data didn't exist before, but what it allowed us to do is uncover 4 different novel targets that we're able to advance their drug discovery pipeline. And so they spent time and effort and money investing in these types of projects. But one project alone, and if you just think about like the return on investment on a single asset in an immunotherapy franchise, we're seeing ROIs calculated by our customers. In this particular example, be 30 to 50 of what they spent. And this measurement of value is kind of the common, I would say, motion for us in our collaborations, so that we cannot just make sure we're delivering value now but also it's why many of our customers have expanded with us over time. The second example was exactly the question you asked earlier, which is the most sort of significant investment decision that these companies are making is a go-no-go decision in refining the trial for a Phase III global study. And so here it was a different global pharma company that was faced with this kind of critical decision. And so here, we were actually trying to better understand what was going to be the comparator arm? Can we actually understand standard of care and make sure we establish a good baseline for what those patients are facing and how their outcomes are performing today, but also to make sure we're stress testing the inclusion and exclusion criteria is not just based on clinical measures, but also looking at can we understand the tumor biology of those patients to make sure that there isn't a heterogeneity or surprises in our Phase III study. And so again, in colorectal cancer, this is essential. And so we were able to be able to not just deliver this type of insight with our biopharma company, but it is essential to be able to derisk a decision and ultimately creates a net present value for our customers. of somewhere north of $500 million. And so you can start to see each project starts to stack up from an ROI perspective and how we build over time. The last piece is really another global biopharma company that was faced with a slightly different decision which is really around not just investing in a global Phase III study, but do I go first line, do I stay in the second line, that type of decision is a high-risk, high-reward type of play. And so again, we start to look at what data sets do we have, what multimodal data sets we can build so that we can analyze the molecular distributions of patients that have high PD-L1 versus low PD-L1 because this ADC sort of decision for that particular drug was going to go up against that type of landscape. And so we've built the data set, we worked with the teams, and we were able to derisk a number of the decisions by building these patient subgroups but also the go decision was made to ultimately allow that customer to proceed. And we talk a lot about like probability of success, but we've been in these collaborations for long enough where our customers have actually seen success. We've seen approvals for the programs that we've supported and that makes it not just a perceived benefit, but an actual ROI metric that ultimately leads to why our business has grown over time. So with those 3, I pass it over to Eric.

Great. Thank you. Okay. So at a high level, I just want to cover this kind of scale and scope of our data business. We're connected to -- I mean we're working with something like 19 of the 20 largest pharmaceutical companies in the United States. That number has helped -- or that metric has held pretty constant over the past several years. The good news being -- we're still working with all these folks. We work with over 250 biotechs. We've signed in excess of $2 billion worth of data licensing deals. Our revenue last quarter on the data licensing side was $87 million. We've now large partnerships in place with not 1 or 2 big pharma companies, but lots. and This number continues to grow as we sign more and more of these kind of multiyear $10 million to $20 million or $30 million a year engagements with folks. And we also delivered an enormous amount of data. I think as one of the things when we talked about the Lens platform, this is -- these capabilities as they scale not just to generate data, structure, harmonize, clean it, analyze it, but also deliver those insights, including the raw underlying the identified data to biopharma. When you think about delivering petabytes of data, it's just not a small task. Ryan got into a bit of the ROI that we are used to measure, but increasingly, folks are looking at this probability of technical and regulatory success and whether or not we're actually generating ROI if they're making investments in their licensing a $20 million of data, are they generating $60 million or $100 million of gain. And we've been through rigorous analyses over and over again with people who are increasing their spend where they're roping in finance they're hoping in [indiscernible], they're roping in other teams to validate the return on this data licensing and over and over and over again, it comes back that this is accretive. And so people increase their spend. We have a long history now of people increasing their spend, which we'll get into in the next slide. This is typically how it works. People very rarely to somebody say, "Hey, I'd like to sign a $100 million deal and license $25 million a year of data for the next 4 years". More often than not, they start with, I'll do a $250,000 project. $0.5 million project or $1 million project. They get the data, they have to try it and test it. Multiple teams are involved. They have internal computational biology resources and bioinformatics resources and biostatistical resources R&D teams, and they interrogate this data and try to figure out, is it good? Is it clean? Can they use it as a representative. So you have to get through all these hurdles before you get to the next project and the next project. We have a long history of going from one program, one project with one asset to multiple assets to multiple assets over multiple years to expanded partnerships and ultimately strategic partnerships. As I have said for a long time, you can -- we -- our pricing model is similar to the large cloud providers, AWS or GCP or Azure. You don't have to sign a big deal to be on Azure or GCP or AWS. You can spend $100 a year with AWS, and they're happy to have you as a client or you can spend $1 billion a year. The only thing that you gain by making a longer-term commitment, both in terms of years and dollars is a reduced price. So if you think about all the people that sign these multiyear, very large agreements with Tempest, the only thing they're getting is a discounted price, meaning they value the data so much, they're willing to make a multiyear commitment because they want to save that money. So I think it just speaks to the value of the product we've built. Obviously, here's a great example that our first strategic partnership with a big pharma, we have several strategic partnerships with big biotech. But it was AstraZeneca that was signed, I think, in 2021. Obviously, that relationship is going strong, and we -- they're funding our foundation model, and that project runs for the next, I don't know, several years. And so we've got a long standing relationship with AZ. GSK was another large partner that came on board. There's a few years left in that agreement and then Merck recently signed up as another large farm who came on board in a strategic way, and that partnership is just starting to kind of grow and prosper in every which direction. And it just speaks to the fact that the biggest cancer companies increasingly realize they need our data to do all the things that Ryan talked about a minute ago. If you look at the -- just the -- we'll get into some of the metrics of the business, but in terms of like overall relationships and concentration, work with about 240 companies in 2025, and that number should be up in 2026. We work with 35 in 2020. So over a 5-year period, we went from basically 35 people in total licensing our data to 240. Back then in 2020, 85% of our business came from our top 5 clients. Now it's 59% and shrinking dramatically. That number is just kind of on a free fall down. So the good news is the business is diversifying itself over time. Increasingly, people don't just want our data, they also want models. Our -- it's actually what we call it data and applications, but in reality, it's not a great name because what's happened to us over the last year is it's rare that people just want our data. More and more and more, they actually want models. You can almost call it modeling and applications. Data in and of itself is interesting, but in a world of large multimodal models, which all of our companies have some exposure to, they want to know how they can use this data to build their own proprietary models or take their proprietary models and hyperscale them with more data and actually figure out how to build something that's proprietary and advantageous to us. Almost every conversation we have now is a blend of license some data and use our capabilities to build models, build those models on the lens platform. We have both CPUs and GPUs connected to that platform at scale or we'll partner with you to build models in some way, shape or form. We're going to -- I'm going to bring Kate to talk a little bit about the benefits of the foundation model on the biopharma side. But again, this is a huge cluster, a massive amount of data that's producing insights. Some of those insights have therapeutic relevance. Some of them have research and discovery benefits, and Kate will cover that.

Yes. So we talked earlier about -- a little bit more detail about our model. And actually, when we say model, we're really moving quickly towards many models. So you can imagine an ecosystem of these models. And then layering on top of that, things like agents that can leverage the models, tools and capabilities, you just saw a Lens. And so we have things there. We mentioned the coscientist type of approach. So future state is that this is moving towards a platform. We talked about it earlier from a diagnostic perspective and how that can help us uncover insights that can become algorithms on top of tools. The same thing is true here for our biopharma partners. And so the ecosystem we have today already includes multiple models. So with the acquisition of Page and their team coming in, we already have several foundation models that are very good at using images and being able to look at different signatures or outcomes. You heard a lot about that from the diagnostic component this morning. You can now imagine how a biopharma company might want to use those same tools and technologies for their trials. So rather than having to run sequencing, they can now use an H&E image to understand which patient that they should enroll and to pull those in. In a similar way, we're actually able to take all of our clinical data that you've heard about this morning and to be able to leverage that and pull that into a model. And so when we talk about multimodal, what we're really saying is that these models can now incorporate things like clinical notes, clinical labs, clinical images on top of the molecular data that we've spent 10 years being able to build on our platform. So when others talk about multimodal, they often are thinking about 1 or 2 modalities. When we talk about multimodal, we're talking about a really large library of unimodal models that we can then combine and start to fuse together into true multimodal. And so the future state that we'll move towards, as you see here, just kind of building upon lots of models that are very focused and specialized. So yes, we can have a genomics model that combines DNA, RNA TCR sequencing, BCR sequencing, we can also then combine that with the clinical model we just talked about in terms of patient trajectory, thinking about patients over time and what's happening to them in the clinical space. And then we can layer in things like images and others. So as we think about how to actually make this useful, we gave some examples already in terms of use cases for our pharma partners. And in some cases, as Ryan showed really nicely, computational scientists, both within Tempus or in our partners are still going in and doing a lot of that work in a more manual way using kind of standard machine learning and data science future state well that they may be able to just ask the model a question and the model will be able with agents and other workflows to be able to produce that analysis. And so what we're really talking about very quickly is the ability to uncover insights that would take perhaps weeks or months to generate to now be able to do that in a very quick fashion. And then to leverage those insights, of course, you will need to do follow-on work to validate them and show clinical utility and make sure that they're actually correct. But the workflows here will speed up all of that process and will allow pharma companies to ask the really critical questions that we just talked about in terms of what type of INE criteria should I think about? How do I design my trial to make sure I stratify patients appropriately for other factors that may impact the outcome of the study? And so we're really excited about -- we've really reached a moment where all of these things come together. They are helpful on the diagnostics side for providers, but they are also helpful for our life science partners.

And again, I think the -- and we've released this morning cover some of this, we've also -- papers coming out that go in much greater detail about how these models predict what they predict and you can take a look at it. But ultimately, we've crossed the major hurdles we had to cross when we entered into the first foundation model agreement with AstraZeneca. They had essentially 2 criteria we had to meet. One was a CNX score for an open for a trial that was it was out in the public while there was a trial that they had data for. We didn't have data for, for which they had trained a very specific model. And so the question was both, could this large scale model replicate trial outcome data that's publicly available and privately available, both where there's no model that's predictive and a highly tuned model that's predictive, and that was the hurdle we had to get over. And at the time -- and that hurdle was not seen as being like easy to get through given this is the first time we were building a large-scale multimodal model in oncology with all of our data. And models just get much better over time. Think about that ChatGPT 0.1 versus ChatGPT 1 versus whatever for on now. And so the fact that it performed this well this quickly, I think, is an indication of what is to come. once you have these kind of models performing at scale or insights permanent scale, you end up saying yourself, okay, and this is the last part of our business is what do you do with them? How do you distribute them to the broader ecosystem. And so we've long focus not just on the diagnostics side of the business and the data side, but also the application side. How do you take these applications or algorithms and distribute them broadly, given that we have this connected ecosystem of 5,000-plus hospitals we're in a unique position to be able to distribute AI into the U.S. health care system at scale in ways other people can't. And there's all kinds of questions that people are answering every day. we know what critical biomarker should I target, what are the therapeutic options I have? What clinical trials my patient eligible for, did I overlook something. And there are also questions that they're not asking like ambient in the background is a mistake occurring that no one knows about where care gap is being kind of broken. And so we built technology once we had these connected rails, and we had date flowing in and out of all these hospitals, and we had the ability to kind of in real time, take that data in, generate insights and put the insight back into the hands of a provider, we chose 2 starting places to focus on. One is, could we use this technology to match patients to clinical trials? And can we use this technology to close care gaps. The third is could be used this technology to develop entirely novel algorithmic diagnostics and distribute those. In the first 2 clinical trial matching, we call Time and our Care gap program we call NEXT. And both of these things are operating at scale. They're operating at scale, they just don't generate lots of money, which I've said many times. But they operate at real scale. This is not like it's not like, "Oh, we've got a few people using these things". We have many, many providers using them multiple care gaps deployed, millions of patients being screened. We are enrolling lots of patients in trials. We are closing lots of care gaps in real time. these things operate at scale. If they were paid for, like I suspect they will one day be paid for, this would already be a large business. So we're fortunate that our 2 main businesses, diagnostics, data generate enough gross profit and enough investment dollars we can invest that we're able to really lean into some of these forward products like our applications business that we think will one day be quite big. In cardiology, 60-plus algorithms deployed across multiple conditions in oncology, a whole body of algorithms. And in radiology, we've got a few in markets today, including our module that operates at scale. And then in terms of clinical trials, we have dozens of providers enrolled in our program. I can't remember, it's 80 or 70, some very large number of providers that spans 1,000-plus, 2,000-plus oncologists. And we have, at any given moment in time, a nice basket of trials that we're able to enroll patients in, in a rapid manner and this program is starting to really scale as we are moving from the space of kind of -- we've been over the last 3, 4 years, proving that it works. And now all the conversations are about enterprise engagements where big pharma is like, "I want to give you 15 20 trials". And we now are in a position where we're like, it's too many. So we're now fortunate that we're able to say to people like we can't take your 15 or 20 trials, which is creating a really unique market condition. All of these technologies that allow you to scan clinical and other forms of data in real time and then generate an insight, essentially are part of this future where there will be ambient AI-enabled copilots that exist in the wild that essentially make good doctors, great doctors and great doctors superhuman. And that is essentially what these products are designed to do. They're designed to kind of be there in the background, structure all this data that historically was unstructurable generate these insights and make sure that they're enhanced people. This is -- not only does this require a connected ecosystem, which we have, which we've invested enormous amounts of time and money in building but requires really intricate technology. We have an entire technology stack surrounding this products like Edge and air and Locker, which are enormous in scope. We can allow a health care provider to give us access to this data without it losing security and protection of that data. We can bring in different forms of multimodal data that don't even reside in their EHR DICOM files, CT scans, pathology slides, other forms of data, like, for example, when you get a 12-week CG, it sits in a different way fall from GE is not even in [indiscernible]. So we can basically with our edge server pull all these different data forms in, structure that data and then airs our technology platform that allows us to basically take the insights and put them back in the hands of doctor through Epic or through other HRs that they're using. So all of these technologies are necessary to kind of pull off this, how do you listen in the background, find an insight and deploy it. In addition to our applications business that is growing rapidly. We also have an algorithmic business, which we call algos. We believe that in the future, these algos will be pervasive. We've talked about that historically. We are now not alone in that conversation. We have, over the last several months, I think, been dragged to Washington several times. People have been here. I think the entire world now is thinking about how do we -- how are we going to manage a world where we don't just generate wet lab diagnostics, but dry lab diagnostics, how do we make sure these things can get validated and paid for, how do they get ordered, all of that. They can't be stopped. There's going to be more of them. Patients are going to want them and the U.S. health care system will have to adapt. And we want to be front and center in that change. We started in a couple of areas where we're most pervasively engaged. One is pathology. We have a variety of pathology, algorithms, some which are FDA approved like our Paige prostate Algo, which actually has FDA approval. Others that are in flight with the FDA, for example, our pan cancer suite has a breakthrough designation is going through that process now. We have a variety of cardio algorithms we've talked about historically, 2 which are FDA approved are ECG-based atrial fibrillation or [indiscernible], algo and our ECG-based low ejection fraction or low EF algo. Many more in flight, many more coming. We envision a world where Tempest is able to basically generate hundreds or thousands of these algorithms. We are convinced they will one day be paid for by the normal reimbursement process. And this will eventually be probably the largest of all of our businesses by far. I'll go through one use case, which we talked about a little bit historically, which is the ECG-based use case because I think people kind of understand maybe oncology or pathology, but this just go to a new area. So obviously, heart attack is the own killer of people in the United States. One of the most common diagnostic insights you would get is from an ECG. This is the thing that's kind of pervasively in primary care. Especially for older patients, you go in and you see your doctor and you might get this as part of routine care. And yet this test that comes back is effectively wrong 3% of the time. So we run a few hundred million ECGs a year in this country and 3% of the time were telling people something that's wrong or sing you're fine, but you're not fine. You're likely going to have a heart attack or stroke within a year, and we don't know it. As you can imagine, the technologies that were built that are most in market today, are now 30 or 40 or 50 years old, so they didn't use AI to make these decisions. And so we've taken millions of ECGs, connected them to outcome response data and other critical diagnostic data like echo cartograms and the like. And so we just built a really powerful portfolio of ECG-based algorithms that are either FDA approved in the middle of being FDA approved. These things are also deployed at scale. We have lots of algorithms to put in lots of hospitals, 140-plus hospitals, touching millions of patients, not small, but again, doesn't generate meaningful revenue yet. Here's the basic use case of how these things will get big. We have a hospital that recently rolled out our ECG platform. It's one of the top academic medical centers in the United States. It's rolled out. They see a large number of ECGs a year. There is currently a code to reimburse part of this world at $128 per ECG, that code relates to part of the population. The code will be expanded to, we think, the entire population. And so you can imagine, as these things get to scale, that one hospital alone could generate a few million dollars of revenue and multiply that by lots of hospitals and lots of ECGs and you get to a several hundred million dollar business just on our API predictor, low EF is even bigger. There's other algorithms coming. And so I would imagine that we will, for sure, be running some kind of algorithmic diagnostic on all ECGs in the future. And somebody will generate $1 billion or $2 billion of revenue just from that product alone. And that same thing is going to happen with echos and CAT scans and MRIs and mammographies and digital pathology slides. Each one of these data modalities that's being generated is an algorithmic diagnostic opportunity that's being missed today. And even at small dollars, $50 or $100 per algorithm the impact you get is enormous. So you're going to spend -- even if you spend $1 billion generating an [indiscernible] insight, you're likely going to save the U.S. health care system, $50 billion or $100 billion of mistake because as these patients don't get caught as they don't get found, they show up with complex disease and the most money we spend is in the last 90 days of life. So just to summarize, the data and applications business is growing rapidly, the bellwether of that business is our data in a modeling business, which is having a moment and growing quickly, and we're excited about the future of apps. On that note, do you want to hit the financials?

Yes. Quick, and we can jump back into Q&A. So as Eric mentioned, we've seen strong growth in data applications, largely driven by the Insights business the data licensing and modeling component of that. So 41% growth in Q1 of this year, the Insights business growing even faster, kind of partially offset by some of the smaller businesses. We announced expanded collaborations with Merck, Gilead, BMS over the last couple of months. As we've said, the pipeline with biopharma remains very strong and engagement with our customers given all the value that Ryan described is in a good spot. In terms of the financial metrics, these are new, but net revenue retention for 2025 was 126%, again, highlighting how these relationships kind of expand over time and people come back and spend more money. The TCV at the end of the year was north of $1.1 billion. So again, in a very healthy spot in terms of kind of that forward-looking visibility of revenue in the future. And then we've previously kind of talked about a 30% growth rate for this business. insights, again, outpacing that offset by some of the smaller businesses for 2026. This is just the same slide that you've seen probably previously around the kind of the breakdown of TCV, $350 million of that $1.1 billion related to 2026. And so again, as these relationships expand and we kind of stack these large strategic collaborations on top of each other, that just gives us a tremendous amount of visibility bolt into this year and next year. And so as we keep adding them, that visibility continues to grow, which is great. And then the 3-year CAGR, again, probably north of 25% for the data business. You'll notice that almost all of this is coming from insights. As Eric mentioned, while we are incredibly bullish on the app space, and we think that these things do get paid at some point in the future, it is not what we're counting on to deliver over the next couple of years. And so while we continue to push those things forward, we do think reimbursement will come in many of those instances, the bread and butter of the business over that time is going to be data licensing and model building. So with that, more than happy to take any questions on the data side.

I have 2 questions. As you develop more AI agents and applications, are there examples you would highlight as practice changing for physicians today, one? And second, Tempes has the largest multimodal oncology data set. Is the Tempest data set comprehensive enough? Or do pharma companies still look to sign additional data deals with other companies?

I'll take the first, Ryan, you can take the second. Look, I think where this is going, we talked a little bit about it a minute ago. I'm convinced where this is going is you -- over the next several years, we will begin to layer real-world data insights on top of every biomarker that is really therapeutically relevant. And so it will no longer be a sequence a non-small cell lung cancer patient to see if they have EGFR ALK. It will be -- once I know someone has EGFR [indiscernible] , what does it mean? Are they going to be in the 1/4 of patients that have almost no response, the 1/4 of patients that will be on that drug for 5 years or the 50% of patients that will be somewhere in between. And I will want to know that because that will determine what I do next. If I know my patient is very unlikely to respond, then I want to bring them in right away. On the other hand, if I know they're likely have been in this drug for a long time, different paths. So yes.

Yes, Eric, if I could add there [indiscernible] immediately practice changing. And we are actually going to present that at the ASCO meeting in a poster. We looked at early stage non-small cell lung cancer, and the frequency of EGFR mutational testing in the face of osimertinib improving survival in that population by 80%. Only 1/3 of patients were being tested. We piloted it in 5 large health care networks across the country. Within 6 months, it was 100%, and that retained over the next 12 months. That's immediately practice changing and has a tremendous impact from patient lives.

Yes. And the second question, not all data is created equal. And so there are many different facets of that data market where there are existing players that have been selling data for decades, right? Those groups are actually addressing a different type of question really around what is happening in the real world, right? It's very descriptive of like looking to see what happened, playing back the news. The reason why our business has grown and why -- what we see is we don't see any competitors in our space is that we're addressing why is this patient not responding to this existing standard of care. And that question, that why question is what's at stake for when they're designing that clinical trial. So when people are making a Phase III global investment decision, again, $200 million to $500 million in that decision, that's at stake. I really need to understand why are these patients not responding? And my hope is that my drug is going to address that. And so really, we're providing data to those types of customers in those use cases and others. But that ROI, delivering that type of value and helping those biopharma companies increase the success rate of those investment decisions is why our data business has grown to a scale that hasn't really been seen in our space. And so that maybe addresses the kind of what we see in the broader market.

And I've said this earlier, like I think per vacate, we said like there will come a time when no Phase IIIs ever fail. And some company like Tempus will be responsible for that. And if you think about it, it's not -- I know it sounds crazy, but it's just like from a tech perspective, of course, it's going to happen. I mean if a large Phase III fails, you either didn't understand the mechanism of action that drives your drug or what you saw in a Phase II is not being seen in a Phase III. There's no other reason. Like literally, it's like that. So both of those are solvable with real-world data at scale. You can understand what drives people to respond to your drug and then you can look at and interrogate at a comprehensive level, the population that's in your Phase I and Phase II and then look at the real world to see is that like representative of what I'm going to see and then track it as you start to enroll patients in our Phase II. In almost every instance, when you have big Phase IIIs that fail, we've gone back and looked at several massive base fees that have failed, and we basically set ahead of time, we could have predicted this failure, like here, you can see it. I could have predicted it from just digitized HNEs. Let alone more complex data. So this -- I think the R&D spends are going to get very efficient as AI becomes pervasive in R&D.

[ Brendan Smith], TD Cowen. Thanks for all the great input today. And I appreciate the color on kind of the monetization of data and applications here. I wanted to maybe double-click a little bit on that. And just can you speak to how you're thinking about evolving the actual monetization of the data business and even the foundation model itself, kind of just within the biopharma customer end market. I guess you mentioned deal size has grown per customer over year, some are bigger than others. But you're kind of integrating new data, you've got all these agents now moving forward. So I guess, how should we think about actual value switches over the coming quarters? And I guess also, is there any differences in terms of revenues to Tempus on what their customer expands with in early discovery, clinical, commercial, just kind of cadence over the next few years?

Yes, I can start. So I think the there's no big seismic change that we see coming. The big seismic change, I think, in terms of like just the general business was -- we sold this very large foundation model deal with [indiscernible], and that came with both a very large data license and some compute and the compute is at a lower margin, right? So that's like, "Oh, wait, that feels a little different than what is normally there". And there was a time when I thought people would sign a bunch of these very large deals. The way the market has evolved, is they're actually signing they'll sign lots of smaller deals, but not, I think, these giant big deals like just you're seeing it now. People are saying, I want to build a lung cancer model. I want to build a prostate cancer model. I want to build a digital pathology model. I want to work with you to build -- to find a new biomarker using Scan Data River. And so I think it's just moving so fast that that's the way it seems to be moving from lots of teams across these big pharma. So I think what you're going to see is the margin profile of our data business will look similar at what it is today, if not better. I think you'll see a blend between models and data that looks at some point, almost -- you won't be able to tell which what, what's what? Are you paying me $2 million to build a model or licensing $2 million of data, it won't matter. And most deals will have some of both.

I think the most sophisticated biopharma companies that we see that are really embedding AI into those critical decisions are actually using the data now at a higher level, which means they're just going to consume more, right? So the AI systems and the use cases you deploy within a pharma company will need that, especially if you're seeing high ROI use cases for every -- for not just 1 Phase III, but you want to see it for every Phase III in every Phase III or every trial. That's what we've seen for the most sophisticated. I think for a majority of the market though, they're not all as sophisticated as the leading early adopters. And many of those use cases are really just using AI to go a little bit faster to build a little bit of a faster car, but aren't addressing like how do I increase my success rate, right? I can go faster, which is still valuable. But when you think about the inefficiency of drug development, where do we waste all our money, it's in the failures. And so if I can increase my success rate, now the ROI can quickly flip on one trial and now -- and this is what we've seen across that expansion like Eric was mentioning, once you see sort of a successful trial readout and you saw what you did differently, now you're asking the question internally, why aren't we always doing that, right? And so that's what we see is like it's just a natural consumption more data. And so not just more in terms of volume, but more in the sense of having real-time data of not 3 years ago, but like literally of last year.

But similar to like the big compute guys, like you don't go to Azure and Azure or GCP. Let me see your menu. How much of this is storage? How much of it's compute, how much of it's large cluster compute, small cluster compute? How much of it's ingress, how much of it's egress. You're like, "I don't care. Just give me the number." Okay? because the margin profile is all pretty similar. Same thing with us. We're doing a deal right now that's -- we're just licensing embeddings, licensing, modeling embeddings. No files are moving, just the insights from those files that exist from a model we built on top of those files. So it looks and smells and feels like a data license, but it's really just someone -- giving somebody something they can use to build a model. So these things are going to merge together.

Eric, just a quick follow-up. Dan Brennan from TD Cowen. Just there's no doubt the interest from pharma appears to have really surged in AI, right, over the last quarter or 2, you listen to the public commentary from them. I'm just wondering you've discussed coming into the year, a really strong TCV and a high conversion rate. So you're enthusiastic about the data growth. Like are you -- kind of how would you characterize the interest today? Because it appears it's gone up dramatically. Are you seeing that will that translate in coming quarter bookings? Or just like -- any way you contextualize that just in terms of the real increase that we are seeing right now?

Yes. I mean I think you've already seen a piece of it you've already seen, which is we've had 3 quarters in a row of $100 million-plus bookings, some significant way higher than that and TCV growth. So TCV is just going up and up and up, even at our scale where we're licensing a lot of data. So I think that certainly is a part of it. But I think it's nothing compared to what's going to come. I think the first step is you have the CEO saying, this really matters. And for that to translate into a signed contract, takes time, and we're in the middle of that world. But I suspect over the next one to whatever, 3, 4, 5 quarters, you're going to see a lot of these folks that are realizing they have to jump in with 2 feet jump in with 2 feet. And I would say -- what I would say to you as a year ago, we had a pool with like 1 or 2 people thinking about diving in. Now we've posed like 20 people think of diving in. So the question is how they're going to dive.

Andrew Brackmann from William Blair. I wanted to ask on the Algos business and recognize it's not a near-term driver for until probably after 2028. But maybe can you just talk about the distribution system that you're putting in place here. In oncology, I get it, right? You have the lab, you have the report, you're giving that molecular information in those contextualized results. But I guess how do you distribute these in those non-oncology settings, what's the look like to these institutions?

Yes. So I mean, let's just take -- we use -- start take Northwestern. So we have spent an enormous amount of time building pipes between us and Northwestern building an infrastructure where that data can flow freely, deploying these algorithms at scale -- by the way, which took like, I mean years and years of effort, legal, IT, not small. Then once you deploy these algorithms to when you bring into the health care system, you inevitably break some part of it. These systems weren't designed for AI. So we began running our ECT algorithm [indiscernible]. And all of a sudden, we were producing an enormous number of patients that needed an echo or needed to wear a patch. Like they don't have like doctors lying around being like, great, so then you're like, well, wait a minute, how do we know, what's the change management side of this. And so you have to go through that. And so that's why it's going to take time for these businesses to actually really scale. You have to lay the pipes or do all that kind of foundational work for years. And then as the revenue starts to come, it really can be like a river. The failure of almost every AI company in our space is that the revenue is always way further than you think, and the cost is always way higher than you think. And so we just had 100 AI companies and looked at their decks you'd be like, "Oh, my God, every one of them was too optimistic and failed?" And most got to business. They just can't not they can't keep investing in that horizon that's always further out. We're just super lucky that we have a business that allows us to make those investments and still generate incremental EBITDA improvement. And that's just like compounding. I mean our xT, our TO, the approval this morning adds like a ton of additional revenue that we either drop to the bottom line or choose to invest. But if you look at the kind of EBITDA generation of the business today and what coming in 2027, we just have a lot of money that we're able to invest and still be EBITDA positive and cash flow positive, where most of our competitors that also make a lot of investments are just burning money.

I think just to add, Eric, a lot of the work that we did to build the integrations for oncology can transfer over to other disease areas. We become a trusted partner within these hospital systems. And so it is easier than being a new company that shows up that tries to connect with somebody like Northwestern, we've been a trusted partner with them for a long time, and so that makes it easier as well.

Ryan MacDonald with Needham. Eric, Jim, Ryan, thanks for hosting this today. I thought it was really helpful to -- that you laid out sort of the use cases in terms of how your pharma clients are using Tempus. And I was kind of curious to understand, as you think about the data business today across the early-stage clinical stage and then the commercial use cases, where does the majority of that revenue why where are you seeing sort of the most demand this year in terms of the use case? And where is the greatest sort of white space for you to go after in sort of applying the platform?

Yes, I -- so really early discovery and clinical development are kind of merged together under the R&D budgets of these pharma companies. And so that's the bulk of our business. Again, we're addressing the why are in certain patients responding to the existing therapies. Every discovery team needs to know that. Every development team needs to know that. And that's kind of the lion's share of the budget that's being allocated to our contracts. Now that being said, we have customers that are also working with us in commercial, but that's a growth area for us, right? And so our data can be used there as well, but again, no one has been able to address the why aren't these patients respond to therapies. And that's why it's -- we haven't really had to compete with others and why we've been so focused on that. Because even though we are working with 19 of the top 20. We're working at the same level for all 19 of those companies. And so our growth expansion still has opportunities just within R&D and then you kind of have expanded customer segments that can grow even beyond just the R&D teams.

Yes. I'll say one thing that we'll jump to financials, then we'll do Q&A again in a minute, so if you have questions. But like this is -- I think people don't -- people have not historically interested our data business, and they don't understand the moat around it. So here's a great example. We have been one of the largest abstractors of cancer patient data for the last decade. I mean, literally, when ASCO decided to partner with somebody, we were one of the 2 partners. We've been doing this for a long time at scale. We also have invested, I don't know, a few billion dollars in technology. We have like 600 or 700 software engineers have been focusing on this problem. So just assume we're a massive abstractor and assume we built unbelievable technology, and we only know oncology, like that's -- if you think of our last 10 years. Okay. Now the world of large language model shows up and you're like, can I just do this in an automated way? Why do have abstractors? That journey, and we are now crossing that journey in 2 of the largest indications. That journey has taken us all this time to be in a world where we can be. We have the ability now to take 100% of our lung cancer patients, breast cancer patients do automated abstraction at scale. So -- which we're doing right now. So all of a sudden, I as a data client can now access 100% of all the notes that exist across this millions of patients, but that's taken a decade of unbelievable time just to get there. And even then you still need humans to ensure that it's right. And so it becomes another powerful tool as we leave early-stage R&D and move deeper into development and commercialization. Should we hit the financials for a second?

Yes. Quick, and then we'll flip back. So I think we've covered kind of the economic model over the course of the session today, but it's obviously a framework for durable growth operating leverage, which we demonstrated over the last 8-plus quarters and long-term value creation. So it starts with diagnostics, where all the data is being generated, but it obviously serves a very important use case for physicians. We have improving margins and a scaled kind of infrastructure that allows us to get that leverage. And then we move over to the data and application side, where we've got a very good backlog of things through our TCV. We have a very broad customer base. As Ryan just noted, still kind of relatively early on in days. And we have a history of high retention rate. So looking at '26, we've given guidance of 1.59 and 1.6, which represents about 25% year-over-year growth, $65 million of adjusted EBITDA. The drivers are exactly what we talked about today. Within Diagnostics, we have very strong clinical oncology growth, improvements in ASPs. We have hereditary in the back half kind of normalizing after lapping some of the share gains. And then in data and applications, it's really just executing on the agreements that we have in place as well as expanding some of those relationships to give us more visibility into '27 and beyond. One other note from a balance sheet perspective, we did the convert a few weeks ago to take out the remaining term loan that we put in place, that gives us about $30 million of annual savings, allowing us to achieve positive free cash flow around the end of the year. This same slide we presented before, just how we think about kind of balancing profitability in growth. And so again, if we expect kind of a 25% top line growth over the next 3 years, the way that we view the world is that incremental gross profit dollars that are generated, we'll reinvest about 2/3 of those back in the business with 1/3 dropping down to the bottom line. And then after that third year, probably flipping that to a 1/3 is reinvested in 2/3 because at that point, you're generating enough gross profit dollars that we can maintain the level of investment that allow us to capture all the things that we think play out over the next decade or so, but still demonstrating operating leverage and significant free cash flow.

Yes, we can go back and forth. So just to wrap up really quickly, then we're happy to take questions. I think just a summary of hopefully what you've gathered from today, we have a diagnostic business that's strong. The integrated nature of our technology platform is driving higher growth than most of the other folks in the space, and it's sustainable. The trends we saw in 2025, the trends we saw in Q1 are continuing into Q2. So it seems to be a long-term pattern of us taking share from other folks. In terms of MRD, we have a comprehensive portfolio of both tumor naive and tumor-informed products. On the inform side, that assay is doing super well. We've got great market traction. And as we unleash more demand because it's currently gated, we expect it to grow pretty dramatically. And we continue to invest in R&D. We've got something like, again, 5,500 patients in studies right now. I think we don't get a ton of credit because we don't spend most of our time focused on the readouts of these kind of interim studies. But at the end of the day, I think it escalate something like 35 posters and papers and [indiscernible]. So the kind of scientific rigor of what goes on here is pretty extreme. And I would suspect that we're able to build a tumor naive product that's quite good over time. The data we're generating is compounding. The moat around our data business is growing. Our AI applications are really taking hold. The foundation model seems to be performing at or above our expectation. We'll certainly, over the next 3, 6, 9 months, have far more that hits the market in terms of insights that flow from that model. And ultimately, the company just has a really good financial profile, which was just made materially better by of this FDA approval this morning. And so we're just in a great spot. We're growing at a good clip, generating leverage, reinvesting it in forward growth. On that note, yes. So I don't know if someone's going to...

Yes. I think I've got it over here. Catherine Schulte with Baird. I had one on data, but we can maybe loop it into financials as well. If we look at some of the comments you made on customer growth and customer concentration in the data business, if we look at kind of value per customer outside of your top 5 customers that went from a little under $200,000 in 2020 to a little over $0.5 million per customer in 2025. So I was just curious, as we think about if we sit here 5 years from now for your data business, how much of that growth is from extracting value from your kind of per customer basis versus that customer growth?

I would think it's -- I would suspect it's actually on this kind of a curve. In other words, I really didn't know that. So if you said it went from $200,000 to $500,000 then I would say, okay, well, in the next 5 years, I bet it goes from $500,000 to $3 million. Like in other words, it's -- especially with biotechs that have no money or smaller or pharma companies that are more budget conscious, that's -- they're far more conservative than a giant global pharmaceutical company that can make a $20 million bet and this isn't the end of the world if it's not a great bet. But to a company that's got $50 million, they raised $100 million, those bets are -- so I would suspect that it goes way up. And you end up -- if we fast forward 5 years from now, you probably have, whatever, 500 clients spending $3 million to $5 million instead of 200 clients spending $500,000...

I think the surprising thing is that since we're addressing this kind of a unique question of like why aren't these patients responding to standard of care therapies, that's not just a big pharma challenge, that's a biotech challenge. And usually in biotech, their whole future relies on that question for their one trial or their one drug. And so what we've been surprised by is the number of biotechs are now signing on where data isn't a nice to have, it's an essential question to address because your future relies on it. And so the more that we can see -- like the market can see the success of how you can apply this makes it sort of an embedded budget line item for not just a big pharma company, but also biotech. And so we expect customer numbers to grow. We're not just the top 20, but also the dollar spend is going to be the biggest in the biggest biopharma companies.

Yes. And another way to think about it, and then we'll go to the next question is like you could almost imagine a world where we're probably there now because I go to every single biotech in oncology in the United States, market cap, let's say, sub-$1 billion or sub-$500 million would be like, hey, we'll give you access to our data, just give us like 20% of your company. And I would think like almost every one of them be like great. Like it's that valuable, right?

Brad Bowers from Mizuho. A bit of a preamble and then a 2-parter here. Just Tempus doesn't really get any credit as an AI company. I think it trades at 4x sales if you pressure the genomics business, especially the drug discovery stuff and the clinical trial benefits here. We saw software bottom over the last couple of weeks. So I guess this kind of gets into my first question, and this was touched earlier, but I'm starting to see GPU counts, compute of pharma peers kind of get shared around really just within the last few days. I think I have Lilly in the lead at about 1,000 GPUs, Recursion about 500 and then Amgen and BioNTech much lower. You talked to 1,000 H200s just for the foundation model. So I wanted to give you another opportunity. You touched on moat a bit, but just to kind of double-click on that and whether you think that the compute will start to enter the dialogue. And then also double-clicking on the Cowen question and maybe throwing it back to you, but it sounds like you talked on our models unnecessarily slowing down. It looks like the data business at 25% is a bit of a slowdown. So I'll throw it back on you why we should expect that slowdown when it sounds like those businesses are firing on all cylinders, and it feels like you kind of built an ark ahead of a flood here.

Yes. So let's talk about the ark for a second. So our compute capacity is probably equal to all of pharma combined. Like literally, we have -- forget our H200s, we have a roughly equal size cluster of GB200s, which are like 4x the H200s. So it's like just -- I mean based on the numbers you just gave, we probably have comparable compute to like all biotech and all pharma in oncology and then some because that doesn't even include all of our other compute, which is equal to or bigger than the 2 clusters we set up for foundation models. So we have a ton of compute and we have a ton of data. And so I think that is the ark. I think the challenge for us in terms of both valuation and growth is, we have always been focused on long-term sustainable growth. We've had opportunities in the past to kind of make decisions that would accelerate our growth in the short term, but may have been hurt us in the long term. And we just tend not to choose those. And so I think we feel very comfortable that the aggregate business will grow 25% or so. We've called out that the data business will grow faster. That could be in the 30s. That could be even faster. There could be other parts of that business that grows slower, like, for example, our CRO business which is, I think, shrinking or relatively flat. So we have other parts of the business that are not growing because we're not investing in them. And so it all kind of -- and we don't get into all the micros of these different businesses. But at the end of the day, our 2 largest businesses, our oncology sequencing business and our data licensing and modeling business are growing faster than everything else because everything else is growing slower and you can see our growth rate. And we don't think that's slowing down. So we're not forecasting like something to actually decelerate materially. We're just saying, hey, for 2026, expect 30-plus percent growth in the data business and the 3-year growth rate is going to be, call it, 25%, and we don't benefit by saying the 3-year growth rate is going to be 30% or 35%. Our stock won't go up because of the first part you mentioned, which is we're caught in this middle ground where technology investors who tend to invest in and think about AI don't understand diagnostics, and diagnostic investors who are deeply steeped in next-generation sequencing don't understand the data and AI. And so you end up in a world where somebody is always worried about the thing they don't understand intimately, and so they just don't know what to do. And so I don't know when that solves itself. It may solve itself. It may not solve itself, in which case we could do things to solve it. But at the end of the day, we've been focused this year on just kind of getting these businesses in the best spot we can, making sure that ark is as good as it can be. And if the market doesn't ultimately recognize the value of our data and apps business, which I -- if we were -- if we took the data and apps business public tomorrow, I would suspect we'd trade higher than the entire market cap of Tempus and could trade at 2x it. So it's not -- I mean it's -- one could argue it's got negative value. So we just kind of look at it and say, eventually, that will either solve itself or we'll solve it.

It's Mark Massaro, BTIG. Maybe flipping back to the Diagnostics business. So your partner, Personalis, I think, has received about 4 Medicare coverage decisions in the last 6 or 7 months. To me, having non-small cell lung, breast and I-O monitoring really is a bit of an unlocking. So when do you expect to sort of unlock the gate, so to speak? And just kind of -- can you give us a sense for what percentage of your reps have been promoting MRD versus when you do unlock, is that going to be a full flip? And then quickly on the rare disease portfolio, you indicated that you plan to launch a whole genome panel this summer. Can you just help us rank order your priorities, like how big of a push do you intend to make in rare disease? Why is that important to you relative to some of the other segments you're going after?

Yes, I'll cover the first. So the number of people currently focused on the MRD portfolio, I don't see Laura. I think she was here earlier, left. But I'm going to say, call it, 15 to 30, somewhere in that range. And we have about 200-plus folks in the field, I think, in just core oncology CGP. So think of it in terms of like if you were trying to understand like percentage, it's 10% to 15% or something, I don't know, somewhere in that range, maybe sub-10%. So it's highly gated. And Personalis is public, so you could look at their financial statements. But if we sent them -- let's just say we sent them 20x the volume of orders we were sending them today, how much cash would they burn? I don't know how much cash they would burn, I don't know how much cash they have, but I don't think that math is sustainable. So it's all been tightly orchestrated to make sure that they -- we are growing as they are growing and that the whole thing works, so we don't end up either breaking their labs or breaking the financials of the business. And so it's in a great spot. I mean what I would -- what we have said historically is the demand is way more robust than we would have thought a year ago. There doesn't seem to be a cap on it. So I'm comfortable that as we continue to invest here and expanding, we'll expand. In terms of...

On rare. So as Tom had mentioned, they had a whole genome offering that was in place. They've largely kind of deprioritized a little bit when reimbursement wasn't there. Obviously, the reimbursement landscape has changed. And so they began working on a whole genome offering just because that's obviously where the market is moving. So it's obviously not the top priority for Ambry given how large the HCT business is, but we do think we can be a player given the relationships that we have with genetic counselors and build out kind of a meaningful business there as well. So you'll hear us talking about it more, but it's still not a significant driver today.

If you also -- if you look at the investments we make, it's kind of interesting, right? You could -- if you think about the decision tree, you could say to us, "Hey, why don't you invest another $25 million and try to develop like some novel epigenetic or methylomic assay, right?" So that you -- someone might say that. But if you look at where our core strength is, it's really at the intersection of technology and diagnostics, not being first to market with something really, really, really novel. And so when you think about rare, I can't think of a single use case that would benefit from a full understanding of the clinical case of a patient than rare. It lives at the intersection of some molecular insight and the clinical diagnostic odyssey, and we're as good as anybody at pulling in that data and making sense of it. So I would suspect long term, if it's not Tempus, it will for sure be a company that looks like Tempus that wins rare. It will not be the company that can generate a whole genome BAM file. That many companies can do quite well.

Maybe we'll do one more.

This is Paul Stewardson from Stifel here for Dan. Just wondering on the foundational model that you shared some early data, hazard ratio, that sort of thing. What's the level of need for prospective validation? You mentioned you have these retrospective, isolated cohorts that you can validate the models in. Is there a need to kind of -- given there may be some survival bias of what models are coming out, is there a need for doing long-term prospective? And to what extent does that gate your ability to move that into a useful application? And then just briefly on the other side of the business, can you talk about the investment? There's this next-generation tissue-naive test. What kind of clinical evidence generation plans do you have that might be bigger than you had before the V2 was being talked about?

I'll take the first. And James, you, can take the second. There Is no -- it's a 2-part answer. There is no gate needed, and we, for sure, will run those studies. There's no gate needed because physicians are free to make decisions based on the data you present them, and we don't get paid for these insights. So if I was -- like if this was Oncotype DX, for example, and I wanted to get paid, I'd have to run a very large study to then convince somebody to pay me and put me in a guideline. But in my case, if I can predict EGFR response or ALK response, I don't get paid for it. And so -- and doctors can make a decision based on what I published as to whether or not they believe that, that is predictive or prognostic and whether they want to pay attention to it. So I think there's no gate. These things will come at scale, but I suspect all of them will turn into -- and they'll have to be analytically and clinically validated before they get on the report. And I suspect all of them will turn into really cool studies over time to figure out how good are they. And eventually, they'll weave their way into guidelines. But the river will come, and we will not gate it. And I think it will be material in terms of how people have to react to it.

And then, Kate, I know you're over there on the V2 kind of studies. I know you briefly hit on it, but...

Yes, we're making those investments now in terms of -- we showed the one slide over the next few years. We've got about 5,000, a little bit more patients in studies today. We continue to enroll. We're enrolling across all major indications, so really pan-cancer, and we're sort of building the technology to be able to -- once we make it through analytical validation, then we'll just be able to start hitting clinical validations across all of those different indications.

I think the way to think about this space is the -- first of all, a significant percentage of the market is CRC. And in colorectal cancer, you have lots of tissue. So it's not a great offering to say, hey, switch from a tumor-informed assay that's really sensitive and come to a tumor-naive assay that's less sensitive when you've got a lot of tissue. So we have all along suspected tumor-informed would win the day in CRC. In other areas, like, for example, lung cancer, where tissue is far more scant, you think, okay, this is a great assay for tumor naive. The problem is those assays haven't performed that well. So tumor-informed is winning the day there as well. So I think we're in this kind of weird zone where -- including our own, which -- whereas the tumor-naive assays aren't performing well enough to win the market at scale and the tumor-naive assays need to go through this consistent R&D process, getting from 500 ppm to 300 ppm to 200 ppm to 100 ppm to 50 ppm. So there's some zone. I don't think you need to get to like 10 or 2, but there is a zone where you need enough sensitivity and specificity and a low enough Illumina detection that you actually like, okay, this thing is -- can play against the market-leading assays like Signatera. We are getting very close to there now. And so we're migrating from Version 1 of our naive assay to Version 2, and now we're going to kind of roll that out at scale in terms of these different studies.

Maybe one more comment to make there, just like something that differentiates us in this space. We've talked a lot about this kind of multimodal data. And what we start to see not just in Tempus, but in the field at large is when you add additional modalities of data, you're able to increase signal-to-noise. And so as we think about the really broad data set that we have that we're continuing to generate, part of the belief is that we will be able to then layer in these other modalities much faster and be able to then improve upon the assays and the technologies that we have. So lung cancer or whatever the indication, we'll be able to leverage imaging and all of the other things to add to that signature.

Yes. I will say one last thing on MRD. I'm first of all a big believer in MRD. I think it's an awesome space we want to play in it. It is unclear to me how this whole thing shakes out. Remember, I think you have to really understand the adoption, right? You turn this test on and then you turn it on in bundles. You get doctors to basically order it in bundles on a recurring basis at scale. And so all of a sudden, you go from like not a lot of volume to what looks like a lot of volume. And if you get even some minimal level of reimbursement, $500 a test, it's like not small. But I don't really know how that all shakes out. Like I feel very good that comprehensive genomic profiling has shooken out. But I think on the MRD side, we have to let this thing play out for another couple of years to figure out what's the cadence, what's the cadence doctors are going to want to order, what's the cadence that's going to get paid for? How does it really work? It will be a great space. But I don't think you can just look at the trend lines today and be like, oh, they're going to continue for the next 10 years. I think there's going to be some movement there to figure out how to rationalize what's going on. You've got some doctors ordering these things every month and some doctors never ordering them. And that's a unique paradigm relative to CGP. There, you just have people that believed or didn't believe in the genome. Like I didn't know -- I don't know think it matters. But once they're like, oh, it matters, the ordering patterns were pretty normalized. On that note, thanks for joining us. I think we're going to do some tours.

Yes, what are the logistics on that?

Please stick around for some refreshments. We're going to do 3 separate groups for lab tours for those who can stay. So we'll collect those in the back. Thanks for joining us.

Thanks, everyone.

Thank you, everybody.