Tempus AI - Earnings Call - Q2 2025

Transcript

Speaker 6

Thank you for standing by and welcome to Tempus AI's second quarter 2025 financial results conference call. All lines have been placed on mute to prevent any background noise. After the speaker's remarks, there will be a question and answer session. If you would like to ask a question during this time, simply press star followed by the number one on your telephone keypad. I'll now turn the conference over to Elizabeth Krutoholow, VP of Investor Relations. You may begin.

Speaker 2

Thank you. Good morning and welcome to Tempus's second quarter 2025 conference call. This morning, Tempus released results for the quarter ended June 30, 2025. The press release, an overview of the quarter, and our latest presentation are available on our IR website. Joining me today from Tempus are Eric Lefkofsky, Founder and CEO of Tempus, and Jim Rogers, CFO. Before we begin, I would like to remind you that during this call, management may make forward-looking statements that are subject to risks and uncertainties that could cause actual results to differ materially. For a discussion of these risks, please refer to our 10-K and other subsequent filings with the SEC. During the call, we will discuss non-GAAP financial measures, which are not prepared in accordance with generally accepted accounting principles.

Definitions of these non-GAAP financial measures, along with reconciliations to the most directly comparable GAAP financial measures, are included in our earnings release, which has been furnished to the SEC and is available on our website at investors.tempus.com. I would now like to turn the call over to Eric.

Speaker 4

Morning, everyone. I'll provide a quick summary. Q2 was a fantastic quarter. As I mentioned in my letter, the company is hitting its stride as we approach our 10th anniversary, which is great. Revenue increased 89.6% to $314.6 million. Genomics revenue increased 115% to $241.8 million on accelerating volume growth in oncology, which increased from about 20% last quarter to 26% this quarter, which was great to see. Our ad testing, which was 32% in the quarter. Aid in services revenue increased 35.7% to roughly $73 million, led by Insights, which is our data licensing, which grew almost 41%. Quarterly gross profit was $195 million, roughly a 160% increase. As the business is growing, we continue to be disciplined about the investments we're making. As such, we saw another sequential improvement in adjusted EBITDA, which went from roughly negative $16 million last quarter to negative $5.6 million this quarter.

We're approaching adjusted EBITDA breakeven. We increased our full year 2025 revenue guidance to $1.26 billion and maintained our adjusted EBITDA forecast of about $5 million for the year, which would be a roughly $110 million improvement over last year. We also improved the balance sheet. We issued $750 million of 0.75% convertible notes, which will drive down interest expense and produce lots of cash savings. This, along with the fact that our cash and marketable securities finished the quarter at about $290 million, and net of that, paying down some of the debt, as I mentioned, we added about $375 million additional, leaves our balance sheet in a really good spot as we approach Q3. All in, the business is right where we want it to be, and we're making great progress. With that, happy to take questions.

Speaker 6

Thank you. As a reminder, to ask a question, you will need to press star, then the number one on your telephone keypad. If you would like to withdraw your question, press star one again. We do request for today's session that you please limit to one question only. Thank you. Our first question comes from the line of Daniel Brennan with TD Cowen. Your line is open.

Great, thank you. Thanks for the questions. Congrats on the quarter. I'll just ask a multi-partner up front. Maybe first, Eric, can you just elaborate on the strong core genomic volumes in the quarter? I know in your script, you kind of talked about sales force efficiencies. Just any color where the strength came from, tissue versus blood, broader market share versus share gains. Next, you know, Ambry, really strong quarter there. I know in the script, you talked about volume outlook of the deal model could be too low. Just kind of what are you seeing there? I know you also talked about the rare kind of genetic disease for kids. Is that contributing today? Finally, Insights, nice quarter there, 40% growth. Any color how Pathos is doing? I know you called it out in the written script that it is contributing.

Just wondering how that's going and if there's any color on bookings, that would be terrific too. Thank you.

Speaker 4

Yeah, so there's a lot there. I'll start at the beginning. Maybe Jim could jump at the end. We saw significant sequential volume growth across our entire oncology testing compendium. It was widespread. It wasn't just solid or liquid. It was really across the board. It was notable. For us, it's a function of a lot of the things we had put in place in terms of sales force efficiency, realigning some of the territories, improvements we had made across our technology stack. We just saw a bunch of those efforts kind of pay off in Q2. The volume growth has just been really strong, and it's great to see, and it's exactly what you want. If you put forth a bunch of initiatives and they start to pay dividends, you have a reason to believe that they're going to be sustained.

We feel like we're in a great spot in our oncology business. The hereditary volumes are also significantly stronger than we expected. They too show no near-term signs of slowing down. The growth is really coming from a few areas. It's in part coming from the fact that some of the historic players in that space have just kind of shrunk in terms of their market share. We're capturing a lot of that volume. It's also that the overall space appears to be experiencing some really good tailwinds. As I called out in my letter, it makes sense. When we acquired Ambry, we had hoped that the space would be quite large over time. People thought it was kind of getting to saturation and maturity, and we historically have said we think that's exactly wrong. We tend to believe that's the case.

There are just far more people that are at risk of getting cancer or at risk of getting disease than those that have a disease. I would not be surprised if the volumes from hereditary sequencing are dramatically larger than the volumes from cancer sequencing or any one disease. They're also making great progress in rare and pediatrics. They're basically the number two player in that space. Even though it's small, it's growing pretty rapidly. I would suspect over time, that business will grow as well. All in, the testing business just was really strong across the board. In terms of our data business, it too is having a moment. Obviously, we signed a very large deal to build a foundation model. That's super exciting. That's under construction. We have great visibility to the balance of the year, which is where you want to be.

It's the law of large numbers, right? Our growth rate is going to, you know, we're not going to grow that business at 80%. It's not going to happen. We told everybody we want to maintain growth rates around 25% or 30%. We grew above that this quarter, which is great. All in, I would say if you look across the two main drivers of the business, sequencing diagnostics and data, we're just in a really strong spot.

Speaker 6

Next question comes from the line of Yoko Oku with Morgan Stanley. Your line is open.

Speaker 2

Hello. Thank you for taking my questions. Given the significant progress you already made in the first half migrating XT volumes to CDX, in addition to ASP drag from XMRAMP, how should we think about cadence of ASP as we think about the path towards the goal of migrating 40% of XT volume to CDX exiting the year? Also, although you have a whole transcriptome RNA panel, some of your peers are adding additional features on their tissue therapy selection test, such as RNA and epigenomic markers. Is this something that you might look to in the future?

Speaker 1

Yeah, I'll take the first question on kind of ASP, and then Eric can speak to kind of future products. On ASP, as we mentioned in the letter, we did see an increase in the volume of XT CDX. It was about 20% of our XT volume in Q1, grew to about 28% in Q2. As you mentioned, we're kind of targeting to get to 40% by the end of the year. In Q2, we did see a mix shift to some of our lower dollar or lower reimbursement assays, both XM, which we don't have reimbursement for today, and then also XT. We would think over the balance of the year, as we kind of approach 40%, you would see incremental gains to overall ASPs. On the XM side, that will continue to grow as well. We are still gating that volume.

We're not anticipating XM creating offsetting all of the gains that we have. We won't see a significant step up as we saw in Q1. We would anticipate small incremental gains over the balance of the year.

Speaker 4

In terms of RNA sequencing, you know, we've been doing whole transcriptome sequencing since the beginning. It's just kind of the way we started when we opened up our lab. We were always doing targeted panel sequencing on the genomic side, but doing whole transcriptome on the RNA side. Our pioneers in the space of running whole transcriptome sequencing from FFP slides. Our panel is just super comprehensive. Not only do we basically report on expression levels on the entire transcriptome, we also do TCR profiling through that assay, BCR profiling, HLA typing. We have a whole body of algos that are also produced from that. At the end of the day, it's just a really comprehensive panel. I think the likely advancements we'll make will be more on the DNA side going forward as we migrate from targeted panel sequencing to whole genome sequencing over time.

We expect, if you kind of fast forward five or ten years, my guess is targeted panels go away largely, and people are just doing whole genome and whole transcriptome and layering in other markers, like for example, epigenomic or methylation markers throughout those assays.

Speaker 6

Thank you. Next question comes from the line of Rachel Vandel with J.P. Morgan. Your line is open.

Perfect. Good morning. Thanks so much for taking the questions. I just wanted to dig into the trials and Insights businesses a little bit. First up, can you just talk about how bookings trended in the quarter for Insights and trials? Are you seeing any weakness from pharma due to some of the headlines related to MFN and tariffs or even biotech customers due to the funding constraints? On Pathos, if you could give us some color there, it looks like that was around $16 million of revenue in the quarter. How should we think about that trending for the back half of the year?

Speaker 4

Yeah, I can start at the beginning. Jim can talk about the Pathos and AC revenue. Our bookings obviously were super strong in the quarter because the AstraZeneca and the Pathos deal was actually an April event. In terms of bookings, it was a large number in Q2. Our total contract value, which we only report out annually, is up materially, as you can imagine, by virtue of that deal. We're in a great spot as we look at total contract value going into the back half of the year. We're ahead of where we wanted to be, which means we're in a great spot given that we had lofty ambitions. The clinical, the trials booking business has never been a large driver of bookings, our clinical trial management business, in large part because, you know, if we help people enroll 10% or 15% of their U.S.

trial, it might be 10 patients or 20 patients or 30 patients or 5 patients or whatever. It's just not a huge dollar driver at the present moment. As that business scales, it might become a bigger number. Right now, it's still fairly small. It's super important in terms of our relationship with biotech and pharma, and it's super important in terms of our relationship with providers. We've now disclosed that we're connected to more than 4,500 providers and institutions across the country, which is obviously a huge number. These connections are deep. These are basically bidirectional connections where people are allowing us to pull data down. We're generating insights or diagnostics and putting that information back in. One of the parts of that is, can we look at healthcare data in real time, basically data from these electronic healthcare records?

Can we do really important things with that data, like identify patients that might be a perfect fit for a clinical trial or identify a care gap that, for whatever reason, somebody missed and the patient's on the wrong path? Increasingly, we're doing that, which means the kind of important role we play continues to go up. I think the clinical trials time business is in a really good spot, as is our next algos business. They're just relatively small today, as we've called out historically from a revenue and bookings perspective.

Speaker 1

Yeah, on the AstraZeneca Pathos deal, in terms of kind of contribution to the back half of the year, it obviously didn't get a full quarter's worth of revenue in Q2, given the timing that we signed that deal. There will be a sequential kind of slight step up, but you would anticipate kind of similar levels in the back half on a quarterly basis.

Speaker 6

Next question comes from the line of Michael Wiskind with Bank of America. Your line is open.

Hey, follow up on that last one. Just wondering, you've made a series of major investments and partnerships with pharma over the last couple of years. Wondering what the pipeline for that looks like in terms of future. I'm talking about, obviously, the data and Insights platform business. Wondering if there's what the pipeline for that looks going forward in terms of the major pharma and their appetite to continue to invest in data, particularly as you're starting to see more and more other genomics vendors enter the field. If I could tack on a second one on that, a lot of focus, obviously, on profitability and EBITDA and get progress in the quarter. Want to hear you talk through your priority for investments going forward. How are you going to balance among the various parts of the business where you're putting the incremental dollar? Thanks.

Speaker 4

Yeah, I'll start. Maybe Jim can also jump in. In terms of our pipeline, it continues to be super strong. We have, as we've called out in the past, the only real kind of market impact we felt was several years ago when biotech funding essentially dried up and we just lost a huge body of customers that in 2022 were flushed with cash because they all went public. We're long past that. That's long in the rearview mirror. If anything, I would hope as the markets improve, some of these biotechs may actually be able to raise capital again, which would create a whole new source of potential data clients. In terms of the big pharma, we continue to have a really healthy, robust pipeline of very big deals that we expect to close in over some horizon of time.

It's essentially because even though big pharma R&D budgets may be getting cut or there might be pressure, data and AI and technology is still such a small component of those multi-billion dollar budgets that we're just not, we're not the thing that's going to get cut. They may not make an investment, but they could not make that investment even in robust periods. If they want to make an investment in AI or data, and they've got a $9 billion R&D spend, they're not going to worry about a $100 million investment over several years. We have not felt any pressure. We also have not felt any pressure from competitors in that space. Literally, every time we lose a deal, basically, we lose it to a company not wanting to go forward and make an investment, not to somebody else that has a product like ours.

Whereas we have very good competitors in the diagnostic space, people that are at scale and do a perfectly good job sequencing patients, we just don't feel like we have those same competitors on the data and AI space. We just don't bump into people. We have not seen them. That hasn't changed recently. It may change in the future, but it certainly is not the case today. In terms of investments, I'll start and Jim can always jump in. This is going to be one of the questions that we have to think through. It was important for us to, as we approach our 10th anniversary, be able to build a business that was growing quickly and still generating operating leverage and getting to adjusted EBITDA breakeven.

As I mentioned in my letter, one of the things I'm most proud of is the fact that you've got a business growing north of 30% at real scale that is still generating improvement quarter over quarter, whereas many others are growing, maybe not even as fast, but not generating that leverage. We're still losing lots of money with no sign of that turnaround in sight. We're doing a really nice job of growing and being disciplined. I expect us to be disciplined until we round adjusted EBITDA positive and round cash flow positive. We're not at a point in time where we're just like harvesting profits. We're still at a point in time where we're making lots of investments in growth.

I suspect we'll make lots of investments in growth over the next several years, given the size of the space and the fact that, as leaders of bringing AI to diagnostics and healthcare more broadly, and given the size of the market, the last thing we want to do is kind of optimize for the short term and miss the opportunity in the long term.

Speaker 1

Yeah, I would just add that we're not doing anything unnatural to get this leverage in the business. We are making significant investments both on the genomics business, data, and on the AI side. We think the level of investment that we're making today is appropriate. We wouldn't accelerate it, obviously seeing the leverage that we're getting out of the business, but we're certainly not starving the business to kind of show this improvement.

Speaker 6

Next question comes from the line of Ryan McDonald with Needham & Company. Your line is open.

Hi, Eric and Jim, congrats on the great quarter. Maybe one on the data business, one on the genomics business for me. On the data side, great to hear things kicking off with building the foundational model and doing some of the training here. Just curious, though, in terms of the incremental demand you're seeing from other partners right now to build something similar here, obviously, you're going to have the sort of first version expected in early 2026, but given the pace of rapid AI investment, can other partners afford to wait for, let's call it, maybe proof of concept to come out in early 2026? On the genomics side, it clearly seems like the MRD portfolio is resonating well within the industry.

Just curious when we could see maybe the bigger unlock here with reimbursement from MOLDEX and when we should expect that to happen if that's a second half of 2025 or a 2026 event. Thanks.

Speaker 4

Yeah, I can start. I think that's one of the great questions that we're also thinking a lot about. We're in very deep conversations with a bunch of folks who are thinking about building similar models to the one that AstraZeneca and Pathos are building. So far, obviously, we haven't announced anything. I think one of the questions we have for that group is the same one you just asked, which is, can you afford to sit on the sideline in a world where these types of models are likely to be transformative, not just to your R&D portfolio, but to the drugs you have in market? That's something I don't think people are really focused on. We kind of called it out a bit in the letter. One of the things we would expect from this model, you're taking an enormous amount of data, right?

We announced we have over 350 petabytes of connected clinical molecular data. It's just this massive data repository. You're essentially running compute for months on a cluster of over 1,000 H200s, which is not a small amount of capacity. You're talking hundreds of billions of tokens where you're running compute. You're likely going to see associations that you just couldn't see until you ran that amount of compute. Those associations are likely going to be things like, when a patient has this particular mutation, even though it might be standard of care to go on X, Y, or Z drug, we know that typically half the population doesn't respond to that drug. Even within that, you have gradients of response. You have 20% of people that might be super responders and 20% of people that never respond and some group in the middle. Right now, these drugs are very brute.

They're like, I have a mutation or there's a biomarker or maybe not even a biomarker, and I get a drug approved and I give it to everybody. What AI is likely going to do, what these foundation models should do, is provide insight to physicians and patients as to who's likely to respond and not to respond. I would suspect that will fundamentally change care and probably much faster than guidelines can adapt because these guidelines are human-oriented, not AI-oriented. I would suspect that AI is going to create some very big disruption within the space. With any type of disruption, you have to kind of ask yourself, do I want to be a disruptor or do I want to be the disrupted? It was very cool to see AstraZeneca, you know, taking a lead in that, and I suspect others will. Over time, I suspect everyone will.

I just can't see a world where people are like, nah, I just won't, I don't need that kind of data. I'll just do it the old way. I think, you know, it's long term, I would suspect many people will build similar models or avail themselves of similar models and move significant dollars from historic chemistry and biology to data and AI. We'll certainly be a, or I would expect us to be a big beneficiary of that. What we don't know is, is that going to happen in a quarter or two or three or four? We just can't, we don't have that kind of visibility. We do have visibility to the business performing really well in the near term because of things we've signed in the past. We're fortunate that we don't need to sign new things in the future to generate really good growth.

In terms of minimal residual disease (MRD), and then I'll give it to Jim, you know, I'll speak to the volumes, he can cover the, you know, when's it going to be a bigger event. We have a really nice MRD portfolio. I mean, it's broad in terms of the fact that we have both tumor-naive and tumor-informed offerings. It's broad from the perspective that we're in multiple different disease areas: breast cancer and lung cancer and IO response and CRC. We cover a significant variety of assays in that market, and it's a really exciting growing market. That said, we are gating volumes until reimbursement, and maybe you can provide some context on that.

Speaker 1

Yeah, we've previously disclosed that we anticipate getting reimbursement by the end of 2025. No changes on our assay. Personalis has publicly disclosed their timing as well. Built into the guide is not a meaningful uptick of MRD revenue for 2025. We would anticipate that occurring more in 2026.

Appreciate the color. Thanks.

Speaker 6

Next question comes from the line of Dan Arias with Stifel. Your line is open.

Hey, good morning guys. Thanks. Jim, on data, can you just maybe orient us on expectations for the back half year? Solid growth, obviously, in the quarter, but it is sequentially ticking down slightly out of the end of last year. The comp steps up quite a bit in 3Q and 4Q. 3Q is actually a pretty stiff compare. Where do you think growth lands in the back half of the year? Is 30% plus still kind of okay to model?

Speaker 1

Yeah, I mean, I think for the year we've talked about around 30% growth in the data business or slightly above. I think the sequentials, Q4 is always the largest revenue quarter for us. Q4 of last year, there's always a step back that occurs in Q1, and it kind of builds throughout the year. As Eric notes, we don't anticipate that business growing at 40% forever. We would anticipate a tick down and landing around or just north of 30% for the year.

Okay. Just to be clear, that's like a 20% growth rate for the next two quarters or so. You're basically kind of halving what you are where you are today.

It will grow slightly. It will grow faster than that. I'm saying it's not going to be 40%.

Okay.

Speaker 4

We don't want to back into providing guidance by business unit on a quarterly basis going forward. The data business is in a great spot.

Speaker 6

Next question comes from the line of Mark Massaro with BTIG. Your line is open.

Hey guys, congrats on the good quarter. I wanted to ask a question about your liquid biopsy business. I recognize that XT in tissue is the majority of the volume, but there was a large company in the space earlier this year that put out some compelling data around the possibility of increasing time points in liquid. My question is on your XF franchise. What are you seeing in terms of demand as far as time points go? How do you believe you're positioned competitively, recognizing that there is another player that's pretty large in the space? How do you think that your position is to compete directly against them?

Speaker 4

Yeah, I mean, at a super high level, obviously, our growth rate in liquid has been dramatically faster than the rest of the market because we're at a real scale. I think we've disclosed historically that reductions are about a third of our volume that we take or something like that, yeah, 25% to 30%. It's significant. It's a significant component of our volume, and we've grown much faster than the market. Net-net, that franchise from a therapy selection perspective is in a really strong spot based on historic performance. We have an asset that we believe is completely competitive with others in the market in terms of size and breadth and so on and so forth. The kind of multiple time point treatment response monitoring space that is certainly emerging is one that we too are looking at.

We have a portfolio across not just measuring minimal residual disease, but also looking at treatment response monitoring. Over time, we would suspect that this is a bigger part of the market, but it also requires reimbursements. You're kind of in the same boat with PRM that we're in with MRD, which is, until you have payers paying for it, you have to gate volume. Otherwise, you're just going to run a bunch of tests and not generate any revenue. We have been more disciplined than that. Over time, these will produce lots of additional tailwind to our unit volume, but they require reimbursement. Right now, they're small, but we have a super competitive product set.

Okay. My second one is just on Ambry, certainly, a lot stronger than we were modeling. I think you alluded to, you know, other players in the space either exiting or perhaps just not having the same level of focus in the past. I'm curious, were there any investments, like, did you guys make any investments into your commercial team in Q2 or prior to that? I'm just trying to get a sense for some of the puts and takes to explain the strong growth in the quarter.

I mean, the growth is basically, I think, at a high level, you could think of it as split 50-50 between market share we're capturing from others who are kind of falling apart and gains we're making just on our own or the average making on its own just by virtue of like having a great product. We're benefiting from investments they've made for years as they have kind of built an absolute best-in-class hereditary platform that is just recognized as absolute best-in-class by a significant % of the market. More and more big systems are just migrating to Ambry because it's kind of the gold standard in that space. People want all that comes with it: the turnaround times, the best-in-class, you know, error rates, the technology they've wrapped around it, the analysis that they provide, the experience they provide. It's just a really good product.

They've also made a ton of investments in Rare and Peds. We're super excited about the growth of that product set, which is still quite small in relative terms, but I would suspect it'll be a big driver over the next several years. It feels to us that the core of the business, the core of the growth feels durable and sustainable. We're not here to say, you know, you can kind of bank on 30% growth for the next five years. We're going to, as we have tried to be historically, we're going to be conservative until we see these things play out.

Great. Thank you.

Speaker 6

Next question comes from the line of Andrew Brickman with William Blair. Your line is open.

Hi guys, good morning. Thanks for taking the question. Pathos has been active on the business development front this year. It seems like more and more transactions or partnerships are happening in the space. Products are finding better homes. For you guys, how should we think about your appetite to continue to do acquisitions or partnerships over, call it, the near to intermediate term? Thanks.

Speaker 4

Yeah, I mean, we have historically been opportunistic, but not overly acquisitive. I think we try to bring the same discipline to the companies we look at, whether it's from a BD or corporate debt perspective, as we've been in terms of running the business. We don't want to derail all of the good organic momentum we have. We're not interested in taking a turn and going left after we've been going straight for a long time. I would suspect we'll continue to do that. There are certainly some number of companies out there that have interesting products or interesting teams or interesting data sets that we look at, whether it's in our applications business, in our data business, or in our diagnostic business. We continue to be measured and disciplined.

I would suspect us to, if you look at our last six or seven years and the companies we've acquired, my guess is we'll take a similar approach in general going forward. The market's changing pretty dynamically. We're also mindful of, as these chess pieces move around, we don't want to find ourselves in a worse position than we otherwise would be in.

Speaker 6

Next question comes from the line of Doug Shanko with Wolfe Research. Your line is open.

Hey, good morning. Thank you guys for taking the question. Just a couple cleanups. Maybe a follow-up to Andrew's question. Is there a good rule of thumb on just, you know, how you're going to think about partnering versus organic? You know, is it how much of it's technology? How much of it is ROI? Just would be good to know kind of maybe a little bit more on the specifics of just kind of like how you almost think about the math there. My second question is really another follow-up on MRD. Recognizing this is going to be more a bigger part of the story as we get into next year. I am curious about, you know, the next few years. How big as a % of oncology volumes would you expect MRD to be? How does that affect the gross margin trajectory over time? Thank you.

Speaker 4

Yeah, I certainly can jump in. I mean, let's start at the beginning. We're not looking to acquire a company. If we have a long operating plan to get to adjusted EBITDA positive, we're not looking to acquire a company that would take a giant step backwards where we're like, "Oh, no, we're just kidding." We kind of start from if we're going to buy something, we don't want it to derail our organic plans. We kind of start from that lens. Not that it would be impossible, but in general, we start from that lens. In addition to that, as we've also said historically, we feel really good about our diagnostic portfolio at present. We think we have an unbelievably comprehensive program from hereditary to solid and liquid treatment selection to MRD and monitoring and response. We really feel like we've got a very broad portfolio.

Not that we wouldn't at some point make additional investments in diagnostics, but certainly right now, we feel like we're in a great spot. Some of the things we look at might be on the data business or the apps business. Again, we're not looking to move in a direction that would change our operating plan materially. You should expect us to be disciplined in terms of what we acquire. You should expect it to be synergistic, plug some kind of hole within the company that we believe is strategic and important and not be some kind of crazy left turn that has us going backwards in massive ways.

Speaker 1

I would just add on that front, a lot of the companies that we historically have looked at were investments that we would have made internally. They have an asset that is obviously additive to what we're building here at Tempus. A lot of those are kind of plug and play technology companies that we've looked at.

Speaker 4

Now, in terms of just overall, really quickly, I do think it's worth noting that AI is going to, you know, again, create some significant disruption in the space. We need to think, we need to think a lot about, you know, what the landscape looks like in a world where these large language models and generative AI are kind of producing unbelievable change. We could tell you, "Hey, this is our plan today," but bear in mind that as the market changes, we need to adapt to it. I want to be, I just want to, I don't want someone to say to me, "Well, you said this." If, like, at the end of the day, we have a plan, but we've also been very good at looking at the overall market and pivoting based upon what we see.

That kind of one of the things you want from a management team is that they're not going to get blindsided over time. In terms of MRD for three seconds, and Jim can jump in, we have a, we have what we believe is a really good portfolio across naive and informed and MRD. It's a massive space. It's a growing space. Right now, it's a relatively insignificant part of our overall business because we don't get reimbursed. We're not pushing it at scale. Assuming we get reimbursement, which we expect we'll get, assuming Personalis gets reimbursement, which they expect they'll get, we will certainly invest more in driving that volume in 2026 and 2027. I would suspect it'll be a catalyst to our unit growth. Right now, we're fortunate that our units are growing significantly without it.

We don't, we don't, we take the same approach to our data growth as we do to our genomics growth, as we do to ASP, which is when you have a business that's like growing this quickly, generating this much operating leverage at this high gross margin, we just don't feel compelled to like, "Oh, we got to get ASP up another $300 as fast as we can," or, "We've got to drive more unit growth as fast as we can right now." We're growing really fast. We're generating tons of leverage. The business is performing super well. As I kind of mentioned in my letter, we actually take a different approach, which is, how do we sustain this not in Q3 or Q4, but over the next three to five years, seven to ten years? How do we do that?

Most of the things we work on as a management team are, how do you build products and take products to market that are going to grow consistently over long horizons of time? If we get and when we get moments of additional tailwind, I would suspect that we'll try to be measured in terms of how fast we, you know, put our foot on the pedal and bring these things to market in a disciplined way, as we always have.

Speaker 1

Yeah, I would just add that from a margin perspective with kind of the launch of MRD. Obviously, we're gating volumes today given we don't have reimbursements. As we do get reimbursements, to the extent that there's any margin impact, we would be mindful of that to continue on the same path that we are in terms of profitability. The same kind of discipline as Eric described that we have today will continue with that even once we have reimbursement to maintain margins.

Speaker 4

Yeah, and just to add some color, we're in a great spot, right? We have a business that has lots of growth drivers. We have a business that has lots of little, you know, additional pockets of potential future tailwind or accelerance in the future, which is amazing, right? We have a data business that has lots going on that could be catalytic. We have an apps business that has lots going on that could be catalytic. We've got MRD that could be catalytic. We've got other products coming to market that could be catalytic, entering new areas in terms of Rare and Peds that could be catalytic. There are many, many things here that certainly over time should drive and propel our growth rate.

I would much rather have a company that grows at 25% for a decade than one that grows at 50% this year and 10% next year. We really want long-term sustained growth. As we think about all the amazing things happening here, we think about layering them in in a way that produces that sustainable growth.

Speaker 6

Again, everyone, if you would like to ask a question, press star then the number one. Next question comes from the line of Subu Nambi with Guggenheim Securities. Your line is open.

Hey guys, good morning. Thank you for taking my question. One is on a recently published paper that showed an AI algorithm that was developed to better stratify diabetes risk for patients with HbA1c levels. Could you talk about the development of this algorithm and any expectation as to when and where you could commercially offer it? Along those lines, as you think about the possibility of moving your Insights platform business into other areas than cancer, how should we think about it longer term? Thank you.

Speaker 4

Really quickly, with all where I was, we have a very broad portfolio of algorithms. As we mentioned, we do not talk about it a lot for a whole bunch of reasons. As you can see from this quarter's letter, we are a part of something like 2,000 publications and posters and papers and on and on. We have very, very deep scientific and mathematical efforts. We have very large product and engineering teams, a very large number of PhDs and MDs. It is over 1,000 technical people here working every day. These are large teams. We work on lots of algos, and these things get published, and they enter market. The challenge with all of our algos is we suffer, most of them, is that we suffer at the present moment from a fundamental flaw in the U.S. healthcare system.

Not blaming anyone, but it is a fundamental flaw, which is we do not have a mechanism today as a system to reimburse for AI or algorithms. We have mechanisms to reimburse for kind of wet lab work. You have chemistry, you have biology, I can pay for it. If you have an AI insight, much, much harder. The system is wrestling with that right now. At a federal level, they are wrestling with it. I suspect over time, they will find a way to pay for these kind of AI and data products, in particular algos, because they can just do amazing things. Every once in a while, you see pockets where it does get paid for.

For example, we have discussed historically that it was really nice to see Medicare, in particular CMS, paying for our FDA-approved algorithms that sit on top of electrocardiograms, of which we now have two. We have atrial fibrillation approved and low ejection fraction. Those get reimbursed at a stated rate of about $120 per algo. That is great. We suspect over time, hundreds of these things will be paid for, as they should be, because not only do they produce unbelievable patient benefit, but they also produce unbelievable economic benefit. You can do the math, right? You can pay for lots of tests at $50 or $100. If they save a $100,000 heart attack or a $200,000 stroke, it does not take a lot to be really accretive to the overall healthcare system.

Until they get paid for, these things are all going to be relatively nascent in terms of our overall financials, and we have called that out. Even though AI is influencing every part of our business, from diagnostics to data, the pure AI-based algorithm part of our business is going to be small until they're paid for. Once they're paid for, if they scale, there'll be really nice economic surprises. We don't forecast that till we see it. In terms of new disease areas, we have very large data sets in cardiology, in radiology, in pathology, and neuropsych is also a growing data set. Nothing compares to the size of the data set we have in oncology, so most of our data products and AI are in that space.

Over time, we would suspect that generating molecular data and producing biomarkers diagnostically will be equally important across most major disease areas. I can't imagine why it wouldn't. Those will also be drivers of our diagnostic business long term and our data business. Today, most of diagnostics is in oncology, and most of data comes from oncology.

Speaker 6

Thank you so much, Eric. Seeing no further questions, that concludes our Q&A session. I'd like to turn the call back over to Elizabeth Krutoholow for closing remarks.

Speaker 2

Thank you all for joining us today. We're available for any follow-up questions. We look forward to updating you again next quarter.

Speaker 6

Ladies and gentlemen, that concludes today's call. Thank you all for joining, and you may now disconnect.