Q2 2025 Earnings Call Transcript

Thank you for standing by, and welcome to Tempus AI Second Quarter 2025 Financial Results Conference Call. [Operator Instructions] I'll now turn the conference over to Liz Krutoholow, VP of Investor Relations. You may begin.

Thank you. Good morning, and welcome to Tempus' Second Quarter 2025 Conference Call. This morning, Tempus released results for the quarter ended June 30, 2025. The press release and 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.

Good morning, everyone. So I'll just 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 tenth anniversary, which is great.

Revenue increased 89.6% to $314.6 million. Our 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, and our hereditary testing, which was 32% in the quarter. Data and Services revenue increased 35.7% to 70 -- roughly $73 million, led by Insights, which is our data licensing, which grew almost 41%. Quarterly gross profit was $195 million, roughly 160% increase.

And 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 a negative $5.6 million this quarter. So 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 is roughly -- would be 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 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. So all in, the business is right where we want it to be, and we're making great progress.

And with that, happy to take questions.

[Operator Instructions] Our first question comes from the line of Daniel Brennan with TD Cowen.

Congrats on the quarter. Obviously, ask a multiparter upfront. 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 this strength came from, tissue versus blood, broader market share versus share gains? Next, Ambry, really strong quarter there. I know in the script, you talked about volume outlook and the deal model could be too low. Just kind of what are you seeing there?

And I know you also talked about the rare kind of genetic disease for kids kind of is that contributing today? And then finally, insights, nice quarter there, 40% growth. Any color on 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.

So there's a lot there. So I'll start at the beginning, maybe Jim could jump at the end. But 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 have made across our technology stack. And we just saw a bunch of those efforts kind of pay off in Q2. The volume growth has 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. So we feel like we're in a great spot in our oncology business. The hereditary volumes are also significantly stronger than we expected and they do 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. And so we're capturing a lot of that volume. And it's also that the overall space appears to be experiencing some really good tailwinds.

And as I called out in my letter, it makes sense. When we acquired Ambry, we had hoped that this space would be quite large over time. People thought it was kind of getting the saturation of maturity, and we historically have said we think that's exactly wrong. And we tend to believe that's the case.

There are just far more people that are at risk of getting cancer or risk of getting disease than those that have the disease. And so 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 #2 player in that space.

And even though it's small, it's growing pretty rapidly. And I would suspect over time that business will grow as well. So 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. So that's super exciting. That's under construction. And we've got 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 -- 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% to 30%.

And so we grew above that this quarter, which is great. So all in, I would say, if you look across the 2 main drivers of the business, sequencing diagnostics and data, we're just in a really strong spot.

Next question comes from the line of Yuko Oku with Morgan Stanley.

Given the significant progress you already made in first half migrating xT volumes to CDx in addition to ASP drag from xM ramp. How should we think about cadence of ASP as we think about path towards the goal of migrating 40% of xT volume to CDx exiting the year? And then 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?

Yes. So I'll take the first question on kind of ASP and then Eric can speak to kind of future products. So on ASPs, 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 about 28% in Q2. And then 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 xG. We would think over the balance of the year as we kind of approach 40%, you would see kind of incremental gains to overall ASPs, and on the xM side, that will continue to grow as well. But we are still gating that volume. So we're not anticipating xM kind of creating -- offsetting all of the gains that we have.

But we won't see kind of a significant step-up as what we saw in Q1, but we would anticipate kind of small incremental gains over the balance of the year.

And in terms of RNA sequencing, we've been doing whole transcriptome sequencing since the beginning. It's just kind of the way we started when we open up our lab. We were always doing targeted panel sequencing on the Genomic side, but doing whole transcriptome on the RNA side. And really, 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. So at the end of the day, it's just a really comprehensive panel. So I think likely the 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.

And we expect if you kind of fast forward 5 or 10 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 methylation markers throughout those assays.

Next question comes from the line of Rachel Vatnsdal with JPMorgan.

Perfect. So I just wanted to dig into the trial and insights businesses a little bit. First up, can you just talk about how bookings trended in the quarter for insights and trials? And 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.

And then on Pathos, just 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?

Yes. So I'll start -- I can start at the beginning, Jim can talk about the Pathos and AZ revenue. So our bookings, obviously, were super strong in the quarter because the AstraZeneca Pathos deal was actually an April event. So in terms of bookings, it was a large number in Q2. And our total contract value, which we only report out annually is not material, as you can imagine, by virtue of that deal.

So 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 want it to be, so which means we're in a great spot given that we had lofty ambitions. The clinical -- the trials booking business that's never been a large driver of bookings, our clinical trial matching business. In large part because 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. And so it's just not a huge dollar driver at the present moment.

As that business scales, it might become a bigger number. But right now, it's still fairly small. But 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 just now disclosed that we're connected to more than 4,500 providers and institutions across the country, which is obviously a huge number.

And these connections are deep. These are basically bidirectional connections where people are allowing us to pull data down. We're generating insights or diagnostics in putting those -- that information back in. And one of the parts of that is can we look at health care data in real time, basically data from the electronic health care records.

And 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 is on their own path. And increasingly, we're doing that, which means the kind of important role we play continues to go up. And so I think the clinical trials time business 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.

Yes. And then on the AZ Pathos deal in terms of kind of contribution in the back half of the year, we obviously didn't get a full quarter's worth of revenue in Q2 given the timing that we signed that deal. So 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.

Next question comes from the line of Michael Ryskin with Bank of America.

Following up on that last one. Just wondering, you've made a series of major investments in 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 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? And then if I could tack on a second 1 on that is a lot of focus, obviously, on profitability and EBITDA and get progress in the quarter. Just 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 incremental dollars.

Yes. So I'll start, maybe Jim can also jump in. But I think in terms of our pipeline, it continues to be super strong. I mean we have, as we called out in the past, I mean, the only real kind of market impact we felt was several years ago when biotech funding essentially dried up, and we just lost like a huge body of customers that in 2022 were flush with cash because they all went public.

But we're like long past that. So that's long in the rearview mirror. And 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. But our -- in terms of the big pharma we continue to have a really healthy, robust pipeline of very big deals that we expect to close over some horizon of time.

And 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 multibillion-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. But 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 $100 million investment over several years. So we have not felt any pressure.

We also have not felt any pressure from competitors in that space. I mean it's like 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. And so we -- 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.

So we have not seen them, and that hasn't changed recently. So 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. But this is going to be one of the questions that we have to think through.

It was important for us to kind of -- as we approach our tenth anniversary be able to build a business that was growing quickly and still generating operating leverage and getting to adjusted EBITDA breakeven and so as I mentioned in my letter, one thing 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. So we're doing a really nice job of growing and being disciplined and I expect us to be disciplined until we round adjusted EBITDA positive and round cash flow positive.

But 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, and 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 health care 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.

Yes. And 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. And so 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.

Next question comes from the line of Ryan MacDonald with Needham & Company.

Needham & Company, LLC, Research Division 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 '26. 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 '26. And then on the Genomics side, it clearly seems like there's -- the MRD portfolio is resonating well within the industry.

Just curious when we could see maybe the bigger unlock here with reimbursement from MolDx and when we should expect that to happen if that's a second half of' '25 or '26 event?

Yes. So again, I can start. So I think, look, that's one of the great questions that I think 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.

But so far, obviously, we haven't announced anything. And I think one of the questions we have for that group is the same when 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. And that's something I don't think people are really focused on. It is -- 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. This is massive data repository. And you're essentially running compute for months on a cluster of over 1,000 H200s, which are -- that's not a small amount of capacity. We're talking hundreds of billions of tokens where you're running compute.

And you're likely going to see associations that you just couldn't see until you ran that amount to 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. And 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.

And right now, these drugs are very brute are 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. And what AI is likely going to do with these foundation models should do is provide insight to physicians and patients as to who's likely to respond and not to respond. And so I would suspect that will change -- fundamentally change care and probably much faster than guidelines can adapt because these guidelines are human- oriented, not AI oriented. So I would suspect that AI is going to have a very big -- create some very big disruption within the space.

And with any type of disruption, you have to kind of ask yourself, do I want to be a disruptor or don't want to be disrupted. And it was very cool to see AZ taking a lead in that, and I suspect others will. And over time, I suspect everyone will. I just can't see a world where people are like, no, I just don't -- I don't need that kind of data.

I'll just do it the old way. So I think it's -- long term, I would suspect many people will build similar models or avail themselves with similar models and move significant dollars from historic chemistry and biology to data and AI and we'll certainly be -- 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 2 or 3 or 4, like 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. And so we're fortunate that we don't need to sign new things in the future to generate really good growth. In terms of MRD and then I'll give it to Jim. I'll speak to the volumes, he can cover the -- when is it going to be a big 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, broad from a perspective that we're in multiple different disease areas in breast cancer and lung cancer and IO response and CRC. And so we cover a significant variety of assays in that market and it's a really exciting growing market. That said, we are gaining volumes until reimbursement and maybe you can provide some context on that.

Yes. So we've previously disclosed that we anticipate getting reimbursement by the end of 2025. So no changes on our assay. And then obviously, Personalis has publicly disclosed kind of their timing as well. So built into the guide is not a meaningful uptick of MRD revenue for '25. We would anticipate that occurring more in '26.

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

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, and the comp steps up quite a bit in 3Q and 4Q, 3Q is actually a pretty stiff compare. So where do you think growth lands in the back half of the year, is 30-plus still kind of okay to model?

Yes. I mean, I think for the year, we've talked about around 30% growth in the data business or slightly above. I think the sequential -- Q4 is always the largest revenue quarter for us. So 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. And so we would anticipate a tick down in landing around or just north of 30% for the year.

Okay. So just to be clear, that's like a 20% growth rate for the next 2 quarters or so you're basically kind of halving where you are today?

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

We also -- we don't want to back into providing guidance by business unit on a quarterly basis going forward. But the data business is a great spot.

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

Congrats on the good quarter. I wanted to ask a question about your liquid biopsy business. I recognize that xT and 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. So my question is on your xF franchise.

What are you seeing in terms of demand as far as time points goes? And how do you believe you're positioned competitively recognizing that there is another player that's pretty large in the space. But how do you think that you're positioned to compete directly against them?

Yes. I mean, so 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 real scale. I think we've disclosed historically it represents about 1/3 of our volume give or take or something that. Yes, 25% to 30%.

So if it's a significant component of our volume, and we've grown much faster than the market. So net-net, that franchise from a therapy selection perspective is in a really strong spot based on historic performance. And we have an assay 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 digital disease but also looking at treatment response monitoring. And over time, we would suspect that this is a bigger part of the market, but it also requires reimbursement, and so 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. And we have been more disciplined than that.

So I think over time, these will produce lots of additional tailwind to our unit volume, but they require reimbursement. And so right now, they're small, but we have a super competitive product set.

Okay. And then my second 1 is just on Ambry. Certainly, a lot stronger than we were modeling. I think you alluded to 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? Because I'm just trying to get a sense for some of the puts and takes to explain the strong growth in the quarter.

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 Ambry is on its own just by virtue of like having a great product. We are 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 percentage of the market. And so more and more big systems are just migrating to Ambry because it's kind of the gold standard in that space and people want all that comes with it.

The turnaround times, the best-in-class error rates, the technology they've wrapped around it, the analysis that they provide, the experience they provide. So it's just a really good product. They're also -- they've also made a ton of investments in rare and peds. And so we're super excited about the growth of that product set, which is still quite small in relative terms but I would suspect it will be a big driver over the next several years.

So it appeals to us the core of the business, the core of the growth feels durable and sustainable. But again, we're not here to say, you can kind of bank on 30% growth for the next 5 years. We're going to -- as we have tried to be historically, we're going to be conservative and until we see these things play out.

Next question comes from the line of Andrew Brackmann with William Blair.

Tempus 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. So 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?

Yes. I mean, so I think we have historically been opportunistic but not overly acquisitive. And 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 like taking a turn and going left after we've been going straight for a long time. So 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, whether it's in our data business or in our diagnostics business. But we continue to be measured and disciplined.

And I would suspect us to -- if you look at our last kind of 6 or 7 years in the companies we've acquired, my guess is we'll take a similar approach in general going forward. But again, the market is changing pretty dynamically. And so 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 be in.

Next question comes from the line of Doug Schenkel with Wolfe Research.

Just a couple cleanups. Maybe a follow-up to Andrew's question. Is there a good rule of thumb on just how you're going to think about partnering versus organic? 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. And then 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 the next few years, how big as a percentage of oncology volumes, would you expect MRD to be?

And how does that affect the gross margin trajectory over time?

Yes. I'll start and Jim can jump in. So I mean, I'll start at the beginning. So 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 takes us -- would take a giant step backwards where we're like, "Oh, no, we're just kidding." So I mean we kind of start from if we're going to buy something, we don't want it to derail our organic plans.

So we kind of start from that lens. Not that it would be impossible, but in general, we start from that lens. And 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.

And so we really feel like we've got a very broad portfolio, not that we wouldn't, at some point kind of make additional investments in diagnostics. But certainly, right now, we feel like we're in a great spot. So some of the things we look at might be on the data business or the apps business. But again, we're not looking to move in a direction that would change your operating plan materially.

So 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.

And 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. And so 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.

Now in terms of just overall really quick limit, I do think it's worth noting that AI is going to again, create some significant disruption in the space. And we need to think a lot about what the landscape looks like in a world where these large language models in generative AI are kind of producing unbelievable change. And so 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. So I want to be -- I just want to -- I don't want someone to say to me, hey, you said this, 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.

And that's 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 3 seconds and Jim can jump in. We have what we believe is a really good portfolio across naive and informed an MRD. It's a massive space.

And it's a growing space. Right now, it's a relatively insignificant part of our overall business because we don't get reimbursed and so we're not pushing it at scale. Assuming we get reimbursement, which we expect to look at, assuming personnels gets reimbursement, which they expect they'll get, we will certainly invest more in driving that volume in '26 and '27. And yes, I would suspect it will be a catalyst to our unit growth. Right now, we're fortunate that our units are growing significantly without it.

So 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 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. Oh, we've got to drive more unit growth as fast as we can right now. We're growing really fast, and we're generating tons of leverage. So the business is performing super well.

And 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 3 to 5 years, 7 to 10 years, how do we do that? And so 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. So 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 put our foot on the pedal and bring these things to market in a disciplined way as we always have.

Yes. And then I would just add from a margin perspective with kind of the launch of MRD. Obviously, we're gaining volumes today given we don't have reimbursement. As we do get reimbursement to the extent that there's any margin impact, we would be mindful of that to continue in the second half than we are in terms of profitability.

So the same kind of discipline, as Eric described that we have today, we'll continue is that even once we have reimbursement to maintain margins.

Yes. And just to add some color. We're in a great spot, right? This is -- we have a business that has lots of growth drivers.

And we have a business that has lots of little additional pockets of potential future tailwind or accelerants in the future, which is amazing, right? We have a data business that has a lot 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 have got other products coming to market that could be catalytic, entering new areas in terms of rare and peds that could be catalytic. So just a lot of -- there's like many, many things here that certainly over time, should drive and propel our growth rate. But again, like 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. Like we really want long-term sustained growth.

And so as we think about all the amazing things happening here, we think about layering them in a way that produces that sustainable growth.

[Operator Instructions] Next question comes from the line of Subbu Nambi with Guggenheim Securities.

One is on the 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 algo and any expectation as to when and where could you commercially offer it? And then along those lines, as you think about the possibility of moving your Insights business into other areas than cancer, how should we think about it longer term?

So really quickly with all of our algos. We have a very broad portfolio of algorithms. I think as we mentioned, we don't talk about it a lot for a whole bunch of reasons. But as you can see from this quarter's letter.

We're a part of something like 2,000 publications and posters and papers and on and on. So we have very, very deep scientific and mathematical efforts. We have a very large product and engineering teams, very large number of PhDs and MDs. It's like over 1,000 technical people here working every day.

These are large teams, and 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. health care system, not blaming anyone, but it is a fundamental flaw, which is we don't have a mechanism today as a system to reimburse for AI or algorithms, we have mechanisms to reimburse for kind of wet lab work. So you've got chemistry, you've got biology, I can pay for it. But if you have an AI insight, much, much harder.

The system is wrestling with that right now. I mean at a federal level, they're rested with it. And 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. And then once in a while you see pockets where it does get paid for.

For example, we've 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, and those get reimbursed at a stated rate of about $120 per algo. So that's great. We suspect over time, hundreds of these things will be paid for as they should be because of not only do they produce unbelievable patient benefit, but they also produce unbelievable economic benefit. And you can do the math, right?

You can pay for lots of tests at $50 or $100. And if they save $100,000 heart attack or a $200,000 stroke, it doesn't take a lot to be really accretive to the overall health care system. But until they get paid for, these things are all going to be relatively nascent in terms of our overall financials, and we've called that out. So 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. And once they're paid for, if they scale, they'll be really nice economic surprises.

But again, we don't forecast that until we see it. In terms of new disease areas, we have very large data sets in cardiology, in radiology, in pathology, in neuropsych is also a growing data set but nothing compares to the size of the data set we have in oncology. And so most of our data products in 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.

And so those will be drivers of our diagnostic business long term and our data business. But again, today, most of diagnostics is in oncology and most of data comes from oncology.

Seeing no further questions, that concludes our Q&A session. I'd like to turn the call back over to Liz Krutoholow for closing remarks.

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

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