Ladies and gentlemen, thank you for standing by. At this time, I would like to welcome everyone to the Tempus AI Third Quarter 2025 Financial Results Conference Call. [Operator Instructions] I would now like to turn the conference over to Liz Krutoholow, Vice President, Investor Relations. You may begin.

Thank you. Good afternoon, and welcome to Tempus' Third Quarter 2025 Conference Call. This afternoon, Tempus released results for the quarter ended September 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 and is available on our IR page. I would now like to turn the call over to Eric.

Thank you. Q3 was a great quarter all around. Our Genomics volume came in super strong with 33% overall growth with Oncology growing at 27% and Hereditary growing at 37%. We expect Hereditary growth will moderate a bit, although we now expect growth to be in the low to mid-20s as opposed to our previous guide of mid- to high teens. Our genomic growth was across the board. Really all of our assays did exceptionally well. And with MRD reimbursement on track, and our planned regulatory filing of our liquid biopsy xF later this year, we expect additional tailwind in that business, both from a unit perspective and revenue. Our data licensing or Insights business grew 38% in the quarter with an additional $150 million in total contract value, which was a super strong bookings quarter for us across multiple contracts that we highlighted in our letter. This is on top of the multi-hundred million dollar foundation model deal we struck earlier this year. So from a bookings perspective, our data licensing business is just really performing exceptionally well. The combination of growth in Genomics and growth in our data business allowed us to generate positive adjusted EBITDA for the first time this quarter, which has been a 10-year goal of ours and a key milestone. This was inclusive of several million dollars worth of additional expense from Paige, which is an acquisition we made mid-quarter. And even with that, we generated a positive EBITDA and would have been close to $4 million in adjusted EBITDA without Paige. So the business is doing exactly what we had hoped. We now expect for the year to be slightly positive adjusted EBITDA, and that's even with the additional several million dollars of drag from Paige. So all in, the business is performing well. We're growing at a rapid pace, and we're managing our costs to generate leverage in the business, which is exactly where we want to be. With that, take some questions.

[Operator Instructions] Our first question comes from Ryan MacDonald from Needham.

Congrats on a great quarter. Maybe, Eric, to start just on the -- in the Genomics business. Obviously, Oncology portfolio continuing to perform very well and a great increase in sort of testing volumes there. Can you just talk sort of click -- double-click a little bit on sort of what you attribute that the great strength in the volume growth here? Are we starting to see sort of a broader market and industry shift to more NGS testing that's sort of just helping see more patients that are just getting sequenced? Or would you say that you're really starting to see a benefit from the execution changes and sort of the sales coverage here with the broader portfolio? Just maybe what sort of Tempus-controlled success, if you will, versus sort of broader industry and market tailwinds?

Yes. So at a high level, look, our success is maybe slightly different than some others. So let me just talk about, I think, what's driving ours, and then we can talk about some macro phenomenon. In terms of our success, it's predominantly related to the fact that our sales force is more efficient today than it was a year ago. We made significant changes to our sales force when we brought in our MRD portfolio. Any time you make changes to sales forces in this space, you kind of cause havoc. I think people don't really realize how much havoc you cause. And then we all talk about the havoc after it's been caused. We certainly did cause some havoc, which was unintentional, and it's taken us several quarters to work through that. Our sales force is now kind of efficiently trained and doing its job, and so we're benefiting from some of that. And the second is that our technology, which is really tightly integrated and allows us to deliver highly contextualized comprehensive results to physicians is picking up steam as more and more doctors want us to deliver results that help them treat patients in a more comprehensive and more efficient manner. So we're kind of benefiting from those 2 trends. What I think broadly, people are benefiting from certainly, I think testing volumes have been healthy as more and more biomarkers are identified, people are looking to make sure their patients are tested. And so I think that's a general tailwind to the space. And then I think certainly, there are some companies who might be benefiting from the fact that they only offered solid or only offered liquid and so maybe they're now doing more concurrent testing or maybe there's some sequential testing. We're not benefiting from nearly as much of that because we've had a comprehensive portfolio in place for years now. So we don't see any of those kind of onetime benefits. So our unit growth, at least to us, looks really healthy and durable by virtue of the fact that we're not being artificially propped up by some kind of onetime benefit in either solid or liquid assays that's driving the majority of that gain.

Our next question comes from Mark Massaro from BTIG.

Congrats on a good quarter. I wanted to ask, Eric, maybe can you just -- there's a lot of interest not only in AI and big data, of course, but there's a lot of interest in MRD testing. And so I was just wondering if you could give us an update on how you're thinking about going to market in the clinic with MRD, recognizing that you have a partner in Personalis. I'm just curious how -- whether or not your team is trained, I believe they are. And just can you give us a sense for how fast you might go assuming reimbursement comes in over the coming weeks or months, how do you plan to sort of leverage your large sales team and go to market against a couple of other pretty significant labs in the space?

Yes. I mean -- so at a high level -- first of all, at a high level, when you have kind of 27% unit growth, leaving aside the Hereditary business, we're operating at a unit growth, which to us is quite healthy. As we've said historically, and we actually, in our letter, have called out that we expect to grow at about 25% for the next 3 years. So that's a fairly exceptional amount of growth. So given our size and scale. And so we don't want to grow 40% this quarter and then grow 20% in Q1 of next year. Like we want sustained long-term unit growth and revenue growth, and we feel like we're in a really good spot to deliver that. So I wouldn't expect us to like get MRD reimbursement and all of a sudden try to like jam as many tests as we can into the market, whatever that means, and kind of artificially buoy our growth rates. I would expect us to kind of dial that up every quarter in a more aggressive manner as reimbursement makes that more affordable. And we will do that. We have a really good portfolio of both naive products and informed products that span CRC, breast, lung, IO. And we've got a whole bunch of -- which we also talked about in our letter, a whole bunch of new studies being run with even a more sensitive version of our tumor-naive assay. So we're investing heavily in the space as is Personalis, and we have a really nice portfolio of tumor-naive and tumor-informed MRD assays. And we will certainly leverage our large sales force. We also have a subset of that sales force that's well trained in MRD, and we'll continue to dial that up. I wouldn't expect us to do anything unnatural in terms of investments in the sales force or anything unnatural in terms of growth, but it will certainly help us. It's one of the elements of tailwind we have that we believe can propel us to 25% growth in that space for the next 3 years. And if you kind of look at the size of our business and go out 3 years, you're looking at a pretty large Genomics business in Oncology at that point.

Our next question comes from Dan Brennan from TD Cowen.

Congrats on the quarter. Maybe just on the new contracts, Eric, the company hasn't really been disclosing, I don't think, new bookings. You had the Pathos deal earlier in the year, but obviously, I think it's been an annual basis. So just kind of walk through the $150 million. You had a lot of details in the press release, all the different customers. But just can you fill us in a little bit about why disclose this? Like kind of why did these come together here? Maybe if you want to update us on what the backlog looks like today since you're giving us the bookings number. Just any more color on this trajectory and whether there was -- were you expecting these this year or next year? Just any more color you can provide since it is a pretty differentiated call out in the quarter this time.

Yes. I mean -- so I think, first of all, we have -- we try to provide some color in previous quarters as to the size of some of these data deals. So we -- this isn't the first time we call out at a customer level or even at a kind of a dollar level, the size of these deals, including the fact that we called out that with the AZ, Pathos deal with several hundred million dollars of additional data licensing. So we try to call these things out when they rise to a level that we feel like we should call it out. So in other words, if we have a -- if we're just closing contracts in a normal cadence, we might just refer to 1 contract or 2 contracts. If we think something bundles together in a way that's worth calling out and worth highlighting, then we highlight. There's no rhyme or reason to why this quarter versus other quarters. We don't want to be in the habit of every quarter being like, oh, our bookings was $56 million or $152 million or whatever, $212 million because it's just -- it creates noise as if that number somehow translates into revenue in the next quarter, and it doesn't because these bookings, like all of our bookings are over multiyear. So if we sign $150 million in data licensing today, it doesn't mean my revenue next quarter or next year is going to go up $150 million. These are typically multiyear deals, and so we try not to cause a havoc. Our total contract value is in a great spot. We'll disclose it at the end of the year. We told people we'll give that number annually. But it's obviously -- we've already told the world…

Our next question comes from Casey Woodring from JPMorgan.

So starting off, just congrats on another strong quarter in core Oncology volumes. You had another competitor come out recently and also report strong liquid therapy selection volumes. So just wondering if you're seeing a similar pickup in xF and more of a marketed shift towards liquid? And then as a follow-up here, you talked about plans to submit xF for FDA approval in 4Q, followed by a full PMA submission for xR. Once you get FDA approval for those tests, I assume they would be eligible for ADLT status. So can you just walk through how you're thinking about the potential upside to the Medicare list price for those tests over the next year? And what we could think about as a benchmark really for the price that you'll try to get for them?

Yes. So in terms of -- so Tempus is unique in that we are now considered strong really across the entire continuum. So we're strong in Hereditary profiling when people are at risk. We're strong in therapy selection, either solid tumor or liquid biopsy, and we now have a strong offering in MRD and monitoring. So people kind of look at us end-to-end. So the interesting thing is we are probably in a pretty good position to see some of these big shifts, and we didn't see that. So we had really good growth in our solid tumor assay. We had really good growth in liquid. Nothing stood out at us as like a fundamental shift from solid to liquid. We had really good growth, certainly year -- prior period over this across both. So that said, I would agree that if with certain studies like, for example, SERENA-6, some of these studies where you might have more repetitive liquid testing, I could see over time, there being some additional volumes to our liquid portfolio that we and others might benefit from. But at the present moment, I haven't seen any seismic shift, although, again, I think the growth prospects for solid are great as more and more doctors order it and liquid probably even better because you're going to benefit from some of that serial testing.

And then, Casey, from a reimbursement perspective, as we've said, we have the long-term tailwinds remain there. xT CDx, we ended the quarter with about 30% of the volume that had been migrated. We now have plans to move the majority of that over to the FDA approved or ADLT version throughout 2026. In the letter, you also mentioned that we're submitting xF to the FDA by the end of this year. Obviously, that's a long process, so we can't speak to specific reimbursement levels. But certainly, ADLT typically provides upside from where we're at today, and that will follow by xR. So our viewpoint, total reimbursement on average is $1,600 for the third quarter, so up about $20 sequentially, but still well below parity with our peers. So given kind of these efforts, these regulatory filings, that certainly will help us close that gap.

Our next question comes from Doug Schenkel from Wolfe Research.

This is Colleen on for Doug. We have a question about Ambry. Ambry continues to perform well and ahead of expectations. We believe that last quarter growth was driven about half by share gains and half by organic expansion. Can you clarify what the mix was this quarter? Also, a competitor reported last night that its Hereditary cancer volumes grew low double digits in Q3 should we, therefore, be thinking about industry growth in the low double-digit range as a reasonable baseline? And within that context, can you elaborate on how Ambry's growth compares to the broader market? And then finally, on Ambry, can you clarify the mix of panels, like larger panels like cancer next versus more targeted panels and how that impacts how we should be thinking about the ASPs going forward?

Yes. So I'll start and then Eric can chime in. So similar to last quarter, about 50% of the gain is coming from share gains. As we highlighted in the letter, we expect that to moderate in Q4. And so we think kind of low to mid-20s is a more likely scenario than kind of where we're tracking today. Obviously, in terms of competitors, we can't speak to the share gains that -- or growth rates that others are experiencing. But Ambry continues to do well, both with bringing on new customers that are previously utilizing our competitors and then also continuing to expand kind of share of wallet with existing accounts. Eric, anything you want to add?

Yes. I mean in terms of the overall market, I would think that -- I think the space is much stronger than people thought. We've said that now on the last several calls. So I think whereas people thought this space might be kind of flat to anemic growth, you're now seeing people be like, oh, yes, we're growing in low double digits, which I think is probably right. We suspect that our Hereditary business will grow in the low to mid-20s, so kind of significantly above that by virtue of the fact that we have kind of the gold standard assays in market today in that space. Look, it is possible that you're going to see growth rate in the high 20s or low 30s. I mean that could easily happen, whether it's in Q4 or Q1 or Q2. And like we have historically, we're going to call out that I wouldn't expect that to continue as a long-term trend. We think a long-term trend, low to mid-20s is -- it feels pretty healthy to us and achievable, and that's where that business is. Do you want to cover the ASP piece?

Yes. And then in terms of kind of breakdown of assays, we don't disclose the assay level detail. The ASPs have been pretty consistent over the last couple of quarters, down a little bit year-over-year as one of our larger payers kind of renegotiated agreements. But overall, pretty stable in terms of the Hereditary space. The only thing that will impact ASPs is the rare business is still a relatively small component of overall testing for Ambry, but that comes with a higher ASP. So as that continues to scale, then that will have some impact on ASPs as well.

And I would just add to that really quickly. There aren't a lot of rare companies out there. I mean, we are now at some size. There's a few others. Obviously, GeneDx is well known. But there's not many. And I do think that we will make real ground over the next 12 to 18 months in becoming a very big player in that space.

Our next question comes from Michael Ryskin from Bank of America.

I want to follow up on the last one on Ambry, but maybe tied into a bigger picture one. Just if I'm looking at the guide, the raise for the guide for the year looks like you bumped it up effectively for the 3Q beat. But just your comments on Ambry just now, if you're going from mid- to high teens to low to mid-20s, by our math, that adds about $20 million of revenue to the full year. So is there something else that's offsetting it where you're taking something out of the legacy Genomics business or maybe data and services? Just if you could talk about the bridge a little bit and sort of how that rolls up to the full year, that would be helpful.

Yes. So I'll start and then Eric can chime in. So the Q3 growth rate was about 32% for Ambry. So we're saying it's going to go from 32% down sequentially into Q4. So not an increase in Q4.

But even still, let's assume that, to your point, if Ambry is outperforming by x amount of money, call it, $15 million or $20 million a year, and that might equate to a $5 million benefit in Q4. We just take the approach that we've always taken, like we try to look at it and say, if we have a beat, beat and a raise, that's great. But we don't need to get ahead of our skis. There's no benefit. We want to be in a place where we're consistently overperforming, outperforming expectation. And we don't need to artificially raise expectation for no reason, especially when the core business is growing at 30%. If we were growing at 4%, we might be like, oh, God, we need to raise expectation. But our business is growing at a really healthy rate, and we want to constantly orient people around whether we grow at 31% in Q4 or 29% or 30%, that doesn't really matter. What really matters is -- can we deliver 25% growth, not just for the next 3 years, but for the next 10 years? If we can, this will be a very, very big business. So we're architected around long-term growth, not short term. That's how we guide.

Our next question comes from Subbu Nambi from Guggenheim.

This is Ricki on for Subbu. There is a bit on this in the letter, but could you share any updates on your work on the foundation model with AstraZeneca and Pathos and maybe what the next milestones we should be looking for here are? And is there any benefit you could speak to from the Paige acquisition in the foundation model work?

Yes. So the foundation model is just finishing the pretraining phase right now. It's going exceptionally well in terms of like the all the -- you run all these small models, both single models and multimodal models and see how they perform and are they predictive and you're measuring them against kind of these common benchmarks like C index to see how they're doing. All that's going incredibly well. The teams feel great. We're kind of entering the phase of large compute over the next several months. And then when that is done, we begin post training later this year, kind of early 2026, and we expect to have kind of the first versions of the model in Q1. In general, the team is super happy with the progress we're making, both on every side. And so there's no kind of red flags. And I would -- we're in the midst of procuring additional GPU capacity. We feel like this is just an advantage we have, and we want to lean into it and double down. And we're going to address our -- if you look at -- and we called this out in the letter, if you look at Tempus relative to other companies, we're going to look and smell and feel like a tech company in many ways, including lines of code we write, amount of money we spend on cloud and compute, number of software engineers we have on staff. And we're in a world where AI is coming and we happen to be perfectly situated, we think, we're investing in that heavily. And I think instead of us taking our foot off the gas, we will continue to press forward. Paige is awesome in that they have their own foundation model work going on in digital pathology. They have a tremendous team and have made really interesting progress there. Those teams are now connected. They're now part of our foundation model team. We're aggregating some of that data and trying to understand the insights. And so there's just quite a bit of good momentum that comes from that. And we're excited to see where it goes.

Our next question comes from David Westenberg from Piper Sandler.

I'll focus a little bit more on the long term. Generally, the reimbursement system, CPT codes, et cetera, have generally worked on reimbursing for what you're doing in the wet lab. Now you've accumulated a lot of data and you have a lot of strong analysis interpretation. Do you believe that the health care system can effectively start to reimburse for really the challenges around data interpretation and analysis? And do you believe there's still a major -- or do you believe there's still maybe a differentiation with what you do in wet lab with, say, air correction?

Yes. I mean -- so look, when we think about the business, and if you look at the kind of guide we laid out the longer-term guide of growing at 25% for the next 3 years, we build that guide almost entirely looking at the growth we can see in our diagnostic business and our data business because those are big businesses, predictable, operating at scale, really good growth rates, really good margin. We understand them. We have a very hard time predicting the growth rate of some of these algorithms we have in market, effectively, to your point, this dry lab CPT code stuff. We have a hard time predicting the revenue associated with that because at the present moment, it isn't well reimbursed, if at all. We believe at some point, that will change. We believe at some point, that has to change or the health care system in this country is in danger of real problems. We just can't afford $5.7 trillion a year, growing at 7.5%. Their only solution to this problem is some amount of intelligence, call it AI, that allows us to understand where error is occurring, where waste is occurring, where mistakes are occurring, where we can be predictive and preventative, that's going to have to be paid for or it isn't going to scale. When that's paid for, Tempus is in a really unique position because we have a lot of this. We invest a lot of money embedded in our results, even with positive EBITDA, generating a ton of algorithms. I mean a lot. We have algorithms in digital pathology, radiology, cardiology, neuropsych, oncology, up and down the spectrum. And so when these things are paid for, we can distribute them across the over 5,000 hospitals connected to our ecosystem very quickly. And many of these things are already FDA approved, and so we suspect our path to reimbursement will be very quick if there is a path to reimbursement. And if -- and I've said this historically, if Tempus ever has its NVIDIA moment or whatever that moment is, it's going to be because one of these things starts to get paid for or 2 of them or 3 of them, and they just scale rapidly. So in the wet lab, you might go from $100 million of revenue to $150 million of revenue, that would be a very heavy lift. But in the algo world, you go from $100 million of revenue to $1 billion of revenue overnight because you're distributing zeros in 1s instead of having to kind of collect biospecimens and run a test and distribute it. So it just scales differently. So I'm hopeful they will get paid for. I can't see any other way out of this mess, and we're well situated.

Our next question comes from Mark Schappel from Loop Capital Markets.

Eric, a question on Paige AI. In addition to their AI pathology applications, I believe they also bring some synergies and leverage to your genomic diagnostics business. I was wondering if you could just provide some additional color or details on how Paige actually complements or works with your diagnostics business.

Yes. I mean it will work beautifully. Obviously, we just acquired Paige like very, very recently. So some of these things are being integrated now, but I'll give you just one example of ways in which digital pathology can enhance sequencing. So first of all, some percentage of the time sequencing doesn't work. It just doesn't work. You can't sequence the patient. Now it's a low percentage, but it's real. It's called kind of QNS. The results just don't -- they aren't delivered. Or some percentage of the time, you don't get enough material to even run sequencing. You just don't literally have enough material, high enough tumor percentage to even sequence the patient. In these instances, today, we say to a doctor, I can't help you. I don't have a result. But in a world where you have these digital pathology algorithms that can be deployed that can predict the most common mutations that might exist from sequencing, and Paige already has some of these in flight with more coming, one FDA approved, others from the FDA, you can basically return results to physicians even when NGS fails. Likewise, you can imagine a world where a certain number of results are really critical to get very quickly. For example, if a patient has non-small cell lung cancer, you want to know if they're EGFR mutated in 1 or 2 days. And so another benefit of integrating these things is we will be able to make some number of predictions very quickly. So we've always thought that the winning answer here was through this kind of multimodal approach to looking at the totality of data that can be generated for a patient and producing the highest quality data-driven insights as fast as possible. And those are never -- typically never single data modality driven. So we want to live in a world where we're every bit as good at generating molecular data as we are generating digitized pathology data or understanding a CT scan or an MRI or mammography. And if you look at our investments, we make investments along those lines. And I think it will over time, similar to the way if you look at Amazon, let's say, 20 years ago, you may have said, oh, whatever, they deliver books or maybe they deliver books in consumer electronics, and they're not that much better than eBay. But if you start to fast forward 5 years, 10 years, you can see the differentiation by Amazon's ability to kind of give you anything you want instantaneously. And that's because of the investments they made in depth of product and speed of distribution. And we're making similar investments or at least the corollary of similar investments in our portfolio today.

And in interest of time, our final question comes from Dan Arias from Stifel.

Maybe one on MRD. You guys have been pretty clear about not having plans to spend a bunch of money on big studies, but it does sound like you're investing there. And so to the extent that, that involves R&D, is there data next year that we should look out for? It does seem like we're going to have a whole slew of high-sensitivity assays coming to the market over the next 12-plus months. So I just want to make sure we have our eyes on the right things and updates from Tempus within that discussion.

Yes. I mean I would say we put out -- and I think this is called out in our investor deck, like it's -- we put out publications posters presentations constantly. I mean it's a crazy number. I just looked at the SITC press release, it's got like 7 papers coming out or something. So we put this stuff out pretty regularly. In terms of big studies, I think we called out in the letter that our -- on the tumor-naive side, we're in CRC today. We're running a non-small cell lung cancer study right now. We likely will go back and look at some of our CRC work. And I suspect you'll get some data coming out about both of those next year. Beyond that, we might bleed into early '27 in terms of other disease areas or other disease indications that we go into. But we expect to have really interesting data in market next year from our tumor-naive assay in both lung and CRC. And we believe we're hitting metrics that are just super powerful on the tumor-naive side that will allow us to kind of go head-to-head against some of the tumor-informed guys by virtue of some of the enhancements we've made internally with -- we have 400 PhDs around here. So it's a fairly large and talented technical team. In terms of tumor-informed, I'll leave it to Personalis to kind of provide you their road map of what's coming and what studies they're doing, but they too are investing, I think, quite heavily.

That concludes the question-and-answer session. I would now like to turn the call back over to Liz Krutoholow for closing remarks.

Great. Thank you. Thanks all for joining us today. We look forward to updating you again next quarter.

This concludes today's conference call. You may now disconnect.