IQVIA Holdings Inc. (IQV) Earnings Call Transcript & Summary

I think we can look to get started. Thank you, everybody, for joining us here at the Citi MedTech Access Day. I'm Patrick Donnelly, the tools and diagnostics CROs analyst here. Happy to have a unique opportunity here with Ron and Mike, incoming CFO, outgoing CFO. So it should be a good conversation. And Ron, maybe we can start with you and then obviously, we'll kind of work our way around. I wanted to start on TAS, not surprisingly, given the level of focus there. Had a pretty good 4Q. I think it grew about 7% constant currency. Can you just talk through the different components of that? Obviously, consulting is a focus of the AI stuff, which I'm sure we'll get to. But maybe talk through TAS and then we can kind of work through there.

No, you're right. We did have a strong fourth quarter. Actual currency, we were close to 10% growth, constant currency about 7%. And strong organic growth in TAS. And that was versus a pretty tough compare in the prior year where we grew over 9%. So we're happy with what we saw in the TAS business. And if you break it down into the components, historically, TAS is going to become a new Commercial Solutions segment with a few changes going forward. But looking back historically, we would break it down and say, okay, the fastest-growing piece historically has been a real-world business in TAS. Now we've shifted a good piece of that business to Commercial Solutions, I mean to R&DS going forward. But in the fourth quarter, TAS, the real-world part of the business in TAS grew double digits. And the other pieces, which is consulting and analytics, info and tech all grew in the low to mid-single-digit range. So overall, a really nice quarter for us. And I'll say, in particular, since you mentioned consulting, we're a little disappointed earlier in the year about the bounce back in that business, and we did see sequential improvement as we went to the third quarter into the fourth quarter. So that's a positive sign.

Yes. And then maybe just for people to contextualize, I think we've gotten a lot of questions recently about just the components of TAS in terms of what makes it up in terms of how big is consulting, how big is real world. Can you guys just help break down -- maybe we use legacy TAS.

Yes. We'll start with legacy TAS, about 1/3 of it is real world, about 1/4 of it is info. And then you have the consulting business, 20% and the balance is the tech business, which is low 20s.

Okay. Perfect. And then, Mike, maybe we can look at the go forward, the new Commercial Solutions segment. I think you're guiding for that to grow 8%. Your R&DS more mid-single, 4%. Can you just talk about the underlying growth components on the Commercial Solutions side? What part of the old TAS pieces are in there? It would be helpful to kind of talk through this.

Yes, I think it's a -- listen, we were -- as Ron said, we finished up the year with a lot of momentum, and we were encouraged by a lot of the leading indicators within the commercial business, things like drug launches are something we keep a close eye on. Our pipeline is good. Client decision-making was improving in that segment as well as within sort of the RDS segment. So I think our range was about 7% to 9% reported, like you said. And again, I think the growth drivers are the same as what we've seen historically within TAS, right? It's now the Patient Solutions segment, which is really the remaining pieces of that commercially oriented real-world business are still growing quite strong as well as the info and some of the other parts of the business as well, kind of in that low to mid sort of single-digit sort of growth range. So yes, we've got a lot of good momentum. And you guys probably saw that the Cedar Gate acquisition that we did at the very end of last year, that was a really nice platform in the U.S. to add to our payer provider segment, which is part of that real-world patient solutions part of our segment. So yes, there's a lot of good stuff happening there.

And then, maybe you mentioned the deal front. We always get questions about M&A contribution in each segment. Can you just help kind of frame up which -- what's the right way to think about M&A in each part for?

Yes. Within our guidance, I think we said about 150 basis points at the enterprise level. And it's still about the same, right? It's still about maybe 2/3 would be towards Commercial Solutions. And then the balance in RDS. I mean, listen, RDS growing at 4% is terrific, actually in this climate. And it was actually somewhat rewarding that we started to get the recognition, I think, from investors that there was a bit of a nice separation last year in the growth rates versus some of the other traditional CRO competitors there. But yes, we feel pretty good about this year.

Yes. And Mike, maybe just given the recent shifts in the segments, can we just talk through Commercial Solutions, R&DS, the margin profiles of both progression and how we should think about the 2 would be helpful.

Yes. From a revenue growth and margin standpoint, I don't -- you guys aren't going to see sort of much difference. The clinically oriented real-world pieces that we moved into RDS, which was ostensibly real-world late phase that people are most familiar with, had a very similar both margin and revenue growth profile as existing RDS segment. So you won't see anything there. And that part of the business had a lower margin profile than some of the more traditional TAS. So if you did nothing else, you would expect the margin sort of uplift. But then we move the CSMS business into kind of this new Commercial Solutions group, and that has a lower margin profile. So net-net, you won't see any changes there.

Interesting. And then, Ron, maybe on the kind of R&DS side, are you seeing pharma clients consolidating vendors more aggressively on the commercial side? How do you guys think about your share capture on the consolidation side? What's the opportunity there for you?

Well, look, we've historically talked in the R&DS business CRO industry in general is being very fragmented. And we've picked up share over the years there, and there has been some consolidation. But when you look at the commercial side of the business for pharma, the vendor landscape is even more fragmented. And it's in part because there are so many different components to the commercial business and in part because there's a lot of local competitors as well. But we are seeing consolidation, in fact, in the Commercial Solutions area. And I would offer up a couple of examples of that to demonstrate why. One is we're increasingly seeing pharma clients come to us and want to outsource chunks of their commercial organization, analytics, and even the end-to-end commercialization of products in certain cases. I think we announced on our fourth quarter call that we had our first deal with a large pharma client in Asia to take over end-to-end commercialization of their products. And the reason we can do this, and it's a good opportunity for us is we have -- of course, we have the data. We have the analytics. We have the domain expertise, but we also have a contract sales organization, which was part of CSMS, that is the biggest part of CSMS that we're now rolling into the Commercial Solutions business. So we can offer pharma end to end. And the reason our pharma clients are looking to outsource more is they're trying to decide how to allocate resources between new drugs that they're launching and existing platforms. And very often for existing drugs, they're looking for someone to help them out so that they can use -- utilize their resources, allocate them more efficiently. And we're in a very good positioned to do that given the assets that we have. And so we're seeing more and more of those sorts of deals come down the pipe.

And we even just got awarded another one of those programs this week. So it's definitely the trend is continuing.

Okay. Interesting. And then maybe we can talk some AI. I know, surprise. We waited 10 minutes to get into it. So Ron, maybe we can start and we'll kind of go through different pieces. But I think a lot of investors are concerned AI as a threat to kind of the core business. The data mode, I know you guys have talked about. Can you talk about just how that can stop AI vendors from partnering with pharma, bypassing you guys in terms of the data. I know that's a big focus. And obviously, Ari.

Sure. And Ari has talked about it on the call, and I'm going to repeat some of what Ari said because I think it bears repeating. I'm not sure it's fully sunk in yet in the market. But look, in order to be effective in deploying AI, you really need 3 things. You need data and you needed at scale and it needs to be fit for purpose. You need domain expertise. And finally, you need the technology and you need to know how to use the technology. And all 3 of those, we think, work to our advantage, not just are they competitive moats against our business being broken into or taken away by AI, but they actually offer us the opportunity for organic revenue growth. Increasingly, clients are coming to us for AI solutions. But let me talk a little bit about why this is a moat that is our data in particular. The first reason is that our data is proprietary. There are certain things out there where you've seen some articles about Anthropic and others putting out variations on their agents that can, for instance, analyze legal cases and things of that matter, maybe displace a LexisNexis or something like that. In our case, our data is not publicly available. In fact, it's all very proprietary data that we've collected over a very long period of time. The other thing about our data is our data is, I would say, I'll use the term, very messy, which is to say you need to not just be able to gather the data, but you need to be able to cleanse the data, bridge, code, anonymize, link the data. What we get in terms of data is extremely messy. I mean it's multiple oncologies. It comes from various spots around the globe. It's not like there's just one big trove of data there. The third distinction I would make about our data is it's very dynamic. It's never changing. It's not static. I mean it changes daily. We process more transactions daily than the New York Stock Exchange. So you have to be up to date with your data to meet your clients' needs. And the last thing I would say is the data is very complex and it's very sensitive. And when I say sensitive in the sense that we need to meet very strict privacy and compliance regulations in order to use the data. And those vary across the world. So you have to have experience working in very specific jurisdictions. It's just not like you can go out and get the same data and use it in the same way in Germany as you can in the U.S. as you can in Japan. You have to know how to work with the data. Now look, theoretically, someone could -- we get asked this question all the time. Could someone recreate what you have in data? And the answer is, theoretically, yes. But practically, it's very difficult because we've been at it for 71 years. We get the data from 150,000 different data feeds around the globe. And we do a lot -- we spend hundreds of millions of dollars annually on that data, and we spend additional hundreds of millions making that data fit for purpose. So yes, theoretically, someone could recreate our data set. Many have tried. Practically speaking, nobody has to date. It's a very difficult thing to do. So we think it's a very durable moat around that data. And then another thing that makes it durable is just knowing how to use the data, and that's a domain expertise. You need to know the business rules. You need to know how to work with the data. And it sounds perhaps trite or trivial, but it is. It's very, very tough thing to do. And the last thing on AI, you need to know how to use AI to use the data. A given process may take dozens of AI agents because you have so many subprocesses that need to be agentified and then you have to have one kind of orchestrating agent over top of it. So in order to use AI data, it's not a simple thing to do. And that's why so many of our customers are coming to us to ask for AI solutions because theoretically, a client could take the data and say, okay, we're going to try to do something with it. But it's very hard for them to do and their economies of scale and having us create solutions like launch planning solutions that we can sell across multiple customers. And in fact, there's another restriction there that when we license data to our customers, they have -- we do it for specific purposes. It's an end product that we're giving them, not all the ingredients that are underneath that. And we have -- they have to -- we have third-party arrangements or agreements that specify exactly how that data can be used. So it's not like a client could just take our data and go give it to one of the AI companies and say, "Hey, have at it", they would need our approval to do that. So there are many, many reasons that we don't feel we're vulnerable here. And actually, on top of that, many reasons, particularly the economies of scale of developing these AI agents for -- and selling them across multiple customers, we think it gives us good upside looking forward. And in fact, we haven't seen clients behavior change where they're not buying stuff from us that they did before. In fact, quite the opposite. They're coming to us and saying, "Look, we want to do things more efficiently". We got this huge staff of people, again, I'll go back to the launch planning example, working on taking various data sources and spending months doing -- planning for a launch, can you do that for us much quicker using AI. And that's just an example of one of the applications. In fact, I think we now by year-end, had done 150 different AI agents covering 30 use cases, and it's changed then since it's gone up. It changes daily.

Yes, this whole shift to AI is fantastic for us. I mean because -- I mean, listen, if you step back and go back to the theory of the merger of IMS and Quintiles coming together, right, the whole ethos of our company was taking that 70-some-odd years' worth of data in disrupting sort of clinical trial development. And what it's done now is that it's put it front and center in our clients' minds. So now the conversations with our clients, we are elevated in their minds, right? They're restrained on capital and they're sitting there going, at the end of the day, can they build some models themselves? Sure. But your models are only as good as the data it's trained on. We have the best data in the industry, both on the clinical and the commercial side. So the types of engagements now are completely sort of elevated, and that's why we have clients coming to us.

Yes, rough estimation of the amount of -- where pharma gets its data is about 3/4 of the data pharma uses comes from IQVIA across the industry. So that gives you an idea of how important our data is.

Yes. And one of the questions I get, and again, I think people are trying to get to the root of the AI question is where does the data come from? Why do you guys have so much versus others? Do you want to just spend a few minutes in terms of the root of the data and why it's such.

So where does it come from? I mean it comes from multiple sources. It comes from scripts. It comes from electronic medical records, claims data, lab data, genomic data. And one of the key is the ability to link data longitudinally on a de-identified basis. So I can go and look at your history, your person X, you're not Patrick. We don't know who you are, but we can trace your prescription habits, look at the electronic medical records, look at the claims data. That's one of the basis for using real-world data is being able to do longitudinal studies across large masses people. And like I say, it comes from 150,000 data feeds, and we are the only company that can do this globally. We have competitors on the script side in the U.S. or Germany or Japan or wherever, but we're the only person that can put together a global view for a brand manager of how the brand is doing. And that's a huge differentiator.

All of those thousands of business rules that govern what Ron just talked about, how you pull the data together that only we have -- our clients don't have access to it, which is why it makes sense for them to come to us.

Yes. Okay. And then another area of task from our conversations with investors, people are looking at is just the durability of the consulting business inside there. I mean what would you guys kind of frame up as a response to the consults?

Look, the first thing I would say is there will be small areas of our business that are going to be eroded by AI. There's no question. But I would emphasize small, limited. And an example of that would be in the consulting business and primary market research, in particular, in places like the U.S. where a lot of this data is publicly available. But to put a number on that, if you look at primary market research is maybe 5% of the 20% -- 5% of the consulting and analytics business globally, which is 20% of the TAS revenue. So we're talking about relatively small numbers here. And even within that 5%, not all of it is exposed. Now there may be other -- there probably are going to be other little pockets of low-end work that are going to be exposed. But the upshot of it is we really feel like the opportunities far outweigh what we're going to lose. Mike, maybe you can expand on some of those.

Yes. So I think kind of sizing me to the numbers Ron is throwing out there, I mean, you want to take a pessimistic case, maybe $100 million. That's the kind of sort of scale of revenue that we think could be at risk. But people really need to start pivoting and think about sort of the opportunities there. So -- and I'll give you a couple of examples on sort of several different dimensions about what we're doing. Let's just say from a revenue protection standpoint for some of that lower level kind of work, right? We launched last year our Agentic AI kind of like copilot kind of thing that we put on top of several of our offerings that we have out there, right? So we're modernizing some of the lower level to keep it relevant and to keep clients sort of stuck into us. Then you guys may have seen a press release we did with Boehringer Ingelheim fairly recently with -- that they were -- they bought our DaaS+, data-as-a-service offering. And that's an agentic AI offering, which essentially allows the client to take their subnational, national and global sort of prescription data with an information management platform to then help hook in sort of data that they have, so they have one single source of truth for all of their data. Now that inherently brings a lot of value to our customers. What it also does for them is that if they want that capability, they have to buy all of our data, right? So if there's -- it protects our data sales, right? But it also in areas where they may have had, let's say, a specific country, they were sourcing that script data in that country for somebody else, well, they have to now buy it from us. right? So that's another kind of example here. And then you have, and Ron mentioned it, kind of the more transformative offerings that are coming out on the commercial side, like our launch planning tool. And pharma employs thousands upon thousands of people who are specialized with, let's say, sort of market research and competitive intelligence and sales forecasting and things like that. At our Tech IQ conference last year in London, we rolled out and we have some clients on it, this launch planning sort of agentic AI module. And our teams came back and said, literally like pharma's jaws were like on the floor because one person can go and say, "I'm going to launch this drug, give me a global sort of launch strategy". And the agents are going through all of our data and all of those sort of verticals, the forecasting, market access, things like that and just spits out like a board-ready sort of like launch plan in a minute. Now that gives pharma options that they can either redeploy those thousands of resources sort of internally, and that brings a lot of value. So to Ron's point, any technology changes, some things are going to go away. And one thing on the clinical side, just to kind of give a corollary there, I was thinking back to like with all this AI hysteria going on at the minute here, if you go back to like early 2000s, right, the narrative at the time, I remember was going from paper case report forms to electronic data capture, the headlines were CRO's business models are going to collapse because they make a lot of money with people sitting there sort of like doing typing. And what happens? Like if you look back in retrospect, CROs didn't collapse, they grew. And the ones that adapted like IQVIA is adapting with this technology grew even faster. And the question is why, right? Well, in that specific instance, right? Yes, certain line items, let's say, of a CRO budget like double data entry, those went away, okay? But what that did do was is that because the data is now in an EDC coming in, in a more contemporaneous fashion, it allows for new services to pop up like centralized monitoring or risk-based monitoring or you can do more interim analyses because the data was coming in more frequently. So with any technological disruption, like there are going to be upshots. And the other thing that happened there is because it was easier to collect data, it actually helped study trials to become more complex. People started adding more assessment in the trials, and that's obviously a net benefit. So whether it's on the commercial side or the clinical side, there's a lot of opportunity there and our moats that we have around our data and our workflows is key.

Yes, it's easier to visualize the downside in AI than it is the upside. And trust us, what we're seeing is that there's going to be a lot of upside here.

Yes. And Mike, you've framed that $100 million number. Is that more -- obviously, Ron talked about the 5% of the 20% of the 50%. Is that kind of...

Yes, that's what I'm talking about. Yes, yes.

Okay. Got you. And then maybe on AI, just overall, improving efficiency and workflows in clinical trials, right? If time lines get shorter, burn gets faster. How -- what are you hearing in terms of how those savings, if you will, would be distributed to CROs, to IQVIA? What's the right way to think about this ecosystem?

There's no large pharma procurement people listening to this call are there? No. I mean, listen, just to give another kind of like framework here on how to think about this. You hear pharma companies out there now AI, AI, AI. And I think when Pfizer had their call recently, people were like, "Oh my God, Albert mentioned it 15x. Like what does that mean?" And most of what we're hearing from our clients is that where they're focusing AI first is like on the drug discovery side, right? So our theory, and it's a pretty prevalent sort of theory in the marketplace is that, that is going to be an area where we start to see the first kind of like step change benefits of AI is in the drug discovery space. So you can easily envision pick a time price in 3 to 5 years, probably somewhere in that window where more molecules are going to be identified with higher probabilities of success, which fundamentally changes the ROI calculations. So there's going to be, in our estimation, a lot of molecules coming through the discovery area into clinical. And by the way, I just mentioned discovery, we announced yesterday we bought a great asset off of Charles River's hands in the discovery space. So again, that sort of played into our thinking on that front. But pharma, and I think Pfizer even said it, pharma is going to reinvest the savings in the economies into improving their pipelines because at the end of the day, that's how pharma companies are valued, and they always have these patent cliffs looming over their heads. So there is going to be a virtuous cycle of this efficiencies. And to get to this specific part of your point about how this is shared, I mean, look, in the clinical space, I think people need to realize a couple of things. One is it is highly, highly regulated. And it moves very slowly. So even we have -- we're looking at quite literally 1,000 individual processes that underpin our CRO for identification, right? And we're going as fast as we possibly can. But once you develop an agent, in some cases, you can't just pick that agent up and deploy it on studies and flight, right? So there's that regulatory break to how fast this stuff can come through. I think the other thing people need to sort of remember here is that at the end of the day, we have -- with large pharma, we have deals and rate cards that are 3 to 5 sort of years long. So the fears of some immediate deflationary shock to the CRM numbers, it's just not going to happen, right? And generally, what happens kind of going back to that EDC analysis is that when we get more efficient, yes, we pass some of that sort of on to our clients in the form of negotiations. And in our budgets, some people kind of misconstrue that CRO budgets because they're heavily labor-based, they're not time and materials contracts, right? You have unit-based rate cards that have certain effort assumptions baked into it. But it's a unit. So if I can identify faster, right, I get to keep a big part of that upside for myself, and some of it will be shared to the negotiations. So at the end of the day, what we think is going to happen is that our margins are going to continue to improve in RDS, and we're going to continue to take share.

And pharma will -- with the better ROI that they're seeing on their investments in research and development is going to -- the big question is, well, are they just going to put that in their pocket and not develop additional drugs. And I would argue in what we're hearing from clients is the opposite. No, improving ROI means we're going to invest more in drugs because we're valued on our pipelines. And to the extent that we can put more drug candidates into our pipelines that have now better return on investment prospects, we will.

Yes. And Mike, you touched a little bit on it, but in terms of that upstream versus downstream, it sounds like your guys' view is AI kind of in the upstream molecule screening piece, it almost is going to increase the number of viable programs. Is that...

That's our opinion. Yes. Yes. And that's what we're hearing from our clients as well.

Yes. But the one thing to keep in mind is that still most of the development that's going on or -- our research work is going on is traditional still. I mean, it will accelerate over time, but it's not like you're throwing a switch here. The second thing to remember is that drug discovery is still, to an extent, serendipitous. For instance, if you look at Viagra and Minoxidil, those were originally -- those drugs were originally indicated for conditions other than ED and hair loss. And it just so happened that in a serendipitous manner, the industry discovered that they had other uses and that were even higher valued. And that happens all the time. And that's going to continue to happen even in the face of more rigorous use of AI for discovery.

Yes. Okay. Maybe we can flip a little bit to kind of the core business. On the R&DS side, can we talk through, I guess, 4Q bookings, cancellations? It sounds like cancellations were a little bit at the higher end. It was a little unclear. Was it 1 or 2 big chunk ones? Was it not? Maybe you can talk on the cancellation side, and then we can dive into the...

You're going to bake me into a book-to-bill ramp here in a second here. But no, listen, I think we've talked previously that 2024, we really saw elevated cancellations about -- in a normal year, the CRO side, we see about $2 billion worth of cancellations. '24 was about $3 billion, and that was really driven by the mass pipeline reviews that were really triggered by the IRA. 2025, we absolutely saw that trend stop. Our cancellations for full year in 2025, even with them being a bit elevated in the fourth quarter, we're really back to normal reasons like drug futility and things like that. And I just want to be crystal clear given the AI hysteria, nothing that we're seeing relative to cancellations, RFP flow or anything is being impacted by anything related sort of AI, right? Like what you're seeing from a cancellation standpoint is just study-specific idiosyncratics sort of events and things like that. So we feel pretty good about Q4 was -- Q4 has some seasonality to it, like Q2 and Q4 are generally the bigger kind of like gross bookings, cancellations, there's no seasonality to them. Whenever they get reconciled and come out, that's when they come out. So P, I just want -- I say that to caution people not to try to extrapolate some linear sort of trend there because kind of Q1 and Q3 are generally lower quarters coming off of a year-end and obviously, the summer sort of holiday period just for some additional context for some of you who overly obsess on this quarterly sort of book-to-bill metric. But no, it was -- there was nothing out of the ordinary to call out there from a cancellation trend standpoint. Book-to-bill is actually being highlighted, I think, because one of our competitors is undergoing some accounting scrutinies and it's calling into question book-to-bills. And again, I'll maybe do a micro ranch here. But listen, it is without a shadow a doubt, particularly now that there are only 4 people who talk to book-to-bill, it is the most misunderstood and missus sort of metric in the industry. It's a long-cycle business. I hope you guys know this. I'm sorry for repeating it. But the way our stock moves with whatever quarterly book-to-bill we put out there, or whatever somebody else puts out there that has nothing to do with us, it feels like our stock always moves down no matter what we do. I'm sure that's not the reality. But we just want to caution people that when everybody -- it's not a GAAP metric and everybody's got a different sort of booking policy. One of the first things Ari did when -- after the merger was -- and Ari comes from manufacturing environments and defense, right? He said, well, what's the industry's standard? Like everybody does it on this as awarded on verbal and stuff like that. And the first thing Ari did was like no, we're going as contracted. So we firmly believe when we stand behind that as contracted basis to put things in your backlog is absolutely the right measure, right? You look at the GAAP measure that everybody is out there, this remaining performance obligation, which is a little bit different than backlog, but it's close and you look at what our CRO backlog is versus what our enterprise RPO number that we disclose every quarter is. And those numbers are really close, right? There's like $3 billion.

And to stress, the RPO is a GAAP number.

Correct. Correct. And you look at some of the other competitors, and they have -- their backlog is far in excess -- I mean, you could drive a truck through -- and it really caused us to ask ourselves sort of a lot of questions, that's why that is, right? So again, at the end of the day, we stand behind sort of our processes. We feel that we have some of the most robust sort of non-GAAP as well as sort of kind of gap disclosures in the industry.

Yes. Understood. I appreciate that the micro ramp. Maybe on just the overall backdrop again, gross bookings seem like they're ticking up. What are you guys seeing from, I guess, both pharma, obviously, biotech has had this nice funding recovery. I know you guys have plenty of exposure there. What are you hearing from customers? And what are you seeing just on that gross bookings backdrop?

Well, generally, if you look at the leading indicators, we indicated -- we gave you in Q4, it all looks good. The pipeline is good. The RFP flow is good. We saw a nice bounce back in the back half of the year and emerging biopharma funding and our conversations with clients have been good. I always caution that, as Mike did there, don't put too much emphasis on quarterly book-to-bill because things fluctuate around. I would look if you're going to choose a statistic to track, track the 12-month rolling anyway. But overall, the environment looks very healthy. Anything you want to add to that, Mike?

No, I would agree.

Okay. And again, it seems like, obviously, we had the pharma, there was multiple headwinds for a long time. It seems like things did loosen up midyear, again, biotech funding picking up. Are you seeing customer conversations kind of steadily improve on both fronts? Is biotech notable with the funding fees?

No, I think they're both good. They obviously are dealing with -- they have their own sort of kind of flavors to them. But there is this general sense of comfort and people more willing to make the decisions, right, now with whether it's sort of policy sort of uncertainty sort of being clarified or funding on biotech. And that's why we've said on both the clinical and the commercial side, we've steadily seen sort of client decision times trend in the right direction.

Okay. And then maybe talk some guidance metrics, which I guess you now own, Mike, so I'll point them towards you. I guess how do you think about maybe down the P&L, the gross margin piece, we get a lot of questions about some of the pricing stuff that's happened last year. Obviously, some of the peers with pass-throughs and pricing have seen noise there. Maybe we start with gross margins, what you guys are seeing and maybe talk a little bit about pricing.

Yes. Let me kind of frame this out a little bit because to be fair, I think given the dynamics and what's going on, we probably should talk a little bit deeper than we have historically about it. If you look back at IQVIA since the merger, I believe every single year, we've expanded margins, EBITDA margins, with the exception of 2020 when the pandemic hit and 2025, right, where we declined 70 basis points on a reported basis. And I'll give you -- I'll use '25 as kind of a construct, right, an illustrative construct. So there's a couple of things going on here. One is you have some very significant nonoperational dynamics going through our EBITDA margins, right? And that's FX and pass-through. When the dollar changes, particularly when, let's say, when it weekends, our revenue goes up, but our -- the denomination of our cost base, IQVIA is pretty sort of muted from any sort of currency changes. So you don't see a lot of change on profit. So FX kind of can throw swings and roundabouts into your margin from that. And that's obviously a gross margin-related item. And then you have pass-throughs, right? I would love if we never had to report on pass-throughs, let alone try to forecast multiple billion dollars of pass-throughs and when they're going to hit and all that stuff. But again, and we saw pass-through tailwinds. So those 2 things just mathematically compressed our margins. If you take those 2 nonoperational dynamics themselves in 2025, I'm speaking in rough generalities there, those 2 things were about a 70 basis point headwind to our reported margins. Then you get into things like mix and a little bit of pricing kind of in the mix. And when I say mix, meaning where we saw things like services were like FSP with lower margins growing faster than the core CSMS growing faster than some of the commercial stuff, things like that and maybe a little bit of pricing kind of thrown in the mix here. That in '25 was about another point of margin sort of headwind. So absent us doing anything else, our margins would have declined on a reported basis, 170 basis points. But we have our productivity programs, right, that we've been -- that we do sort of all along. And to be honest with you, agentic AI is just another tool in that toolkit there on the horizon. But that added about 1 point of tailwind, right? So if you look at our 70 basis points and how you should think about IQVIA's EBITDA margins going forward, really, our productivity programs kind of offset mix plus or minus. And that's what gives us confidence when we're constantly asked about that long-term outlook we disclosed in our Investor Day that, look, in any given year, we expect EBITDA margins to be between 0 to 30 basis points of margin expansion. That's kind of really what we're talking about. Could FX and pass-throughs kind of distort that? Yes. But kind of operationally, that's really what's going on in our business. So we're very disciplined there. And we're really trying to get people to focus more on the EBITDA dollars than the margins just given some of those distortions.

Yes. And you mentioned the mix pricing piece. I know that was probably a bigger concern and discussion maybe a year ago at this time. It seems like things have eased up. What's the pricing environment?

Pricing environment is stable. I mean, really. And one of the reasons why when people were pushing like aren't you worried about sort of the pricing. And we're not being dismissive of it, right, because pharma is very sophisticated. But here's the reality. If we make, a, we're very disciplined when we make sort of very smart strategic disciplined decisions in our pricing. But when we do make a bet on something and lean into a particular sort of price, it goes into our backlog. It goes into a $32-plus billion backlog. And that study is going to burn over 3 to 5 years, right? So we have our ongoing productivity initiatives, and we consistently monitor, like on a full-service trial, we monitor the margin that, that study was sold at versus how it's being delivered. That's actually one of the beauties of the EAC methodology at any given point in time, you kind of -- that's what you stare at. And we consistently deliver on a portfolio basis, 1 to 2, 3 points higher than what we sold at, right? So we're confident in our ability to continue to drive productivity.

Okay. And then maybe the balance sheet leverage, I think, 3.6 at this point. Can you just talk about where we're heading on that front? Again, you guys have done a good amount, you mentioned the Charles River deal. You've done some deals, share repos. What are the priorities? And how should we think about the leverage piece?

I think we continue -- nothing's changed on that front is the short answer. I mean, leverage within the 3 to 4 kind of turn range is something that we're more than comfortable with. We always try to do acquisitions because they are obviously platforms for growth. If you asked me that same question 3 weeks ago before the AI drama sort of ensued, I'd say we'd probably be more opportunistic in M&A. But certainly, our share is a screaming buys -- our stock price is screaming buys sort of right now.

Yes. We are opportunistic, which is to say what -- how we divide up our capital deployment between share repurchase and acquisitions depends on the strength of our pipeline and pricing, what people are asking for and then also our stock. And obviously, given where our stock is today, it's -- you can imagine where our head is about the balance between the 2.

Yes. Yes, that's fair. And then maybe in the last few minutes, a unique opportunity here with the outgoing and incoming CFOs. Mike, as you take the seat here, is there any different approaches you have versus what Ron did, whether it be deployment, cost savings, et cetera?

Yes. What a devilish question has asked with Ron sitting right next to me. So Ron, why don't you put on some ear muffs here for a second here. Listen, the short answer is no. I mean the word that comes to mind is continuity, right? I've worked with Ron and Ari for 9 years. We're going to continue to be sort of very disciplined sort of financially. We're going to continue to have very high standards across the organization in all areas, including an ethics compliance and things like that. And look, I'm excited to where this agentic kind of runway we have in front of us and to be able to support taking the business forward. I don't know if you want to...

Well, look, I would -- I'd be disappointed if Mike didn't do stuff differently than me. That's one of the advantages of making a change. When you've had a CFO or an executive in a seat for a long period of time and you bring somebody new in, you always get new ideas, and that's a good thing. So I'm sure Mike will be making some changes. But as you said, we've worked together very closely. So it's not like bringing somebody in off the street. We've kind of work hand in glove and share a mutual philosophy.

It's a great opportunity now. I mean we are -- I was sitting there lamenting kind of the weekly report that the IR team sends us, and we're currently trading at a PE of 13, which I'm not going to speak for other companies, right? But it surely doesn't make sense for IQVIA. Our historical PEs are high teens sort of low 20s, right? So to people who own our stock who we talk to and are frustrated, well, look, this is not -- this does not make sense where we're being sort of valued right now. Our fundamentals are incredibly strong. I mean, look at the guidance that we put out there relative to what's -- especially relative to what other people are starting to put out there. And the entire IQVIA team is laser-focused in just keeping their head down and executing on that. And kind of to those people who don't own our stock right now, I would say, this makes no sense right now, right? The fundamentals are strong, and we're laser-focused on continuing to kind of keep delivering sort of value here. So we're pretty optimistic.

Yes. And look, we see AI as an opportunity, not a threat. It's clearly being priced into our stock right now is a threat. We feel that's inappropriate and see it as an opportunity, and we'll just have to prove that out over time because it's very hard to disprove a negative, a generalized negative like we're dealing with right now on AI.

Thanks for having us.

Yes. That's a great spot. Thank you so much. Ron, enjoy your retirement.

Thank you.

[indiscernible] and the trip.

Thank you, guys.

Thank you.