All right. Why don't we get started? For those who don't know me, my name is Chuck Napkin. I cover the payments and transaction services space here at Stephens. Joining us today from NCR Atleos is CEO Tim Oliver and CFO Andy Wamser. Gentlemen, really appreciate you being here this week.
Yeah, my pleasure. Good to be back.
Yeah. Thank you. It's been an interesting few years, as we were talking about earlier. It's been two years since Atleos split from NCR. You've completed the heavy lift: 700 customer connections migrated, 200 TSA exits, and a clean balance sheet established. Now that the separation is behind you, what does Atleos 2.0 look like, and what are you focused on strategically?
Yeah. Do I need to pull this mic around, or is it?
Yeah.
Pick it up? Fine. Personally, it was a lot harder to get done than we thought. A separation of a 140-year-old company was difficult, and I think it was a distraction, a necessary distraction, but it consumed a lot of resources to get that work done. We've got a lot of resources now we can redeploy to go do things that are much more thoughtful and strategic. We've been exceptionally tactical for the last several years. We sat down with our board in the last board session, and they said, "Tactical is fine. You've been successful for seven quarters doing what you said you're going to do. You'll be there for an eighth, we suspect. What are you going to do for us now? What's the strategy?" You've earned the luxury of having a strategic outlook for this company that you didn't have before.
We're spending a lot of time thinking through not just what we're going to do. The original model was grow 3% or 4% or 5% a year, grow profitability twice as fast, convert cash flow better every year such that you get closer to 50+% conversion rate by the time you get a couple or three years out, get the balance sheet in order, buy back your shares, and all of a sudden you create a tremendous amount of value over the next three years. I think that's still an underlying case for us, right? That can still take place. I don't think we need too many people to help us get that done. I think we can do that ourselves. We're on pace to do that. The board said, "That's mildly entertaining.
What else are you going to do?" There are three things that came out of that session, and we're going to have a big strategy session in February with our board. The first was we have assets that are remarkable and differentiate us in the space that we must protect. Our installed base of devices is some 500,000 strong around the world. The real estate that those sit on and the customers behind them are incredibly powerful. No one has a bigger one. We've bypassed our nearest competitor in the last year. We now have the largest installed fleet in the world, and it's growing more quickly. We're actually putting more devices into place. I suspect that lead will increase. Protect that installed base and the real estate that it occupies because that's what we'll monetize over the next several years as we grow our services business.
Protect that services footprint. That services footprint of 8,000 CEs all over the world that deliver incredible service to our customers, best-in-class service to our customers. Make sure you protect that and do not let anybody get inside of that business. Lastly, we have a fleet out there of our own owned and operated devices that gives us a knowledge set into how banks have to run their machines. We probably run them better than banks. Protect all those three things. Discover what you are going to do with that infrastructure. We think there are a couple of avenues of growth for us. First, the kiosk has to become more capable. Physical to digital or digital to physical transactions. We will do a lot more of them. We can do your driver's license renewal. We can do your passport renewal.
We can do a lot of things for governments and for others that require physical presence and verification of an individual at a device and safekeeping of something in that device. Then use that incredible service organization to do more, either for banks and retailers who ask us all the time to do that, and it is kind of the underlying growth strategy and the base case, let's call it, or do it in other places. Do it in other devices. Do it for folks who need more service support on electrical mechanical devices with the need for very high uptimes. That is the beginnings of a, let's call it, a three- and five-year strategy that goes beyond that kind of base case that still is pretty compelling.
Great. We'll drill into the products over time, but I did want to go back to one of the points you'd made about cleaning up the balance sheet and essentially doing what you said you were going to do. From a finance perspective, you've achieved your goal of sub-three times leverage ahead of schedule and reinstated the shareholder returns. What's enabled you to do that, and how should we think about how are you thinking about capital flexibility and capital allocation going forward?
The model as we rolled out said that we would generate significant cash flow over the first couple of years. At a conversion rate, it was a little too low because of the high debt and the high tax rate spinning out. We've worked on both of those things. We're a little ahead of schedule on cash generation. At the time, we said, "Let's get down to three times leverage," because that felt like a number that was less offensive to most people. As we sit here today at three times, the world would like us to get a little bit lower, and we're perfectly willing to do that. We'll have an opportunity in October of next year to refinance a tremendous amount of our debt.
I'm very hopeful that the rating agencies will look upon that cash generation and be constructive with us, and that perhaps come October we can do something pretty different with our debt and continue to work that down. I think there's plenty of room now to go buy our shares back, right? That's something we talked about when we were on the road proposing spinning this company out. At the time, some favored a dividend. It had become pretty clear that small-cap companies shouldn't pay dividends unless they're really high dividends, and we don't intend to do that. A share repurchase, particularly at the levels we've continued to trade at, makes very good sense. We'll accelerate that forward.
Those cash conversion rates, as we lap some of the tax rate and as we lap interest rates and as our profitability becomes a bigger percentage of the total, we should be able to convert very, very nicely. I suspect our conversion will go up five points or so next year from 30% to 35%. I think it can go as high as 40% the following year and should continue to improve from there. I feel good about the cash generation over time. That is what this business is all about, really, right? It is a capital-efficient business. We do not have to invest a tremendous amount of capital back into it. We ought to be able to run M&A—I am sorry, CapEx—relatively close to depreciation and use M&A to add more assets, add fleets, add capabilities to the device.
Nothing big and transformational, just simple bolt-on stuff that's accretive and instantaneously pays for itself in a year and a half to two years.
Right. I wanted to double-click on the ATM as a Service strategy. I know you touched on protecting your base, emphasizing your service capabilities. Could you maybe talk about that offering, the value proposition it brings to your customers, and how it enhances revenue predictability and provides an uplift to ARPU?
Yeah. It is an uplift to ARPU if you think about our traditional business being about 35% of the total lifecycle spend in the ATM. Put in there, the cash and transit guys do about 30 cents on the dollar. There is 65 cents of what—and then the other 35% banks are doing for themselves, or they are cobbling it together and going and chasing down. For us, it is more than doubling of the ARPU available to us on the device when we take it as a service. We are now at, probably 35,000 machines that we are currently—38,000 machines we are currently doing as a service on soup to nuts. We have got another 90,000 machines behind that that we are doing enhanced servicing on. We call it as a service only when they help toss us the keys and we get everything.
Many of our competitors who describe as a service would say any service other than their traditional service makes it as a service. I'd say, look, we're at 125,000 machines currently. We run in some way with an extended service set for our customers. It's a very compelling sale for most customers. If you have fewer than 500 machines, I know I can convince you it's the right thing to go. Fewer than 1,000, probably almost as easy to convince you. I'm going to save you a lot of money. I'm going to save you 20% of it, at least 20% of what you're spending yourself to support the machine. I'm going to give your client a better experience in the device. I'm going to make sure that that device is up and running more often. I can deliver uptimes better than your internal IT team can, for sure.
I'm going to guarantee that you have less cash in that device than you've ever had before because my algorithms are pretty sophisticated. Since it's a cost to me, I can help you get cash out of the device. Believe it or not, it's still early innings from an adoption perspective, but the product itself is maturing and is becoming more universally accepted.
Got it. To further that point, I wanted to talk about the geographic footprint of the company and that business in particular. Obviously, I think overall we're still in the early stages of adoption, but some geographies are ahead of others. You recently entered LATAM, Middle East, and Spain, and announced some new outsourcing deals there. Could you talk about what regions you're seeing more success in than others and maybe drill into the stages of adoption across your footprint?
Yeah. We've had much more success with the ATM as a Service strategy in places where we have our own network, an owned network. That's 14 countries. We think there's about 32 countries that we have sufficient scale to be effective as a service to make sure we know we can drive the cost savings that our customers expect and that we can guarantee the uptime. Sufficient density of engineers that we can deliver the uptime. The margin conversion in every country appears to be a multiple of what it was going in. In every instance, ATM as a Service has been accretive to local margin rate. Admittedly, North America, Western Europe, we have our highest margin rates in all that we do, and the ATM as a Service margins there have been the best.
They've been significantly accretive relative to the average margin rate in those regions. The same is true in Spain or in India or in Brazil. When you win those deals, it's very accretive to overall margin rate in the region, but admittedly, rates in Brazil are a little bit lower than North America. The mix of devices in the ATM as a Service strategy matters to the rate that we describe for the business itself.
How differentiated is your offering relative to the competition? I know there's various areas of the value chain. You play in most, if not—you play in most, but at the same time, some of your competitors are in certain areas but not in other areas. Could you maybe touch on the competitive set and where you're differentiated from a product standpoint?
Yeah. We started talking about ATM as a Service in December of 2021, grew it from nothing to almost $300 million today. When we pitch ATM as a Service—and by the way, since then, a lot of other folks said, "You know what? They're right. That is where the growth is going to be." That wave of outsourcing associated with the cash ecosystem is going to be very lucrative for somebody going forward. Not so different than the core system outsourcing that some of our FIS and Fiserv have been soft for a period of time. I think we're about to catch that wave. A lot of folks who offer the product of ATM as a Service, and then they'll cobble together a solution of which we're a part. When Fiserv wins at—we're their exclusive provider. When some other companies win—we're a provider.
When Brink's wins an ATM as a Service deal, we hate that they win it, not us, but we win too because they take us with them. We are bumping into atypical customers, competitors when we go into that space, but no one else can put it all together themselves. The only thing we do not do—and we do a little bit of cash and transit, but for the most part, we farm out our cash and transit business. We do everything else around the device. Others coming into the space do very little around the device other than their particular expertise. Cash and transit, in the case of Brink's, the core system tie-in for FIS, and they come to us and ask them to help to solve that.
If you want a single throat to choke and you want somebody who can integrate the solution completely without including others, we're the place to go. If you want to incrementalize your way there, maybe some of those folks can help you get there. We have every right to win when we go into these competitions. It's frustrating if we don't, but yeah, no one else can put it all together themselves. The learning and the efficiency that comes with that, I don't know how you—I don't know how you commit to an as-a-service transaction when you don't have a service organization. Seems risky.
You've now posted five straight quarters of double-digit ATM as a Service growth. As this business scales, how sustainable is that 30-40% growth trajectory? I guess we talked about what's driving that momentum, but how reliant or un-reliant is that business on overall ATM growth around the world?
Don't need growth. We've got plenty of customers to convert. We've got that base of 500,000. We've converted 35,000 of them, right, or 38,000 of them. There's plenty of runway for this strategy. When we rolled it out in 2021, we thought it'd go a lot faster. We had growth rates in this thing that were 60-70%, right? We were ahead of where we are today in the original model. The 40% rate relative to that model is mildly disappointing, right? But it's a rate we can sustain. It's a rate that we've been able to execute against. We've been able to onboard customers quickly enough. The backlog is—the selling effort has created a backlog that supports that 40% growth. I feel very good about the 40% growth rate.
I wish it was a little bit higher, to be honest, and it matters a little bit where it comes from. If I get that all in North America, it's nirvana. I think we're going to have a really good quarter this quarter in North America. We'll love that come Q1 and Q2 of next year. But 40% growth from here should be something we would anticipate next year and the year beyond.
I think the only thing I also had to add to that is when you look at—as we've gone throughout this year, you've seen the ATM as a Service accelerate in terms of that growth rate, but what you've seen most importantly is the flow-through associated with it. For example, in Q3, we were just below that 40% at 37% growth, but you saw the flow-through in terms of being at 65%. We're seeing really good flow-through associated with it. It's not just the top line. It's incrementally getting that software and service revenue as well.
Got it. Okay. I did want to touch on hardware and tariffs as well. It's a relevant piece of discussion. You've said you'll put 20 more devices into service this year. Where are we in the hardware cycle? I know maybe five, six years ago, there was a refresh cycle. Where are we today, and how are you viewing the longevity of the trend?
In 2019, you saw a 35% increase in the number of devices manufactured by the big three, right? This year, our largest competitors are relatively flat. We're up about 15%-20% in the number of devices we're putting into place. I think part of that is our hardware set is very competitive now. It may not have been as competitive as it needed to be, quite frankly, over the previous two years. I think part of the extraordinary growth we're seeing this year in devices putting into practice and into service is because we left a little on the table in the previous years. We're getting back our rightful share, and our customers are treating us well as we've started to distinguish ourselves on the recycler in particular, right? I also think we're taking a little bit of share.
I mean, we have to be if we're growing this quickly. As I said before, we established ourselves as the largest fleet in the world. At the end of last year, we'll extend that lead this year. It feels like that 35% pop that we saw in 2019, we've seen six or eight points of that this year, but it's not been 35%. I think it's going to be the next several years. We can see our order book out a couple of three quarters. It looks really good going into the first half of next year. I can't see beyond that, but I have every expectation that after really what is going to be a record year for us from a hardware perspective, save 2019, but a more normalized year, I think we're going to beat it next year.
I think it's going to be another good hardware year in 2026. Early innings, I think this is probably a three or four-year kind of mini cycle, if you will, and we'll participate handsomely across that whole time.
Maybe just one other thing I'd add to that is hardware is still only 20% of the total portfolio, but I think the important point about getting that install base and what we've been able to do this year and what we frankly are able to project for Q4, we talked about being mid-20s, and we can see out that Q1 is going to be a really good quarter as well. The important point then is that we also get the service and software business associated with that, and that comes with then multi-year contracts. We really do view, I would say, the service and software in aggregate because as we get more, I'll just say incremental services, whether it be software services or traditional services, two men in a truck, it gives us more predictable revenue streams for the next several years.
Got it. Tim, I know you touched on recyclers, and I think the point that's worth emphasizing here is that relative to the past cycle, you have a stronger offering in that area, and you're better positioned to participate. Could you talk a little about that product?
Yeah. The recycler is a singular throat for both deposits and dispersing. It's a machine that takes cash back in and reutilizes it such that truck rolls to the device are minimized. It's a great technology. Large banks love it, and they're embracing it now, and most of them are turning that recycling capability on and finding out where it works most economically for them. What they've decided is the world's going to recycling, and the lion's share of the larger banks, when they order their devices, they're ordering recyclers. Smaller organizations are treading a little bit lightly. They will follow the big banks once the big banks show that it's successful. About 20% of our devices this year will be recyclers. I think that's a lot more than it was in the previous year.
Our recycler product was not good enough as we sat spinning out two years ago, and we were behind from an uptime perspective, performance criteria. We were behind in terms of number of devices we could produce, and our costs, frankly, were too high. We've worked on all of those over the last two years, and in particular, the last 18 months, and now we're leading in performance. We're leading in number of devices delivered, and we're able to make enough devices to keep our customers happy, and we're working hard on the cost criteria. With the absence of tariffs, as you brought up, those costs are moving in the right direction. We've just got one cost called tariffs that's moving a little bit the wrong way.
Okay. Let's talk about tariffs. That's my next question. What are your current expectations?
See, I already did that. I said it right off.
Yeah. Good segue. What are your current expectations for tariff impact as we head into next year? Could that be—could we see a potential—I know it's tough to say what's going to happen from here, but is your bias to the upside or downside, or how should we think about it?
No, this is going to get taken care of in relatively short order. It is painful right now. It is a 50% tariff rate coming out of India. We do all of our manufacturing, most of our manufacturing now in India. The only wholly owned plant by us is in Chennai. It is an excellent plant with very low costs, great employee base, and really good fixed cost leverage because we do so many devices through a singular plant. It makes no sense to move that facility or change that up in the short run until we know what is going to happen. Most all of us who make ATMs buy our components from all over the world, and all of us are getting tariffed on the components that make up 70% or 80% of the cost of the device.
Where you do the labor, the assembly does not really matter that much from a tariff perspective. At 50%, it is a big deal. $200 million a year coming to the U.S. at 50% is a big tariff. We will absorb, I think, this year between $25-$30 million of tariff hit, and we are going to cover that and still hit our guided ranges. It has been painful. We have taken a lot of cost out to get there, but we will scramble back to a decent outcome. I think that tariff rate ultimately is 15% or 16%. Everything that we hear from our experts in India, where we have gone and met with government officials there who have offered to help us offset some of these costs.
I've been in Washington twice working the halls on both sides of the Capitol building trying to figure out how we can get this to be helpful. Of course, most of them say there's only one guy who makes these calls, so feel free to talk to us. It does feel like 15-16% of where it's settled out. Right now, it is still at 50%, and most of my shipments in Q4 are going to be at 50% because they're on the water already. I'll go into my planning session next year, and Andy will presume something like 25% if we don't have a decision. When the court case comes down, and I think it could be any day when the Supreme Court rules on the legality of the tariffs, Modi will then move.
He is not going to move in anticipation of that because they may throw them out altogether, and he does not need to negotiate. I think we are weeks away from a decent outcome there. When we give guidance for 2026, we will be very specific as to what we have assumed for tariffs so that you all can guess with us.
Okay. Makes a lot of sense. Before we move to networks, I want to stop and see if there's any questions in the audience.
Just your comment about being a little bit short of where the original expectations were on ATM as a Service. Where's been the friction point, or where has it been different than what reality became?
Demand is fine. The selling cycle is a little longer than we anticipated. The decision point isn't as front-end loaded with new machines. We presume people will make this decision when they've made the next order of new machines, and they don't see that as a point they need to make the decision. They're perfectly willing to make the decision of as a service after they've replaced the device a year or two later. It is a different selling process inside the organization. It has been a little extended. The biggest delta for us has been the ability to implement devices quickly enough. In most instances, we sell a machine to Bank A. We send in a crew that does a kitchen remodel, right? They rip the old machine out of the wall. They put the new one in. They put the spackle around it.
They put the machine in, and they spin the machine up and go. When it's as a service, you have to integrate into their systems, into their IT systems. You have to touch the IT department inside of these banks in order to make this switch, and it's taking longer for the bank to prioritize the switch and to get us up and running. What we thought would be three and four months is turning out to be eight or nine months to get a customer onboarded once we've sold them the products. The backlog has been decent. We're pretty good, actually. Demand is reasonably strong, probably as good as we would have thought. Our sales guys are getting paid decently well to sell this product because they're all selling it forward, right? We've just not implemented it as quickly as I would have thought.
I think the backlog will stay strong. We have the capacity now to about 12,000 machines a year. We're doing 1,000 a month right now implementations. That probably needs to get higher to say 15,000 a year, so roughly 1,500 a month. We'll keep working that up. It is implementation in concert with our customer that's been the most difficult timeline.
The only thing I'd add is when Tim talks about the implementation in terms of being at 1,000 a month or accelerating, I'd say when we have all geographies going at the same time, when we think about North America, CLA, EMEA, and APAC all going the same, then we have the capacity to even do more than that in the future.
Yeah. It can be bumpy, right, in some regions, some quarters. Yeah. Thanks for that.
Switch to networks. Traditional withdrawals make up the bulk of revenue, but deposit transactions are up 90% year over year and are your highest margin use case. How significant could deposits and digital wallet activity become over the next few years?
Yeah. We'll have a digital wallet, I suspect, by the end of this year and end of the upcoming year because the LibertyX product is, well, Bitcoin is not our future. Stablecoin is. Having a wallet that allows people to move between currencies is going to be crucial. Do not be surprised if you hear us talking about a digital wallet. I do not think we'll have our own currency, but we'll reach out for naming of that customer. Yeah. The transaction model has been typically a decline in surcharge transactions, more than offset by an increase in surcharge-free transactions with new clients on the network. Add to that new transaction types like deposits and Ready Code and others. That was working really well until we got thumped on the prepaid card, the payroll prepaid card environment, down 15%-20% this year.
A million and a half customers just disappeared. It used to be at our devices, habitually, no longer there. Probably went home. I'm not getting paid differently. We're not really sure. The decline you've seen in the last two quarters is entirely related to that trend. There's a little bit also of DCC in that people aren't traveling to the U.S. like they did before. Both Canada and Mexico, we've not seen as much cross-border vacationing and the like. Those two phenomena taken together cause us to be down modestly. The other underlying trends are true. To your point, deposits are growing at 90-100% year over year. The Ready Code business had an air pocket. We had a contractual issue with an acquisition that took place of a customer, and we needed to go offset that. We fixed that now. That's growing.
That's back and growing again. Surcharge-free transactions are still growing really nicely. We've stabilized. Those two negative patterns have stabilized at not great levels, but they're not getting worse. They won't be a headwind next year. They'll be maybe neutral. I think that business will get back to growth in Q4, flat to up slightly in Q4, and then back to growth in Q1. Much like ATM as a Service has been too small to move the needle for the totality of the self-service banking business, so is the case of deposit transactions and the overall economics of the network business. That's changing. It's a very lucrative transaction. We get paid both a fixed and a variable fee on the deposit business. We're really excited about deposits from small retailers.
It appears the small business folks are looking for an easy way to deposit cash. I think we can be that solution.
On the topic of expansion, you've announced expansions with Circle K, 7-Eleven, and Access Cash in Canada. What's the next wave of footprint expansion? Are you targeting more rural areas or deepening your density with existing retailers? How should we think about that?
We'll keep looking global, right? Italy and Greece have gone really well for us. Those machines are very productive. There's not enough of them, and we've entered those two markets without an acquisition. We entered and started putting machines in ourself, and that's a slow way to get up to speed. I think we're north of 300 machines in each of those countries, good machines, learning our way through. If there were fleets available there, we'd buy them. Underutilized fleets that could be put to good use. The KCs is a great example of places that are more rural where we have really good coverage in the Allpoint Network in the United States that's about five miles within 85% of the U.S. population. There are other locales around that that we could service better. KCs is more of a rural customer.
They're great for us, and we're great for them. We can prove out to Circle K that we're driving a lot more foot traffic through their stores because Allpoint's in there. The typical Circle K has 23,000 more visitors a year because my machine's in there. I can prove that out to Circle K that it's going to work. Others like Circle K, we'd love to have them on board. If there's going to be a machine in those truck stops in those locations, it might as well be ours that's networked because it's good for everybody to make that happen. I think you're going to see more countries. You're going to see us go to more countries. As I said, it pulls ATM as a Service business when you're able to do that.
You're going to see probably more of these underutilized fleets in parts of the world where it's a little more difficult to service the devices that we know we can drive foot traffic to.
Got it. I wanted to double-click on a couple of the areas of the business you mentioned, specifically Ready Code and the prepaid card. First, on Ready Code, could you talk about what that is and how that fits in the overall strategy? Secondly, on the prepaid, how should we think about some of the headwinds you're facing in that business entering 2026?
Yeah. Ready Code is, think about a cardless, accountless transaction. It is somebody who just uses our device and the cash in the device to extract what they need when they need it. In other words, that is because they lost their credit card or because their mom wanted to send them money, cash for the bar in whatever town they are in. Or a gig worker wants to go extract their daily earnings to go have some walking around money for the weekend. Whatever it might be, Ready Code can do that. We have started signing up lots of customers and getting them on board who want to push cash one way or the other to folks who may or may not have an account, may not have an account with us, and do not have a card, do not want to have a card.
The obvious one we've talked a lot about is the gig workers who have a payroll system. We contract with a payroll company. The payroll company makes the API available on the phone of the app available on the phone of the payroll recipient, and they go to our device. We push a code to them. They push it back to the device, and we put cash out to them. It is a really good way to extend. You think about all the use cases for that. It goes way beyond gig workers. The Western Unions, anybody who's pushing cash one way or the other, we can do that at devices that already have cash in it and not have to involve anybody other than the person with the code in the phone. I'm sorry. The second question. You had a second question there I didn't answer.
No. I think that was it. You touched on the Ready Code, and you touched on the prepaid. Next question was, how should investors think about the impact of rate cuts as it relates to your vault cash within the business?
Yeah, I would first off and say any sort of rate cuts for us is beneficial, right? You think about how we have the 80,000 machines we rent around the globe, just call it $3.7 billion that goes into our machines. When we think about the impact of rate cuts, and particularly within the U.S. and what the U.S. Fed is doing, we've had two cuts so far. When you think about if we have $2.6 billion within the U.S., you directionally would get about $14 million sort of savings or less expense. If we get a third cut, it'd be close to $20 million. Clearly, with next year, with the network, particularly if we do get that third cut, we should certainly see improved, I would say, less vault cash expenses, which is our biggest expense within the network.
We've tried really hard to get the amount of cash in those devices down. We've done a decent job, as Andy said. It used to be closer to $4 billion around the world in different currencies. It's down to $3.5 billion, maybe even a little bit less, of which $2.5 billion is in the U.S. Rate cuts matter in the U.S. They also matter in other parts of the world. It can be helpful too.
Sure. I want to circle back to free cash flow before we finish up. I know you touched on it a little bit earlier. You're expecting a step up from 30-35% in 2026. What's driving that step up? I know you talked about capital allocation priorities, but if we think about just cash flow efficiency and conversion over the long term, tying it into the AI theme, what are some ways that you're looking to improve the efficiency of the overall business, whether it's dispatch optimization or any ways you could incorporate AI or greater efficiency into your outsourcing operations?
The sources of cash are increasing more quickly than the uses of cash, right? We don't need our interest expense will come down. Our tax expense should be flat to down slightly. We don't need to use a lot of cash to support revenue growth. I think working capital may even be a modest source of cash next year. We don't need to chase new business with overspending depreciation with CapEx. We can be very, it's a very capital-efficient model in the absence of some of the cash uses we brought over from the separation. We'll lap those to a certain extent and leave those behind. Remember, in 2023, we had like $50 million of cash expenses just from getting free.
I think we can lap those, which means you should see modest upward pressure on conversion rates, presuming we keep that capital light, and I think we can. The only use of cash that's a little bit outsized relative to the rest of the business would be investing back into our own fleet, right? We have to refresh those devices as we go, or ATM as a Service that we're perfectly willing to take on when the returns, the kind of returns we've seen. You should expect that cash conversion rate to accrete up over the next several years.
Yeah. Maybe the only thing I would add is next year, when we think about the conversion of at least, say, 5% better next year, we're obviously going to have less interest expense. We expect still strong earnings growth, and there could be a potentially source of working capital. You sort of look at the benefit of 2027 and beyond. That is where we touched on the potential for the refinancing that we have in Q3 of next year. There is an opportunity there where you kind of look at where our notes trade today, where we have our revolver and our term loans, and you can see the potential to at least get 2.5 percentage points lower on our bonds, which is at 1.35, and let's say 25-50 basis points on the revolver and term loans.
That in itself should be another $40 million of incremental sort of interest.
You asked about efficiencies, cost efficiencies. It is on the back end of that question. Yes, there is a lot of them to be had. We have deployed AI in our service organization this year, driven really terrific results, both cost savings. We are down about 700 people across our service organization with all-time high service levels. That is working great. We need to take those same AI tools and take them to our selling organization, and we need to take them to our corporate functions. We will do that next year. Our costs are still a little bit too high. We spun out with a little bit too much cost. As you know, we are a pretty big, complicated company for only $4 billion in revenue, right? We are trying to figure out how we simplify that and automate more stuff. I think it will be a good productivity year in 2026.
Got it. Any questions?
That 35% debt's conversion of PIVOTOP. Can you talk about it?
Yeah.
Yeah. I was going to ask about your cost and margin structure. What are your biggest opportunities and levers that you control as you kind of think forward?
Yeah. We have the highest margins in our space, right? We have to be careful about pushing too hard on what margin rate the customer sees. I think we want to compete aggressively in parts of the world where margins are a little tighter. I think there may be, from a mixed perspective, as I grow in India and grow in Brazil and other places, maybe some modest downward pressure on margin rate in aggregate. I have to offset that and help. I want to improve margin by a point a year. I got to go get costs out all over the place to make that happen. We were very NCR Legacy, we love to pretend we were a very big company. You get lots of layers of management with that.
You get big selling organizations that have pre-sales and then post-sales, and you have the specialists, and then you have we've got to kind of go through. We're hiring another consultant. We had a consultant come in and help us with two of our corporate functions. And we offshored a lot of jobs, and we have to take a lot of folks out of the organization with their help. We'll do the same. We're actually interviewing firms right now to do that on three other of our corporate functions, and in particular, go look at our selling effort and how to be much more effective from a selling perspective. You can make a salesperson far more effective using AI today than he or she has been in the past, much like we just did with our customer engineers.
They're far more effective with the AI tools we've given them than they've ever been before. The number of calls they can do in a day are up. The number of times they get the first fix right is way up. The number of times they arrive with the right part of the truck is way up. We are going to have a huge opportunity, I think, to go after the selling effort and the support effort with AI.
Is it a multi-year kind of journey for you then? It's not done at this point, two years in?
You want to be really careful when you do this stuff that you don't disrupt. What we distinguish ourselves on is the service level that we provide to our customer in the uptime. We roll these tools out slowly to make sure that they work before we go disrupt the entirety of the operation. When we rolled out the most recent AI tool, we went to Canada first because they're nice up there. We got it right in Canada. Then we rolled out in the U.S. region by region. Now we're moving it to Western Europe in the first quarter. We will start yet another tool in Canada in the first quarter of this year. We will continue to use this cycle to roll it out globally.
Yeah.
A little less risky that way. Thank you.
I was looking at your stock. It's 7 PE or so on a board basis. What do you think is the most misunderstood aspect or what could get some of your analysts like Chuck here to really fall in love with it? Or kind of what do you think is misunderstood?
Chuck's already in love. He just doesn't, it's a quiet love. Here's what I said. When we came out, we said we need to generate predictable results quarter after quarter for a period of time and put a track record together. Seven quarters in the track record, I get that, but at least we're starting to do that. We needed to get our debt down under three times to make some big investors who are long-term, long holders, to get them interested in the story and allow them to own the shares. We're there, and we're talking to those big accounts. We need to get share of mind up. I think people have far less excuses to not pay attention when those two things are fixed, when we have a track record that's defensible and we have our debt where it needs to be.
The cash generation that this business can take on in the next couple of years is really powerful. What it can do for our debt and for our outstanding share balance is going to be huge. At some point, people are going to sit up straight and notice that. I think we had a firm, a buy-side firm, publish a report a quarter or so ago where they looked at their assumptions that are a little aggressive, but from their pencil to God's ears, right? The direction in which they were pushing us is exactly the right direction to push. They see very, very strong cash conversion in the out years. I think that they're right. The question becomes, what is that inflection point? I thought we saw one, and we get to $42 or something.
It felt like maybe we were on the path to re-rating a little bit and getting an extra multiple point. All of our whole sector has come back down a little bit over the last several months. I think we will be back there again. I think when we give guidance for next year and we tell people that we are going to generate free cash flow that is high and that we are going to return a lot of that to shareholders, I think having just done it for the two quarters that preceded that guidance, I think that will be helpful.
I think when we start to give people a flavor of 2027 that has $500 million in EBITDA, $1 billion in EBITDA and $500 million of free cash flow with a multiple like you just talked about, and a track record of two years preceding it that suggests that the third year will absolutely come true, I think we have a breakout moment. At some point, people will sit up straight and go, "This is just a cash machine. It's just going to generate a lot of cash," right? It's not going to be a sexy high-flyer growth. We're going to grow 4% or 5% a year. In a space that no one thinks is growing, we're going to grow 4% or 5% a year.
Top line.
Top line. We are going to grow profitability twice that fast because we can expand margin and use tools to get more efficient all the time. As I said earlier, the base case of this thing is halfway decent. If we can add some acquisitions onto that, do $100 million or so, $50-$100 million a year of smallish bolt-on acquisitions that extend the network or bring more functionality to the device, I think it could get pretty exciting.
Yeah. Let's expand on that a little bit. We've got a couple more minutes. In terms of the M&A strategy, any particular products or solution areas you would look at from an inorganic standpoint?
We need to be ready for Stablecoin. We need to be ready for the transactions that take place around Stablecoin. We have Bill Pay capability around Stablecoin. We need to make sure that people can get to the device, take a wallet of different currency types, convert that to cash or not, pay bills with it, transact however they want. Any functionality at the device can help us do that, we are willing to invest in. I will tell you, this stuff is not expensive. It really is not. I mean, the acquisitions we have done, we have put single-digit millions of dollars into a few companies, got nice equity positions in them, and exclusive rights to some of their capability that have really been helpful in some of the product we have rolled out. It is good for those technologies. Remember Clip Money? We did the Clip Money thing.
We got a really nice piece of Clip Money and a really good relationship with them because we could make them real overnight. When you give a product like that access to 80,000 machines, they go, "Oh, that changes the trajectory of the company." I think that's a very obvious place for us to put money. We're going to put some money into the device itself to make the device more capable, to make sure TAP is ubiquitous, to make sure that the sensors in the device have all the biometrics necessary to make sure the life of the device is long. We are going to buy fleets. There are so many fleets. You talked about Canada earlier, right? We stepped into an underutilized fleet there that had been asking us for years to help them out. We said, "We will help you.
I will just own you. Incredible group of employees. Those machines are already generating significantly more profitability than they did before because the network is pushing, our network capability is pushing the Cash Access brand and allowing people to go to those devices. I think you're going to see us in Italy or Greece or other countries find underutilized fleets that we can add on and make accretive like overnight. Those are the most obvious acquisitions for us.
Yeah. Great.
Tim, Andy, really appreciate you joining us again this year. Thanks to everybody in the audience. We will be around all day in case anybody has any questions. Look forward to seeing what is next out of you guys.
See you, buddy. Thank you.
Yeah. Thanks again.
Thank you. Be good.