Our team has been preparing for this. And one of the benefits of the long lead time that we have in that business is we can be really thoughtful about making sure we plan the work and engage the work and as well as looking for ways to do it more efficiently. The team has done a really good job of that. So, the job shrinkage and the reduction of cost has really not come from that segment of the business. And I believe we’re prepared to deliver the growth that we’re talking about without material changes to the cost structure on the upward side. On your second question around the discretionary CapEx, it’s hard to give you a really long-term forecast about that because we haven’t paired that with what we think the demand is going to be in the amount of activity.
At a given level of activity that’s similar to what we’re doing in 2024, I would say, yes, we would expect the CapEx to be in and around that level if that’s the level of activity that we’re operating with. We’re continuing to see the business move and navigate towards a greater percentage of colocation nodes. Those returns, as I mentioned to an earlier question, have come in at levels that we would expect — we expected them to come in at. So, we’re seeing the multi-tenant model, multi-tenant systems deliver returns that were in line with expectations. And then, as we go out a long way, our view is generally that the carriers are going to need more small cells than what they’re currently taking today as they densify the 5G network. And we believe that densification will continue as consumers use the network to an even greater degree.
So the total addressable market and the need for small cells, we believe will have upward trends on it. And as those upward trends come, I think it creates the opportunity for us to put investment opportunities back through that rigorous process that I talked about in my comments around do these investments in those — in particular markets that may have opportunity in them, do they make sense for us relative to other alternatives? And we’ll just have to see how that unfolds to see whether it makes sense for us to pursue those or not.
David Barden: Got it, thanks so much.
Jay Brown: You bet.
Operator: The next question is from Greg Williams with Cowen. Please go ahead.
Gregory Williams: Great, thanks for taking my questions. Just echoing the comments for Dan, I wish you all the best, and thank you for the support. Just the first question is on the higher cost for ’24. It looks like it’s up $30 million to $60 million even after the $35 million cost savings. I’m just wondering if you can break out the piece parts there if it’s ground lease escalations, et cetera. And then the second question is just on the comments around the small cell returns. You’re saying it’s just as good as towers on a multi-tenant system basis. So, is there any update to kind of lease-up rates in small cells as we think of that vis-a-vis towers? Thanks.
Dan Schlanger: Yes. Thanks, Greg. Appreciate the comments. And then, I’ll start with the first question on higher cost in 2024. You hit on a bunch of them. So, the major cost increases that we experienced, we do have ground leases under our towers as a single large line item that we have in our P&L on the expense side. And those ground leases increase at about 3% per year in cost. That has to be baked in. Secondly, we, like every other company, is faced with — are faced with increasing cost for labor for people who work here, people we’re hiring because the cost of people is going up with inflation. And then, lastly, we had some onetime savings in the back half of 2023 that won’t occur going into 2024, which show a little bit more of an increase also. When you add those things up, you get to the type of cost increases you were just referencing.
Jay Brown: Greg, on your second question around the returns, small cells have been historically continue to be, in our view, would be — will continue to be initially and with the first colocated tenant and the second colocated tenant on a particular system actually better than what we’ve seen historically from towers. When we put capital into the ground for small cells, we are at about double the initial yield on invested capital to what we are with towers. Whether those towers were acquired historically or built, our initial returns are more than double what a tower is. When we get to the second tenant on a small cell system, we’re in the low double-digit range. Generally, with towers in order to get to those kind of yields, as you can see in the supplement, we’re well over two tenants in order to get to low double-digit yields on invested capital.
And then, when we get to a third tenant on a system, we’re high teens, low 20% yields. You could see some of that in the disclosure that we gave last quarter around some markets that we’ve been in for a long period of time and have a significant number of multi-tenant systems. In some places where we’ve got the three tenants, we get the very attractive returns that would exceed those of even towers historically. So we’re at the early stages of colocation. So, it’s — we’re not multi-tenant across the entire system yet, but we do believe over time, like towers, over 25 years of adding tenants, we’ll continue to see growth in those returns and yields. And I believe over time, this business on the small cell side will continue to trend towards what we’ve seen in towers and will create a significant amount of shareholder value over time as we’ve been able to build assets in the best markets in the United States, those with dense populations and a lot of data demand.
And believe as the carriers densify the network, the assets that we have are going to result in a lot of colocation over many years. And that will consistently drive increases in yields and growth in value creation for equity holders.
Gregory Williams: Great, thank you.
Operator: Next question is from Matt Niknam of Deutsche Bank. Please go ahead.
Matthew Niknam: Hey guys, thanks for taking the questions. Just two quick ones, first, on AFFO per share growth, maybe if you can help us think through the moving parts driving the expectation for second-half AFFO per share in ’24 to be better than the first-half. Is it improved leasing? Is there anything on the cost side in terms of ramp-up of savings to be cognizant of? And then secondly, on services, if you could just help us think through the progression from the, call it, 25-ish percent range in the third quarter to around the 50% exit rate by the end of next year. Is that linear? Is that more of a stair step higher? Just how to think about the path there? Thanks.
