Crown Castle (CCI) Q3 2026 Earnings Call: Guidance Raised on Cost Savings

Crown Castle (NYSE: CCI) raised its full-year 2026 outlook after reporting continued organic growth, additional cost savings and benefits from a $275 million customer agreement. Management reiterated that 2026 is expected to be the low point for organic growth, although core leasing activity is likely to finish near the bottom of its guided range.

Key Takeaways

• Third-quarter organic growth was $35 million, or 3.5%, excluding Sprint cancellations and DISH terminations. Growth was 3.8% when prior-year DISH revenue was also excluded.
• Crown Castle increased its full-year 2026 outlook by $20 million for site rental revenue, $25 million for adjusted EBITDA and $5 million for AFFO.
• The company paid $275 million to acquire a residual economic interest in a subset of towers, extend a major customer’s lease terms and limit long-term churn.
• Crown Castle expects to purchase more than 400 ground leases in 2026, more than double the 2025 total. Discretionary capital expenditure guidance increased by $40 million to $240 million.
• Management still expects more than 200 basis points of margin expansion by 2030 versus 2026, supported by operating efficiencies, systems modernization and ground lease buyouts.
• More than 100 edge computing sites have been deployed, but management said the opportunity remains in the trial phase until Crown Castle determines whether it can scale profitably.

Key Financial Data

Business and Operating Performance

Crown Castle generated $65 million of annualized run-rate operating cost savings after aligning its cost structure with its tower-focused business model. Since the initial 2026 outlook was issued in February, full-year adjusted EBITDA guidance has increased by $20 million from run-rate cost savings after excluding revenue increases and the reduction in services contribution.

The company is modernizing project management and field services systems, optimizing its facilities footprint and reducing corporate nonlabor costs. It also expects to exit the transition services agreement associated with the sale of its fiber and small cell businesses, which management believes will further simplify corporate functions.

Ground lease buyouts remain a central efficiency initiative. Crown Castle expects to acquire more than 400 leases in 2026. Management said owning more land beneath its towers should improve margins, increase operational control and support faster customer service.

The $275 million customer agreement is expected to increase site rental billings by nearly $4 million in Q4, or $16 million annually. It should also reduce prepaid-rent amortization by $2 million in Q4, or $8 million annually. Management expects approximately $4 million of annual AFFO accretion, excluding prepaid-rent amortization, and $20 million of additional straight-line revenue in Q4.

Crown Castle has deployed more than 100 edge sites, compared with very few at the beginning of 2026. The company has identified about 5,000 existing structures or shelters in varying condition, many of which could accommodate customers quickly. Its current target deployment size is generally 0.2 to 0.5 megawatts.

A review of 25 sites indicated potential access to as much as 100 megawatts of aggregate power within approximately nine to 11 months and with limited Crown Castle capital. Management emphasized that edge computing remains an early-stage opportunity.

Management Guidance

Crown Castle raised its 2026 outlook by $20 million for site rental revenue, $25 million for adjusted EBITDA and $5 million for AFFO. The changes reflect $20 million of additional straight-line revenue from contractual term extensions and a $5 million reduction in site rental operating costs.

Full-year organic growth remains projected at $135 million, or 3.4%, excluding Sprint cancellations and DISH terminations. The rate rises to 3.6% at the midpoint when prior-year DISH revenue is excluded.

Management continues to expect 2026 to represent the low point for organic growth. It cited existing master lease agreements, recent spectrum transactions, potential satellite-sector terrestrial networks, future spectrum auctions and rising mobile data demand as possible longer-term growth drivers. However, the company did not provide specific 2027 guidance.

Crown Castle also remains on track to deliver $2.1 billion of AFFO at the midpoint across the second half of 2026 and the first half of 2027. Management continues to target more than 200 basis points of margin expansion by 2030 compared with 2026.

Risks and Watch Items

• Core leasing is expected near the bottom of the $60 million-$70 million range because some opportunities are materializing later than previously anticipated. Management described this as a timing shift rather than a change in demand trajectory.
• AT&T’s 600 MHz deployment has progressed more slowly than Crown Castle initially expected as network design, engineering and equipment-vendor selection continue.
• The scale, timing and architecture of any SpaceX terrestrial network remain uncertain. Crown Castle said deployment speed would depend heavily on whether SpaceX uses existing towers or builds a denser microcell network.
• Edge computing revenue remains small relative to the overall business, and management is still evaluating whether deployments can scale profitably to hundreds or thousands of sites.
• DISH-related bankruptcy proceedings have been paused for court-ordered mediation. Management reported no specific progress update and said mediation would continue through early November.
• The timing of upper C-band clearing and deployment remains uncertain, including potential interference considerations.

Analyst Q&A Highlights

SpaceX and 800 MHz: Management confirmed that some equipment from DISH remains on Crown Castle sites. DISH had installed equipment at more than 12,000 Crown Castle locations, with the prior ability to expand to 20,000. Crown Castle said its approximately 40,000-site portfolio could offer a faster route to market if SpaceX pursues terrestrial coverage through existing infrastructure.

AI-driven network demand: Management said carrier network traffic could move over time from roughly 90% downlink and 10% uplink toward a more balanced profile as AI applications increase data uploads. This could require equipment upgrades, additional tower fiber and more edge computing and storage capacity.

600 MHz deployment: Management characterized the spectrum as highly valuable for indoor, suburban and rural coverage. It expects the customer to seek rapid deployment after completing network design and equipment selection, although the rollout has started more slowly than anticipated.

Customer agreement and churn: The transaction extends an existing tower portfolio agreement and limits long-term churn on the affected sites. Management said it is beneficial but not large enough to change Crown Castle’s long-term normal-course churn framework of 50 to 100 basis points.

Services activity: The company expects services activity to improve in Q4 after a weaker third quarter. Management said the business is inherently uneven by quarter and that delayed 600 MHz activity was the largest specific factor.

Full Earnings Call Transcript

Complete Earnings Call Transcript

Management Remarks

Operator

Good day, and welcome to the Third Quarter 2026 Crown Castle Earnings Conference Call. [Operator Instructions] Please note this event is being recorded. I would now like to turn the conference over to Mr. Hamilton West, Vice President of Corporate Finance and Treasurer. Please go ahead.

Hamilton West

Thank you, Chuck, and good afternoon, everyone. Thank you for joining us today as we discuss our third quarter 2026 results. With me on the call this afternoon are Chris Hillabrant, Crown Castle's President and Chief Executive Officer; and Sunit Patel, Crown Castle's Chief Financial Officer. To aid the discussion, we have posted supplemental materials in the Investors section of our website at crowncastle.com that will be referenced throughout the call.

This conference call will contain forward-looking statements, which are subject to certain risks, uncertainties and assumptions, and actual results may vary materially from those expected. Information about potential factors which could affect our results is available in the press release and the Risk Factors sections of the company's SEC filings. Our statements are made as of today, October 21, 2026, and we assume no obligation to update any forward-looking statements. In addition, today's call includes questions of certain non-GAAP financial measures. Tables reconciling these non-GAAP financial measures are available in the supplemental information package in the Investors section of the company's website at crowncastle.com.

With that, let me turn the call over to Chris.

Christian Hillabrant

Thank you, Hamilton, and good afternoon, everyone. As I reflect on my first year at Crown Castle as CEO, I am pleased with the meaningful progress we have made to sharpen our strategic focus, simplify the business and build a strong foundation for future growth. Over the past year, we have taken several significant steps towards becoming the best-in-class U.S. tower operator in addition to completing the sale of our small cell and fiber businesses. .

First, at the beginning of the year, we took decisive actions to align our cost structure with the current business model, while preserving our ability to support customers, deliver best-in-class efficiency and enable future growth. These actions resulted in $65 million of annualized run rate operating cost savings. Second, we launched an enterprise-wide transformation program to create a more efficient and scalable operating model that supports our long-term growth objectives.

Our transformation program spans the entire business and includes initiatives such as modernizing our systems, optimizing our facilities footprint and reducing corporate nonlabor goals. This year, we launched a new project management system, improving the application to air process and a new field services system, enhancing our asset information and utilization data. We are also on track to exit our transition services agreement related to the previous sale of our fiber and small cell businesses, which we believe will allow us to further simplify our corporate functions.

Third, we continue to scale our ground lease buyout program while meeting or exceeding our target returns. In 2026, we expect to purchase more than 400 ground leases more than double the number completed in 2025. We believe acquiring land under our towers improves our margin and increases operational control of our assets. allowing us to deliver more value to the customer by meeting their needs more rapidly.

Finally, we are also making progress strengthening our customer relationships and driving customer engagement. We resolved a long-standing commercial issue in the quarter with one of our largest customers that advances our strategic relationship and positions us for future growth opportunities. We acquired a residual economic interest on a subset of our tower sites and extended the customers' lease terms for those sites. We believe this agreement creates value for both parties by simplifying economics increasing predictability and limiting long-term churn risk.

We believe this progress positions Crown Castle to capitalize on an increasingly supportive long-term backdrop for wireless communications infrastructure. Mobile data demand, one of the clearest indicators of demand for our infrastructure continues to grow at a healthy pace and is showing no signs of stopping with the emergence of AI-enabled applications physical AI and the acceleration of data-only wireless devices. We are also seeing continued investment in spectrum by our customers and others in the wireless industry.

Over the past year, carriers and other wireless offers have committed significant capital to acquire spectrum assets, including AT&T's acquisition of 3.45 gigahertz and 600 megahertz spectrum. Verizon's acquisition of AWS-3 licenses and SpaceX's acquisition of AWS-4, H Block, and 800 megahertz spectrum. These transactions demonstrate the commitments of the wireless industry to aggressively invest and to secure spectrum resources needed to support future terrestrial wireless network growth.

Notably, as part of its 800 megahertz spectrum acquisition, SpaceX said it aims to "become the first network operator to deploy both satellite and terrestrial spectrum." We believe SpaceX's mobile network ambitions could provide an additional list in two ways: First, to invest directly by leasing towers to build their previously announced terrestrial network. And second, as a catalyst for network operators to achieve investment in their mobile networks to enhance performance and coverage against new entrants in the wireless marketplace.

Looking ahead, we expect the FCC to auction approximately 800 megahertz of spectrum, consistent with Chairman Car's efforts to advance spectrum policy and maintain U.S. leadership in telecommunications. This includes approximately 160 megahertz of upper C-band spectrum expected to be auctioned by July 2027. Historically, new spectrum deployment has been an important driver of network investment, and we believe the industry spectrum pipeline remains supportive of long-term infrastructure demand.

We continue to see momentum in edge compute infrastructure as demand for storage and compute continues to accelerate while many large data center deployments face multiyear construction and power delivery delays. Crown Castle now has 100-plus edge sites deployed in several concurrent trials underway. While initial results appear promising, we continue to evaluate whether these edge solutions will scale.

Crown Castle is positioned well to serve this demand in a capital-efficient manner through its nationwide network of tower sites, each with existing power and our band connectivity and can provide select distributed move-in ready locations. Our talented and dedicated teammates continue to play a critical role in our progress transforming the business, and we are seeing rising levels of engagement with our clear and focused strategy.

We are committed to investing in their continued development through additional training, individual development plans and training in the cost-effective use of AI. In conclusion, over the past year, we have taken decisive steps to simplify the business, strengthen our operating model and build the right leadership team. This progress allowed us to increase our full year 2026 guidance and positions us to capitalize on the sector tailwinds and opportunities.

We believe our pure-play U.S. tower strategy, investment-grade balance sheet and capital allocation framework positions Crown Castle to maximize long-term shareholder value. With that, I'll turn it over to Sunit to walk us through the details of the quarter.

Sunit Patel

Thank you, Chris, and good afternoon, everyone. We continue to execute on our strategy towards becoming a best-in-class U.S. tower offer in the third quarter. positioning us to increase our full year 2026 guidance. Starting on Page 3. Third quarter organic growth, excluding the impact of sprint cancellations and dish terminations was 3.5% or $35 million. Third quarter organic growth increases to 3.8% if DISH revenues are excluded from prior year segmental billings. This growth was more than offset at site rental revenues by $5 million of Sprint cancellations, $61 million of dish terminations and a $17 million decrease in noncash straight-line revenues and amortization of prepaid rent.

Third quarter services margin was $15 million, which we expect to increase in the fourth quarter. Turning to Page 4. We are increasing our full year 2026 outlook by $20 million for site rental revenues, $25 million for adjusted EBITDA and $5 million for AFFO to reflect first $20 million increase in straight-line revenues related to contractual term extensions as a result of the agreement Chris highlighted earlier; and second, a $5 million decrease in site rental cost of operations as we continue to drive efficiencies and increase our ground lease buyout activity.

Excluding the impacts of the transaction, Full year 2026 adjusted EBITDA and AFFO are each increasing $5 million. I would like to take a moment to further highlight our progress on driving cost savings. Since issuing our original full year 2026 guidance in February, if you set aside the increases to our revenue and the decrease from our services contribution, we have increased full year 2026 adjusted EBITDA by $20 million from run rate cost savings or approximately 50 basis point as a percent of our site rental revenues.

On top of this improvement, we still expect to expand margins by more than 200 basis points by 2030 relative to 2026. Turning to Page 5. Full year 2026 organic growth, excluding the impact of Sprint cancellations and dish terminations remains unchanged at $135 million or 3.4% of the point. Full year 2026 organic growth increases to 3.6% at the midpoint if dish revenues are excluded from prior year site rental billings. We continue to expect 2026 to mark the low point for organic growth.

Our full year 2026 organic growth outlook includes $60 million to $70 million in core leasing activity. And based on our current view of the leasing pipeline, we expect to finish the year towards the bottom end of the range. Now I'd like to provide additional detail on the customary agreement we completed in the third quarter. We paid $275 million to acquire a residual economic interest in a subset of towers originally acquired many years ago, extend the customer's lease term for those sites and limit long-term churn across this portion of our portfolio.

By acquiring the residual economic interest, we expect to increase site rental billings by nearly $4 million in the fourth quarter or $16 million on an annualized basis, and decreased amortization of prepaid rent by $2 million or $8 million on an annualized basis. We're not changing our full year 2026 guidance ranges for site rental billings, amortization of prepaid rent and interest expense as these fourth quarter impacts are captured within the existing ranges.

On an annualized basis, we expect the impact of the transaction to be approximately $4 million accretive to AFFO, excluding the impact of prepaid rent amortization. We expect the term extension will result in an additional $20 million of straight-line revenues in the fourth quarter. Turning to the balance sheet. We ended the quarter with 97% fixed rate debt with a weighted average maturity of 7 years and leverage at 6.5x and net debt to last quarter annualized adjusted EBITDA.

Net debt to last quarter annualized adjusted EBITDA decreases to 6.3x when including the impact from the incremental straight-line revenues related to the term extensions commencing in the fourth quarter. Lastly, we're increasing our outlook for discretionary CapEx by $40 million to $240 million to reflect the higher volume of ground lease buyouts we expect to complete in 2026. Discrinary CapEx, net of anticipated prepaid rent additions of $40 million at the midpoint is expected to be $200 million at the midpoint.

To wrap up, the actions we've taken this year to improve efficiency and strengthen the business, support our increased 2026 full year guidance. We also remain on track to deliver $2.1 billion of AFFO at the midpoint across the second half of 2026 and first half of 2027. We believe our investment-grade balance sheet disciplined capital allocation framework and focus on becoming a best-in-class U.S. tower operator positions us well to generate attractive long-term shareholder returns. With that, operator, I'd like to open the line for questions.

Operator

[Operator Instructions] And our first question for today will come from Rick Prentiss with Raymond James.

Question-and-Answer Session

Ric Prentiss

A couple of questions. First, as you know, we always say follow the spectrum really not the CapEx, and so 1 of the most interesting things, obviously, you've touched on it a little bit in the last couple of weeks was brain selling spectrum to SpaceX the 800 megahertz spectrum. So a couple of questions in regards to that. We've heard some to maybe all of the towers out there that DISH rented might already include radios and antennas or some equipment for the 800 megahertz frequency that SpaceX is buying. Can you help us again, is that something you understand that might actually already be on your towers?

Christian Hillabrant

Rick, the answer is there is some equipment that is on the site what specifically the radio types and bands are, I couldn't tell you off the top of my head, but that is correct that the equipment is up there from when DISH abandoned it.

Ric Prentiss

Okay. And obviously, it's also the old Nextel stuff. So our understanding is also it's still in the handsets for a lot of handsets that are out there. So that's also good have, but are you -- is it still like to you that is in the chipsets in a lot of the devices?

Christian Hillabrant

The 800 megahertz spectrum, again, I could comment specifically on what the carriers have out there. But obviously, I think more importantly, as you and I both know, 800 megahertz does solve one of the critical challenges for an eventual SpaceX deployment, which is you need low band to get into buildings, you need low band to be able to provide breadth of coverage. And so it's obviously a huge milestone in their acquisition of that spectrum. .

Ric Prentiss

Makes sense. Little confusing process, obviously, with the escrow set up at the FCC with EchoStar selling spectrum to AT&T. Can you update us as far as we've been hearing some chatter when we were in D.C. for the wireless history foundation tender a couple of weeks ago that the tower industry might be getting closer to figuring out who's going to get how much out of the escrow, and maybe an update on what are you hearing from MLAs with satellite companies? Any progress?

Christian Hillabrant

Okay. Well, let's handle the DISH 1 first. So here's a quick update. So the current proceedings in the bankruptcy court have been placed on pause. While the parties attempt to negotiate a settlement in a court order mediation. There's no specific update to report on the progress of that yet. But as we communicated, I think, on prior earnings calls, the fact that this has now moved into a bankruptcy process, will likely result in an accelerated time line compared to a traditional core proceeding. So that's -- I think that's a good news story and the parties will remain in this mediation through early November. And then sorry, your second question was just around...

Ric Prentiss

Update on any MLA progress from selling companies, whether it's Base, Amazon or anyone else? Just how are we doing there?

Christian Hillabrant

Yes. So I don't think I want to get into a customer specific detail, but you can assume that we're talking to all parties.

Ric Prentiss

Okay. And last 1 for me, a lot of all these interrelated questions. The equipment that was abandoned by DISH on the towers who -- if somebody wanted to buy that for pennies on the dollar or take ownership of that? What's the process? Would they come to you guys and different tower companies? Do they need to go to the bankruptcy court so they need to go to DISH. Like if somebody wanted to say, hey, there's some equipment up there that's interesting to me. What's that process look like? .

Christian Hillabrant

Yes. I think it's probably a question best left to the lawyers in terms of what that is. I mean from our perspective, dishes abandoned equipment, they walked away from it. We've given them every opportunity to claim that equipment and take it off our towers, which they have elected not to do and they've effectively abandoned it. But in terms of how that would work contractually, I think that would be something that'd have to refer out to the lawyers. .

Ric Prentiss

Yes. I guess Shakespeare wasn't right about first thing we do about the lawyers. But.

Christian Hillabrant

We love our lawyers, right? Thank you for the questions.

Operator

The next question will come from Brendan Lynch with Barclays.

Brendan Lynch

Available infrastructure is a company that I identified you as their vendor for deploying edge data centers. Can you elaborate on what you see as the potential growth for this relationship?

Christian Hillabrant

Yes. Let me start saying we're excited to partner with available infrastructure. As you know, they are building a nationwide network of edge new cloud data centers. And while edge data centers are not new to CCI, we look at this as an exciting niche space and look forward to seeing how it evolves over time. More broadly, I guess, an update on -- we've had this trial now for a number of months.

As you heard me announce earlier, we've now eclipsed 100 sites deployed. Part of what we're looking at is to learn from these trials and understand both the demand profile from potential customers, but also just gaining an inventory of what we actually have and what we can actually market to our customers. And as we've gone through this process, we've identified -- we have something like 5,000 existing structures or shelters that are already existing, that are out of our sights today in various levels of condition, but in many cases, able to take customers to move in right away.

We've been researching power, and I think you've heard us talk about before. The sweet spot for us is between, say, a 0.2 to 0.5 megawatt installation, utilizing the existing power on sites customers to be able to utilize it. Some customers have come back to us and have asked and inquired whether we would be able to bring additional power out to the sites. And just as a sampling, what we found across like 25 sites that we just recently completed. What we found out was across those 25 sites there would be an opportunity to bring up to 100 megawatts in aggregate across those 25 sites.

And this is power that could be delivered in approximately 9 to 11 months with very limited capital required from Crown. So I think from an optionality perspective, it opens up the possibility of providing even greater service to our customers. But again, until we've finalized on how we can scale this profitably, I would still characterize this as a trial phase where we're still learning a lot. And this is -- this update I'm providing with you now is indicative of what we're learning along the way and how this might emerge as a business opportunity over time.

Brendan Lynch

Great. That's helpful color. You've also suggested that you've hit an inflection point on organic growth. When we think about the cadence into 2027, can you break down where you anticipate an acceleration or deceleration between the components of core leasing to escalator as Sprint churn case and normal course churn? .

Christian Hillabrant

Yes. I just start by saying, we reiterate that we believe 2026 is the low point for organic growth in our business and across the industry. The reasons why we believe that are the following. One is the MLAs that we have in place that gives us visibility to that future contracted activity. We announced in this past quarter, we resolved this long-standing customer issue, which both strengthens our relationship and clears the path for future growth opportunities with the customer. the recent spectrum transactions, albeit somewhat delayed in the AT&T acquisition of 600 megahertz.

Verizon's acquisition of AWS-3 and now SpaceX with the 800 megahertz. These are all positive catalysts for towers. And then longer term, we've talked about the opportunity of 6G and the fact that the FCC and administration has made it a priority to earmark 800 megahertz of spectrum. And more importantly, stating 6G is a battle to be won and is a kind of a matter of life and depth for the country to win that battle versus others. And starting that spectrum acceleration auction in July 2027.

And then the edge infrastructure ecosystem that is now starting to emerge in a the underlying factor that mobile data demand continues to grow, and is even expected to double over the next 5 years and the powering of AI inference that's driving a much higher percentage of like traffic. All of these are activities of why we believe that 2026 is a low watermark for growth. But obviously, that has to be borne out into 2027 and beyond, but these are positive tailwinds for the industry and for Crown as we look towards our future growth, without me speaking specifically towards our 2017 guidance, which I'm not prepared to share today.

Sunit Patel

Yes. And keep in mind, now we've had some of the satellite players also declare the interest in a taste network. So I think that's a positive also. .

Operator

The next question will come from Michael Funk with Bank of America.

Michael Funk

So first, I'm going back to the 800 megahertz and SpaceX comments, if they were to deploy that nationwide, if you give some approximation for number of tower sites for a minimum level of service with 800 megahertz. And then your thoughts on the timing, the timing of a build of that scale, given any kind of labor constraints equipment purchasing, just maybe think about the framing for timing and tonal contribution to Crown Castle, if they do deploy the 800 megahertz. I have one more follow-up.

Christian Hillabrant

Yes, Michael, I think it's a tough question to answer because it really depends on the strategy in which SpaceX decides to go to market. If part of their goal is to move as quickly as possible across the terrestrial network, that would lead them towards maximizing the number of co-locations on existing towers and rooftops that exist today, and they have to accept the heights of what's available in the ad centers.

If on the other hand, they decided to go alone and to build thousands and millions and billions of microcells that might take quite a bit longer based on the scale of not only identifying the land available, but also the time, cost and effort of bringing backhaul to every site and power to every site. So that's a very good question. It really depends on what SpaceX decides to do in their network design.

I think what we would offer to them is the fact that we've got approximately 40,000 sites. In many cases, we do have prior installations that have already been through the zone and permitting process so that it could help them along with others in the industry of getting to market faster and getting that 800 megahertz deployed.

Sunit Patel

I mean 1 indication, it took DISH like 4 years. And as Chris pointed out, some of that is already up. So it'd probably be -- if they want it to, depending on the ambition. -- you could do it in a much more compressed time line depending on how many sites they want to get to, but they can move quickly if they want to. .

Michael Funk

Yes. So 2029 time line ballpark it for wireless service, utilizing dislike is obviously easiest line of sight to get there versus doing the micro cell strategy? .

Christian Hillabrant

This is the industry cheat code that exists in that is just built this network. I know the people that built that network, many of which were former colleagues at T-Mobile. They're very good smart deployment people. They move through very fast companies like Crown leaned in very aggressively to help them and staffed up in the services side to actually help enable that. So that does represent a huge time-to-market advantage if they go down that route. But in the end, it's going to depend on what route is speed of essence or some other model more appropriate for them.

I think you have to wait to hear from them. But I think on our side, we stand ready and available. And obviously, as one of the lead providers of tower space to dish, obviously, have the scale to really help accelerate that.

Michael Funk

One more, if I could. How are you thinking about the 600 megahertz contributing to leasing activity and timing, whether that's going to be a 4- or 5-year normal maintenance activity from AT&T or if you expect that will be new equipment, shorter duration for deployment?

Christian Hillabrant

I think if I'm being honest here, I think it's first, approval of 600 megahertz of the transaction, obviously, took additional time as they went through the FCC approval process. And now I think the customer is very much in the throes of doing the design engineering and selection of their OEM vendor, which would provide the equipment for them to put up. And so it's probably been a little bit slower of a rollout than we anticipated.

That said, 600 megahertz is the equivalent of beachfront property is the way that I described this. It's propagation is such that it helps to drive in building coverage in urban and suburban areas. It extends coverage in the rural areas. And the customer just paid $23 billion for the spectrum, it's highly valuable. I would think that they would look to deploy this as quickly as possible, given the value in the overall spectrum portfolio of what they've just acquired. And again, with the handset manufacturers having already these frequencies in for other customers. it's something that they can actually put to good views very, very quickly.

Operator

The next question will come from Nick Del Deo with MoffettNathanson.

Nicholas Del Deo

First, Chris, you noted that you're supporting over 100 sites. What would that number have been a quarter or 2 or 3 ago. And what portion of that 100 are supporting GPU deployments versus other use cases?

Christian Hillabrant

Yes. Let me start with the second part. I can't tell you exactly what is in those -- that these have been build-outs that have continued to accelerate through the course of the year. Starting at the beginning of the year, I think we're effectively at 0 or very few sites. So we've been able to build that number in that period of time. Again, the key learnings for us has been understanding what the requirements are of edge data customers that is different than our traditional tower customers, and making sure that we're able to provide them with that timely basis information to help them accelerate their deployment.

And so it's been pretty much a steady build. I think it peaked. We peaked in the third quarter. It's built up steam. And again, the key for us is this something that's scalable so that we're not just talking like 100 sites, but we're talking hundreds or even thousands of sites over time is when it really starts to make sense for us as a new line of business.

Sunit Patel

Nick, it's small in the scheme of our overall revenues, but it's a good sign. And I would say compared to the beginning of the year, we are talking to more players. And I think just doing some more research on our sites and power availability and supply lines, as Chris was saying, we can -- it looks like on certain subsets of sites can provide a bigger power footprint, which I think gets more in the sweet spot of what edge needs are. But again, it's still early days and we'll have more to talk about this in subsequent quarters.

Nicholas Del Deo

That's good to hear. And then maybe one on the leasing outlook. You noted it's going to come in at the low end of the expected range probably a little bit of a change from prior comments that leasing would be back-end loaded. So is that driven by leasing getting pushed into 2027 for some reason? Or are you seeing a shift in the underlying trajectory or something else going on? Anything in general would be great.

Christian Hillabrant

Just more broadly, I think we're seeing interest from many different customers, including our traditional mobile network operator customers in these edge data center operators. In order to finish the year above the low end of the range that we guided to, we would have to see these opportunities materialize earlier than we currently anticipate. So it is -- it's more of a timing issue. .

Sunit Patel

Yes, it's more a shift. I would say the excitement around what people want to do is probably build up, but it's a timing shift as opposed to a trajectory shift. .

Operator

Next question will come from Cameron McVeigh with Morgan Stanley.

Cameron McVeigh

So Elon yesterday said that he expects AI agents to drive orders of magnitude of growth in bandwidth demand, and I'm curious as we think of the extent that AI agents are driving higher uplink traffic. How much can the carriers accommodate with their existing equipment and potential software upgrades. And then to that point, have you had conversations with customers looking to upgrade equipment related to higher uplink traffic so far?

Christian Hillabrant

Maybe to put this in context and share at a recent analyst conference where we sat with the three CTOs of the operators. And what they described was a very interesting and exciting future around AI and the impacts on the carrier networks, predominantly that it would take the current network profile of where, say, 90% of the total mobile traffic data load is on the downlink and only 10% is on the uplink that longer term, that would move closer to a 50-50 between the downlink and the uplink, meaning that you have people in the field ingesting data through video, through voice, through a text whereby they would be adjusting that data and uploading it in order to gain real-time information.

And the implications of that are twofold. One of which is, to your point, there's some re-architecting of the network, both in terms of capacity and then the configuration of how to balance that traffic load up and down. Two is, it starts to talk to latency and it does become a bigger issue. And as you think about the edge of data compute is something where, over time, as those applications evolve the need to have very fast response times lends itself towards the potential for future growth in the edge data centers. So those are the two implications that we've heard directly from customers in terms of the difference in the customer profile. The broader context of the data growth, again, the latest data that I saw published was a doubling of total load over the next 5 years. This was the CAGR that was implied.

Sunit Patel

Yes, I think it's going to be a big growth driver as he was pointing out, I think it also means that fiber infrastructure to towers, I think he also talked about the implications for a need for a lot more edge compute and storage. So these are all things that should be good drivers for our business. And to your point, it is -- I don't think it's just a software upgrade, if you just have raw growth in bits and it takes part of more bits, it might need upgrade of equipment to over time. .

Operator

Next question will come from Michael Rollins with Citi.

Michael Rollins

Two different topics, if I could. So first, going back to the customer dispute, can you provide just a little bit more background? You said it was long standing. What was me about that dispute versus the normal types of relationships that you have with your large carrier customers. And the payment that you made, is that related to the deferred purchase price associated with some of the carriers to leasebacks where there's a certain amount of money that gets deferred into the future. .

Or is there just -- I know you mentioned some of the benefit you expect to lease in dollars, but any other context that you can provide there would be super helpful. And then just separately, back to the topic of DISH. Can you share how many of your 40,000 locations did DISH get to with their infrastructure before they hit the stop button.

Christian Hillabrant

Yes. Let me start with the second 1 first. I think the easy answer on the DISH was they had 12,000-plus installed sites with the ability to go up to 20,000 over time as they develop their network. And then in terms of, again, not so much a dispute, but just more of a long-standing customer issue. This is -- I would characterize it as just part of the normal back and forth that tower companies and their customers have over time. But it was something that, by resolving it opens up the door for additional growth, that was the key for us as a catalyst in getting this sorted.

Sunit Patel

Yes. So a couple of additional things. You asked if we were purchasing futures, this wasn't that at all. This is more to do with the fact that this residual economic interest is that we're getting benefits from some of the revenues. So this just was a simplification. There was a win-win for both of us. I think for them, it simplifies things for them on their side. For us, it was an accretive transaction for us. if you look at it on a multiple basis. And also, it came with a term extension and an agreement to limit churn on a longer-term basis. So I would say, if you look at it in terms of -- from a balance sheet perspective, yes, it's $275 million, of which $158 million we put on other intangible assets and then the other $118 million in just other current activities, but that's the value of the contractual term extension and limiting the churn. So that's what it was. It didn't really have anything to do with any purchase options or anything like that.

Michael Rollins

I appreciate all that detail. And if I could just follow up with one other then on that. So given what you described on the churn side, does this make the case stronger that annual churn ex the Sprint decommissioning side of it could stay below 1% per year?

Christian Hillabrant

Yes. I mean, so as you know, we said our churn is 50 to 100 basis points, 50, probably call it. But -- so it impacts churn. But as I said, it's for a small -- a subset of the towers, so not big enough to kind of impact our guide -- long-term guidance we've talked about with respect to churn, basically beneficial.

Operator

Your next question will come from Mike Ng with Goldman Sachs.

Michael Ng

I just have 2. First, I was just wondering if you could comment on some of the lower network services activity in the quarter and if you expect that to kind of return back to more of a normal cadence after this quarter.

Christian Hillabrant

Service Yes. I mean, look, if you look at our guidance for the service business this year, it's $85 million. So if you look at what we reported in the third quarter, that implies we'll see a nice pickup in the fourth quarter, a nice increase in the fourth quarter. So yes, it returns to more of a normal cadence, probably not at the levels we might have guided to, but yes, in that close enough in that vicinity, at least in the fourth quarter.

Michael Ng

Great. And I was just wondering if you could talk a little bit more about the operational efficiencies in the quarter and some of the kind of key drivers of the 200 bps of margin expansion over the next few years. Any change in terms of key things that could help you there?

Christian Hillabrant

Yes. I mean, so Chris talked about a couple of systems implementations we've done. We've got more of those coming up next year and the year after that. And then we also talked about exiting our transition services agreement that will lead to further simplification in future periods. So I think we've been consistent about saying our margin will improve 200-plus basis points between here and 2030. But the fact is we are -- we made more progress on that on this year and still sticking to that guidance, which I think should say on the margin, we are doing better than what we had indicated at the beginning of the year.

Sunit Patel

Well, particularly the GLBO as a driver of our business, we identified at the start of the year that we had a gap to what our key competitors had in terms of owning land underneath our towers and embarked on a strategic program to accelerate those activities, which we effectively doubled, more than doubled year-over-year and continues to be a large driver of the efficiency and cost savings that we're investing in today.

Operator

Your next question will come from Matt Niknam with Truist.

Matthew Niknam

I'll keep it at 2 as well, but I'll keep them simple. First, on services. I know you had maybe alluded to the expected downtick that we actually ended up seeing in the third quarter. Is there any indication of broader softer leasing from carriers ahead of next year's upper C-band auctions? And then secondly, on the edge data centers, can you speak to whether there's a general type or size of customer you're speaking to for these trials? And if there are any sort of specific use cases they're typically trying to solve for?

Christian Hillabrant

Yes. I mean, in terms of the services, you heard me refer earlier to the fact of that 600 megahertz have been pushed out. And so this was something that when we led in the guide to would have an uptick in 4Q is now that the transaction has been completed and applications are starting to flow is that is the largest individual component from a services perspective. In terms of the edge data centers, again, the traditional installation has been in a sub-1 megawatt in a 0.2 to 0.5. It's providing edge data compute at the cell site in the existing structures that we have, supporting use cases of kind of both a combination of capacity supply in a high-demand data center environment, combined with folks that need the low latency that's provided by the fact of our cell sites being as close to the edge as they are. Those are the primary installation types that we're seeing in that space.

Sunit Patel

Yes. And a couple of things. I mean, remember, I think we've said that the service business has always been lumpy for us in any given quarter. We've talked about that in the past. From a long-term trend perspective, and I think we talked about this earlier, we've seen a time shift, but the interest level in general continues to be pretty strong and active and maybe helped by the change in the competitive environment a little.

Operator

The next question will come from David Barden with New Street.

David Barden

I guess the first, if I could, would be the language around what SpaceX seems to be trying to do seems to be very prospective. Like it's going to maybe probably eventually someday build a network. Your services business has been very much about planning and permitting and those sorts of things. Usually, when companies are planning on building something they kind of come to you guys first and then they buy the spectrum, they don't buy the spectrum first and then figure out how to use it later.

Could you guys weigh in on that a little bit on how you really think what's actually happening there? And then the second question, if I could, please, would be -- regarding the upper C-band auction, we all know the auction is going to happen next year, but that the deployment is going to be delayed because of the potential interference with the FAA optimeter issues.

So when do you think that, that would actually begin. It's probably not going to begin at the last minute in 2030, 2031. It's probably going to begin sooner rather than later. So if you could give us a sense as to the conversations you're having on that topic, it would be hugely helpful.

Christian Hillabrant

Sure. Thanks for the question. So let's start with the SpaceX and I think you raised a great point. For those of us that have spent their careers designing, building operating networks, which I spent more than half of my career doing for Verizon predecessor in T-Mobile is you realize how complex these things are, how long it takes. And what I think Crown brings to the table is a strong pedigree of experience in helping customers get on air quickly from the time an application comes in to the time that we give them an NTP to get up on air. .

And so if SpaceX as serious as it sounds like they are in building a terrestrial network, I think we are a great partner for them in being able to help take a concept and turning it into reality of how to navigate through building a terrestrial network, which is, again, very, very different, obviously, than billing a satellite network as we'd be the first to admit.

So that's something that we believe we can add value to. And again, based on our portfolio and the fact of the partnership that we had with initially, that we have the majority of our sites in top 100 markets where the vast majority of POPs are located. This is something that we can help to solve for, I think, in a significant way. And so we stand ready when and if that time comes. Then more broadly on your comments on the upper C band. Yes. So there's always been this issue where spectrum goes to market an auction. It takes the time that it takes to do an auction, then the aftermath of that is the spectrum clearing before it can be put into use.

What I think I've shared in the past is where there's a will, there's a way. And so if you think back to when the original C-band spectrum was being put into place and there was issues with potential adjacent channel interference with the ultimates is that the industry found a way to work around that and to accelerate those activities. And so I don't know if you can put exact date on when it is. There's typically a deadline by which the FCC sets for having the spectrum clear, but oftentimes, that's not the actual data the spectrum is cleared and put into service.

So I believe it's 1 of these things where stay tuned to the space, but I would not anticipate that it takes that long. There is a desperate need to put additional spectrum to work based on the data growth that we've talked about before. And therefore, again, where there's a will, there's a way to find a way to clear it and get it up on towers as quickly as possible.

Michael Rollins

I really appreciate those comments. And I just want to be crystal clear that when you say that you stand ready to be a partner, the reality is that that SpaceX really hasn't reached out to do anything and that we're just kind of waiting to find out what the plan is. Is that what you're saying?

Christian Hillabrant

What I say -- I'll just reiterate what I said earlier, which is we won't discuss individual customer discussions, and you can assume that we're talking to all of them. .

Operator

And the last question for the day will come from Batya Levi with UBS.

Batya Levi

Just wanted to get back to the discussion about the dispute you had with the customer. Is there any reserve that we need to factor in as a reversal as we forecast ahead maybe a cash component that you are expecting to receive? And is this an extension of an existing portfolio option of a new lease from the key customer and any color in terms of how long that lease is and if it has a normal escalator would be helpful.

Sunit Patel

It's an extension of an existing portfolio. There are no reserves or anything in the future. I think it's sort of like what I described it. We are paying $275 million to acquire a residual economic interest on the revenue side. and that's what it is. But in addition to that, they extended the term of the contract for this particular set of towers, if you like, and agreed to limit churn on a longer-term basis. So I think those are the critical value components of this transaction.

Batya Levi

Is it a typical 5-year extension? With a 3% escalator? .

Sunit Patel

Specific extension, but yes. I mean you get some sense of it from the straight line revenue treatment, I can't give you more specifics on that, sorry.

Operator

And this will conclude our question-and-answer session as well as our conference call for today. Thank you for your participation. You may now disconnect.