Earnings Call Transcript
BCE INC (BCE)
Earnings Call Transcript - BCE Q3 2025
Operator, Operator
Good morning, ladies and gentlemen. Welcome to the BCE Q3 2025 Results Conference Call. I would now like to turn the meeting over to Kris Somers. Please go ahead, Mr. Somers. Thank you. Good morning, everyone, and thank you for joining our call. With me here today are Mirko Bibic, BCE's President and CEO; and our CFO, Curtis Millen. You can find all our Q3 disclosure documents on the Investor Relations page on the BCE website, which was posted earlier this morning. Now before we begin, I would like to draw your attention to our safe harbor statement on Slide 2, reminding you that today's slide presentation and remarks made during the call will include forward-looking information and are therefore subject to risks and uncertainties. Results could differ materially. We disclaim any obligation to update forward-looking statements, except as required by law. Please refer to our publicly filed documents for more details on assumptions and risks. Now with that out of the way, I'll turn the call over to Mirko.
Mirko Bibic, CEO
Thank you, Kris, and good morning, everyone. Last month, at our Investor Day, we unveiled an ambitious and exciting 3-year strategic plan, which positions Bell for the future. This includes clear and transparent financial targets to drive long-term shareholder value and a refreshed brand that better reflects the full breadth of our customer segments. We also took the opportunity to introduce our investors to several members of the executive team and conducted deep dives into each of our operating businesses. As we execute against this plan, we're leveraging our proven ability to drive efficiencies to generate strong, sustainable free cash flow growth and total shareholder return, supported by a disciplined capital allocation strategy. Earlier this year, I shared our intent to focus on 4 strategic priorities, all underpinned by our unique and highly differentiated assets in fiber, wireless, media, and enterprise. In Q3, we continued to execute diligently against all 4 of those strategic priorities, and you can see this in our results. I'll start first with the customer. We have a reenergized focus on customer service, and it's really paying off. Thanks to the investments we've made, we reported a second straight quarter of significant postpaid churn reduction, and this is the direct result of customer service improvements, increased product intensity, and effective real-time retention offers. A few weeks ago, we launched new wireless plan tiers, each offering distinct value propositions. This innovative approach moves beyond traditional data bucket sizes, introducing differentiation based on network speeds and video quality, roaming and long-distance features, varying levels of device discounts, and content offerings. The response from our customers has been very positive. This construct gives customers more choice and reduces churn, all while leveraging our owner economics and content. Our premium Bell branded postpaid wireless loadings are significantly higher than our consolidated reported postpaid net adds, with Bell brand postpaid year-over-year growth exceeding 100%. This is a clear indicator of strong customer demand and the strength of the Bell brand in the market, and it's completely on strategy. We're also continuing to make meaningful progress on transforming the customer experience with initiatives like the AI-powered virtual assistant that we showcased on October 14. This and other AI-driven applications serve as a technological foundation for a next-gen customer experience. I'll move now to delivering the best fiber and wireless networks. As you know, for well over 100 years, Bell has built and operated the best networks in the country. And now, we also have one of the best network growth engines in the U.S. with Ziply Fiber. This was our first quarter of operating this asset, and its results are reported in our new Bell CTS U.S. segment. We're very pleased that Ziply Fiber's financial results continue to exceed our original investment case. Again, as you saw on October 14, Harold and his team are engaged and excited by the tremendous opportunity ahead. With the formation of the Network FiberCo partnership now complete, Ziply is well positioned to accelerate its fiber build and expand beyond its current 4-state footprint. Construction is set to ramp through 2026. Currently, Ziply's Fiber network passes 1.4 million homes in the U.S., and we expect to reach approximately 3 million locations by the end of 2028. Over time, we intend to leverage the Network FiberCo partnership to expand our U.S. fiber footprint to 8 million locations, and we'll do that in a cost-efficient manner. Including our U.S. operations, we added 65,000 net new fiber subscribers this quarter. In Canada, fiber continues to be a key driver of multiproduct penetration through mobility, Internet, and content cross-sell opportunities. We're focused on increasing the number of subscription services per household, with content bundling playing a central role in that strategy. In Q3, product intensity was up approximately 7% year-over-year, fueled by growth in content subscriptions. And as we shared at Investor Day, we plan to increase product intensity in the next 3 years by approximately 25%. So we're off to a good start. Our fiber advantage will grow with the availability of Wi-Fi 7, which works best on fiber and is going to improve the product intensity momentum. Turning to wireless. The environment has stabilized, and we expect this trend to continue. Wireless service revenue and ARPU both declined by less than 0.5 percentage point, while postpaid churn improved by 15 basis points. We also recently announced a partnership with AST SpaceMobile to deliver direct-to-cell satellite service. This breakthrough technology will expand our network reach, bridging the gap between the terrestrial 4G and 5G networks in Canada's most geographically challenging areas using powerful and reliable low-band spectrum. The initial launch of our service is scheduled for late 2026. The service will include voice, video, text, and broadband data capabilities using base stations owned and operated by Bell within Canadian borders, and it will be accessible with an ordinary smartphone. The partnership with AST will enhance network reliability, resiliency, and security for all those choosing Bell. Turning now to enterprise and leading with AI-powered solutions. We all know that the Canadian economy is changing, our industry is changing, and technology is advancing at an unprecedented pace. The AI revolution is in full swing, and it has the potential to change how we work, how we live, and how we connect. At the same time, global instability is rising, and Canada and other countries are reassessing long-standing relationships that, in some cases, seem far less solid than they once were. Against this backdrop, as we've been sharing, we've reshaped our strategy, and we're well positioned for growth in this new environment. In just the past year, we've launched 3 game-changing AI-powered solutions businesses, and they're all foundational to our long-term growth strategy. That's Ateko, Bell Cyber, and Bell AI Fabric. Each of these businesses is expected to deliver significant top-line and bottom-line growth as we execute against our 3-year strategic plan. I'm pleased to report that revenue from AI-powered solutions grew 34% year-over-year. Most of that is organic growth, and it's a strong validation of our strategy. Canada is having its AI moment, and it will be distinctly sovereign. According to a recent survey by the Harris Poll commissioned by Bell, 75% of large Canadian businesses consider AI to be a strategic enterprise-wide priority, with 91% of them prioritizing data sovereignty. This is where Bell holds a clear advantage. Bell's AI Fabric is precisely engineered to meet these exact needs. Our purpose-built AI data center business and the full stack AI alliance we've assembled with other Canadian tech leaders continues to have a deep pipeline of interest, and we expect to announce more growth in this space in the coming months. The public and private sectors share a fundamental role in building Canada's sovereign AI ecosystem, and the renewed commitment to AI that we saw this week in Budget 2025 is an important step forward that will support adoption, strengthen the economy, and help Canada compete globally. I'll turn now to the fourth strategic priority, which is building a digital media and content powerhouse. We recently introduced our new streaming bundles for Bell Mobility and Internet customers, which feature Crave, Netflix, and Disney+ all together in one bundle. Our commitment to sports content also remains strong. We announced long-term broadcast and streaming rights extensions for regional coverage of both the Montreal Canadiens and the Winnipeg Jets, reinforcing our leadership in Canadian sports media. We're also continuing to ramp our digital media capabilities. Our long-term partnership with iHeartMedia was expanded this quarter to include Canadian representation of iHeartRadio's extensive podcast portfolio, significantly enhancing our digital audio offering. Additionally, we entered into a strategic ad distribution partnership with Tubi, one of the largest and fastest-growing free streaming platforms in Canada. All in all, these initiatives will create new opportunities for digital advertisers to reach Canadian audiences across Bell Media's audio and video platforms. In short, we're executing with discipline, and we have momentum across all 4 of our strategic priorities. This focused path will continue, positioning us to deliver long-term sustainable free cash flow growth and enhanced shareholder value. As shown at our Investor Day, we have a highly coordinated and energized company that's fully aligned and ready to continue to execute. With that, I'll now turn it over to you, Curtis, for a review of our Q3 financial results.
Curtis Millen, CFO
Thank you, Mirko, and good morning, everyone. I'll begin on Slide 7 with BCE's consolidated financial results. Total revenue was up 1.3%, driven by the acquisition of Ziply Fiber completed on August 1. Ziply Fiber's operating results are reflected in our new Bell CTS U.S. segment, while our Canadian wireless and wireline operations are reported under Bell CTS Canada. Overall top line growth was moderated by retroactive revenue adjustments at Bell Media related to contract renewals with certain Canadian TV distributors in Q3 of 2024. Adjusted EBITDA increased by 1.5%, also reflecting the contribution from Ziply Fiber. This led to a 10 basis point margin increase to 45.7%, our strongest result in more than 30 years. Excluding the contribution from Ziply Fiber and normalizing for the aforementioned retro benefit of Bell Media last year, overall BCE's EBITDA grew by 0.4%. Net earnings and statutory EPS were up significantly over last year. This was largely due to the $5.2 billion gain from the sale of our minority stake in MLSE on July 1 and lower asset impairment charges compared to Q3 of last year. These non-cash charges were related to Bell Media's legacy properties to reflect the ongoing digital transition of the advertising ecosystem. Adjusted EPS was up 5.3%, supported by higher EBITDA. CapEx was down $63 million this quarter, bringing year-to-date CapEx savings to $551 million. We anticipate a year-over-year step-up in overall spending in Q4 as Ziply Fiber executes its fiber build-out, consistent with our 2025 capital intensity guidance of approximately 15%. The combination of lower CapEx, higher cash from working capital, lower severance payments, and the flow-through of higher EBITDA drove a $171 million increase in Q3 free cash flow. Turning to Bell CTS Canada on Slide 8. Internet revenue was up 2%, solid results showing we're striking a healthy balance between subscriber growth and disciplined pricing supported by fiber. Our business markets continue to build momentum with strong demand for our unique and differentiated suite of services. We saw sustained strength in AI-powered solutions, where revenue increased 34% year-over-year, driven by rapid growth at Ateko and Bell Cyber. We're excited about the opportunities ahead and remain on track to generate approximately $700 million in AI-powered solutions revenue in 2025. Wireless service revenue declined modestly by 0.4%, in line with the 0.3% decrease in Q2. When normalizing our Q2 results for the non-recurring revenue benefit related to the G7 Summit, Q3 service revenue performance showed notable improvement compared to last quarter. Wireless product revenue was up $41 million this quarter. This year-over-year increase was driven by greater sales of mobile devices. Our EBITDA result was in line with plan, with a notable 10 basis point margin increase over last year to 46.8%. This reflects our continued focus on cost management as evidenced by a 0.6% reduction in operating costs this quarter. Turning to our new Bell CTS U.S. segment, which reflects Ziply Fiber's operations for the 2-month period following the acquisition on August 1. As a reminder to investors, Bell CTS U.S. financial results are reported under IFRS accounting standards, consistent with Bell's other operating segments. We're pleased to report a strong start, with financial results tracking ahead of the expectations we set at the time of the announcement. Total revenue reached $160 million, driven by the strength of Ziply's Fiber-to-the-prem platform. Internet revenue grew 15% year-over-year, supported by continued expansion of Ziply Fiber footprint and strong customer penetration. Bell CTS U.S. delivered $71 million in EBITDA for the period, representing a robust 44.4% margin. This performance reflects both higher operating revenue and the benefits of Ziply Fiber's efficient cost structure and customer-centric operating model. The impact of Ziply's customer-focused model is evident in higher NPS scores and cost efficiencies. While the customer base continues to grow, customer contact rates are declining, now among the lowest in the U.S. market. Looking ahead, with continued operational discipline and a significant growth runway, we expect strong EBITDA growth for Bell CTS U.S. over the coming years, in line with the 3-year plan presented at our Investor Day. On the subscriber front, Ziply added 9,000 net new fiber customers in August and September, underscoring the strong momentum in expanding Ziply's Fiber customer base. Notably, fiber now represents 87% of total retail Internet subscribers. Total retail Internet net adds totaled nearly 5,000 subscribers, which reflects competitive losses in copper areas. Over to Bell Media on Slide 10. As projected, total revenue was down in Q3, decreasing 6.4% year-over-year. Excluding the one-time retroactive sub fee adjustment in Q3 of last year, the decline was closer to 1%. Despite strong digital ad growth in both video and out-of-home, total advertising revenue was down 11.5%, reflecting continued softness in traditional advertising demand for non-sports programming, as well as the impact of the previously announced divestiture of 45 radio stations. While Crave and sports direct-to-consumer streaming continued to grow, subscriber revenue declined by 5.2%, primarily due to the aforementioned retroactive revenue adjustments we lapped from last year. These adjustments were also a major contributor to the 6.7% decline in Bell Media's EBITDA this quarter. Excluding this one-time item, Q3 EBITDA was up 11.3% year-over-year. We're also pleased that OpEx was down 6.3%, showing you the focus we have on business transformation. Looking ahead, despite near-term headwinds on linear advertising demand, we remain confident that Bell Media will deliver positive revenue and EBITDA growth for the full year. Our focus remains unchanged for Bell Media to consistently deliver annual revenue and EBITDA growth while contributing meaningful free cash flow to BCE. Turning to Slide 11, our balance sheet is very healthy with $3.6 billion of available liquidity and a sizable pension solvency surplus totaling $4.5 billion. Our net debt leverage ratio at the end of Q3 was approximately 3.8 times adjusted EBITDA. This reflects the acquisition of Ziply Fiber, which closed on August 1 and was funded using the $4.2 billion in net proceeds from the MLSE sale received in early July, along with cash on hand. In late August, Ziply Fiber's outstanding debt of $2.7 billion was redeemed, partially funded by the $2 billion public debt issuance we completed earlier in the month. I'd also highlight that BCE's nominal net debt at the end of Q3 was $40 billion, which, despite the Ziply Fiber acquisition, is lower than the $40.3 billion reported at the end of 2024. Looking ahead, we remain sharply focused on reducing our leverage ratio to 3.5 times by the end of 2027 with a clear path toward 3.0 by 2030. We expect to reach these milestones through a combination of organic EBITDA growth, free cash flow generation, and near-term monetization of noncore assets. To conclude on Slide 12, we remain sharply focused on our 4 strategic priorities to drive growth across our key business units alongside our company-wide transformation to enhance efficiency. With the year-to-date consolidated financial results tracking in line with plan, operating momentum across the business and our consistent proven execution in a competitive marketplace, I'm reconfirming all of our financial guidance targets for 2025. I'll now turn the call back over to Kris and the operator to begin Q&A.
Krishna Somers, Vice President of Investor Relations
Thanks, Curtis. With that, we're ready to take our first question.
Operator, Operator
The first question is from Vince Valentini from TD Bank.
Vince Valentini, Analyst
Can you help us unpack the federal budget a bit? It's still not clear to me if these tax breaks for accelerated depreciation could apply to the money you spend on typical CapEx also on the data center front. Do you take that as they would co-invest in facilities with you, easing your investment burden, or that they would be a big customer and spend more on their own needs and do that only with sovereign providers like Bell? Those would be the 2 main things out of the budget. If you have any comment on the tower siting and sharing fiber builds stuff too, if you think there's any relevance there. Anything you can tell us would be helpful.
Mirko Bibic, CEO
Yes. I think as a general comment, Vince, I'd say that at a macro level, the budget is certainly positive in terms of having a number of initiatives industry-wide. I don't mean the telecom industry only, but just generally speaking, spurring more investment in the Canadian economy is a decidedly good thing. And I think there's a lot of pro-competitive, pro-investment initiatives in the budget that should be looked upon favorably. On the capital allowance initiatives, Curtis, I'm sure will add to what I have to say, but those are always looked upon favorably because I think they're a direct mechanism to continue to encourage companies to invest. On the AI side of things, still a lot to unpack there, Vince. So I can't answer your questions specifically just yet until we do more unpacking. But I would say that the initiatives that are outlined in the budget, the upwards of $900 million for sovereign AI and sovereign cloud shows that this government is committed to seizing the AI moment and encouraging AI adoption. And this is where we're at. We're at a moment in time where we need to move from AI science to industrialization at scale across the Canadian economy using Canadian tech leaders. I think that's the signal you should draw from the budget. So in that regard, it's a good thing. AI infrastructure in Canada by Canadians for Canadians. And I think we'll be able to capitalize on that in general terms because that's a measure to increase adoption in a sovereign way. So view that as a directional positive for Bell AI Fabric. And with the sovereign AI alliance that we've put together with Bell AI Fabric and Coveo, Cohere, and ThinkOn, and a number of other Canadian tech leaders, we're in good shape there, and you're going to see some growth at Bell AI Fabric in the quarters to come. I'll turn it over to Curtis to unpack some of that for you.
Curtis Millen, CFO
Vince. Just on the tax side of it, as Mirko said, it's certainly helpful. I think over the medium term, it's ultimately accelerating a tax shield. A similar proposal was instituted in 2018. We will see a benefit from that over time. I wouldn't expect a benefit in 2025 or 2026 based on the wording where it qualifies once the budget is actually enacted and given timing of spend. But in 2027 and 2028, we would expect a benefit.
Vince Valentini, Analyst
I guess you can't quantify that, Curtis, even in ballpark terms?
Curtis Millen, CFO
No, not yet. We've got to work through that, but we'll get back to you with more details when we have them.
Operator, Operator
Our next question is from Drew McReynolds from RBC.
Drew McReynolds, Analyst
Just on the, I guess, Internet competitive landscape. Obviously, a lot of focus continues to be on the TPIA regime. So just Mirko or Curtis, can you just provide an update on how you think the competitive environment is evolving here in Eastern Canada? And maybe an update on whether you've started out West with some of your initiatives highlighted at your Investor Day. Then just a quick second one on Northwest Hill. Is there an update just in terms of potential timing of getting that deal across the line?
Mirko Bibic, CEO
Thanks, Drew. I'll address those points. Let's begin with the Internet situation in the East. I want to reiterate what we communicated on October 14. Our approach is going to be twofold. In the East, our primary goal is to protect Bell's retail position, and we'll implement the integrated strategy we outlined on October 14, which I summarized today. Ultimately, we believe that fiber resellers will, overall, take more market share from cable. This will allow us to enhance our retail position with the Bell brand while simultaneously increasing fiber network penetration in the East. This is typically how the market functions. Moving to the West, our focus will again be to protect our mobility base, as I mentioned on October 14. We plan to do this by offering more services in a disciplined manner. This involves emphasizing our wireless tiers, leveraging our distribution strength in the West, and introducing no set-top box 5 TV or streaming bundles to boost wireless sales and reduce churn. When necessary, particularly for our most valuable customers, we will resell fiber Internet. Overall, we will become more competitive in the West. We currently have a trial in Kelowna and anticipate a full launch of fiber resell in the West by January. We will approach all of this very thoughtfully, and I believe that’s an important point to highlight. Regarding Northwest Hill, the buyers are still in discussions with the federal government to secure funding, and we are actively working to finalize that transaction. It appears that closing it will be more likely in 2026. However, considering the time this has taken, I want to emphasize an important point: we definitely want to close the deal, but we are also content to operate Northwest Hill and serve the residents in the North. It’s a strong asset. Whether we close the deal or not, it has minimal impact on our efforts to reduce debt. We were not looking to sell Northwest Hill due to debt concerns; our reasons were entirely different. We are eager to finalize it, and that remains our top priority.
Operator, Operator
Our next question is from Jerome Dubreuil from Desjardins Securities.
Jerome Dubreuil, Analyst
First one, a very helpful Investor Day a couple of weeks ago. One of the comments we've been receiving is some investors were expecting to see some margin growth down the road. We don't necessarily have a problem with that given the mix shift. I don't know if you can share, but do you expect margin growth on the Canadian telecom business between 2026 and 2028?
Curtis Millen, CFO
Yes. Jerome, it's Curtis. So yes, at Investor Day, we did talk about '25 to '28 revenue, 2% to 4% growth, EBITDA 2% to 3% growth. So at the high end of revenue, are you looking at margin compression? But ultimately here, what we also announced at Investor Day was our continued focus on operating cost reduction. So $1.5 billion of cost savings and frankly, more thereafter as we continue to leverage technology and our internal digital transformation. So I think you're going to continue to see a focus on cost containment. As you mentioned, I know your question was Canada-specific. Ziply, as we accelerate our footprint with the PSP partnership, those margins will decline over time, but still at a very 44.4% margin starting point for Ziply. So pretty healthy margins all around. Ultimately, in the range of flat margins is more what I would say.
Jerome Dubreuil, Analyst
Okay. And on the AI Fabric, I would like to maybe dive a bit more into the timing of the impact on your results. So maybe if you can reiterate the level of investment there, the expected financials and returns, and the timing of flow-through in earnings.
Mirko Bibic, CEO
Yes, I'll start and then Curtis can add any relevant points. We were quite open about the growth targets we outlined on October 14. These targets are based on monetizing 73 megawatts of power, although our ambitions exceed that. I would consider this a conservative growth projection since it only involves monetizing 73 megawatts, which is expected to generate between $100 million and $150 million in annual EBITDA. This creates a solid business with a strong demand pipeline. As I mentioned earlier, we anticipate some announcements in the coming weeks and months that will demonstrate how we are activating the sovereign AI alliance we established. It's very promising. Once again, it’s worth noting that we achieved 34% year-over-year revenue growth in Q3 in the AI-powered solutions segment. Most of this growth is organic and comes from Ateko and Bell Cyber. In Q3, there was no contribution from Bell AI Fabric. The advantages of AI Fabric were evident in Q2. There were no new announcements in Q3, but we expect AI Fabric to experience growth in the upcoming weeks and months, and we are very enthusiastic about this growth opportunity.
Operator, Operator
Our next question is from Maher Yaghi from Scotiabank.
Maher Yaghi, Analyst
So Mirko, I speak with a few U.S. domiciled data centers with subsidiaries in Canada. The view is that whatever they're offering in Canada is sovereign. The reason I'm saying this is, do you think the government is going to formalize what is considered to be sovereign AI to make it very clear to enterprises in Canada what constitutes sovereign AI and what does not? Because it's kind of like what she said, he said type of thing right now in the marketplace.
Mirko Bibic, CEO
Thanks for that, Maher. I'll respond in two parts. First, it's important to clearly understand what sovereignty means. Sovereignty goes beyond merely having a data center in Canada; it is a complex concept. It involves where data is stored, how it is transferred, who has control over that data, who manages the computing resources, and who governs access to the technology itself. This encompasses action, movement, storage, and governance. The second point is that I encourage everyone to review some of the federal government ITQs and RFIs that are available. These documents provide significant insights into the strict definitions of sovereignty as they relate to technology. To meet certain federal requirements, it's essential to be fully sovereign Canadian. Regarding Bell AI Fabric, we assure our customers that their data will remain in Canada. If the data needs to be transferred, whether from St. John's to Vancouver or any locations in between, it will always stay within Canada. This is a significant advantage for us.
Maher Yaghi, Analyst
I hear you. The reason I'm asking this question is I have looked at the RFIs that the government has put out, and they have their definition. But how can that permeate into the enterprise market? Because I mean, I agree that getting a contract from the government is likely to be easy — their definition of what constitutes sovereign AI is probably the highest level of conservatism, let's say. But what about the general enterprise, the Canadian enterprise market where, so far, it seems like it's still not clear what constitutes sovereign AI? So do we need some form of formal definition by a government agency to kickstart this new era of sovereign AI in Canada? Or do you think it might happen without some formal regulation?
Mirko Bibic, CEO
I think it will happen without formal regulation. The market will speak, Maher. The enterprise market in this space will be like it is in all the other enterprise vectors we operate in. I think our customers are going to rely on the providers that they trust, that they have deep relationships with. I think there's going to be a preference for Canadian, and that's with or without the geopolitical concerns that permeate today. I think the geopolitical concerns just actually further help market demand being tilted towards Canadian providers. If you take a step back, in our case, we have the deepest enterprise relationships. We have the most long-standing enterprise relationships. We're speaking to our enterprise customers on AI-powered solutions and growth vectors at the same time as we're talking to them about the core business relationships we’ve had for a long time. So as they seek to lean into AI workloads, we're in good shape in terms of having the infrastructure ready now. A key advantage in AI is time to power and time to compute, and we can deliver that. We can deliver data centers that are connected to the best networks, all located in Canada from a company they trust that has provided reliable service to them for over 100 years. So on the enterprise side, I think at this point in time, we're just going to rely on our unique market advantages.
Maher Yaghi, Analyst
Okay. Maybe just my follow-up question on wireless. You mentioned something interesting in your prepared remarks related to the amount of postpaid subscribers you loaded on the Bell brand. I'm happy to see how you guys are pivoting your offering, making less emphasis on data buckets and more on quality, content, and value-added services in wireless. Can you maybe just dig a little bit deeper into what happened in Q3 as you made that pivot? Because when you look at the numbers from a big picture point of view, we see a lot more prepaid and postpaid loading this quarter versus last year. And so I'm trying just to make sure to understand what you meant by in your prepared remarks.
Mirko Bibic, CEO
Thank you for the question, Maher. Our primary focus remains on the financials. We are working to balance subscriber growth and the associated economics. This is reflected in our financial results, including service revenue, ARPU, and a notable improvement in churn. Additionally, you can see this in our increased product intensity. There are numerous factors that have shown significant improvement. Regarding the postpaid wireless figures, if you examine them closely, the Bell brand postpaid net additions are substantial. While the total consolidated number is 12,000, it's important to note that the decline is mainly in the flanker brand net additions. The strategy revolves around focusing on the financials, the Bell brand, and multi-product offerings. Furthermore, the new wireless tier plans are expected to drive positive subscriber growth in the upcoming quarters, along with strong churn and financial metrics.
Operator, Operator
Our next question is from Stephanie Price from CIBC World Markets.
Stephanie Price, Analyst
Hoping you can give us a bit of further color on the U.S. Internet environment now that you've had a full quarter of Ziply. Are there any changes to your thoughts on the pace of the U.S. fiber rollout here?
Mirko Bibic, CEO
No, everything is on track, as Harold outlined in detail on October 14. So we're very pleased with Ziply's performance. We're looking forward to continued growth. The key thing is that it continues to perform ahead of our investment case.
Stephanie Price, Analyst
Okay. And then a follow-up for Curtis. Just on free cash flow growth. It was very solid in the quarter and year-to-date, but full year guidance is obviously maintained. Hoping you can talk about the puts and takes here as to get you to the bottom and the top of the full year range for free cash flow.
Curtis Millen, CFO
Yes. Stephanie, thanks for the question. Q3 specifically was strong. There are some timing impacts I'd just say off the top, still confident in the 6% to 11% full year. CapEx, we still expect it to be in the 15% range for the full year. It was a little bit lighter in Q3. Those are timing items, and we expect Q4 to be a heavier CapEx spend, which is probably what you're picking up on. So really nothing but reiteration of our full year expectations and timing in terms of CapEx and a few working cap items.
Operator, Operator
Our next question is from Sebastiano Petti from JPMorgan.
Sebastiano Petti, Analyst
I guess just maybe following up a little bit on Jerome's question regarding the cost savings. I guess, can you update us where we're at in the $1.5 billion? Yes, I think at the Analyst Day, you talked about being halfway there. Help us with the shaping of that? And maybe just remind us, Curtis, is that $1.5 billion a run rate exiting 2028? Just some color on the shaping there would be helpful. And relatedly, as we think about the CAGRs on EBITDA, particularly understanding there might be some margin compression as Ziply and PSP ramp up over time. But help us think about the shaping of the consolidated EBITDA growth over the forecast period. Does it make sense that 2026 growth is maybe above trend as we kind of think about the guided range?
Curtis Millen, CFO
Thanks, Sebastiano. A few things to unpack there. In terms of margins to tie up Jerome's earlier question and how that flows through. I think cost savings, you're right, we're halfway through. That will continue to accelerate. So that's lumpy with initiatives but continues to increase year-over-year as we leverage digital transformation and efficiency initiatives. You see relatively flat EBITDA margins over time. Our focus on reducing costs and driving efficiencies allows us to absorb some of the strong subscriber growth in the U.S., which includes COA, and it allows us to fund our other growth businesses, like AI-powered solutions. Those investments in the newer business, not only CapEx, it's OpEx also. So flat margins on a percent basis, increasing margins on a dollar basis as we transform the financial profile to be more heavily weighted toward future-focused products and services while maintaining similar EBITDA percent margins. In terms of the U.S., again, 44% margin we've talked about between 40% and 45% in the U.S. So again, you'll see offset of efficiencies and scaling benefits in the U.S. with incremental costs and expenses to drive subscriber growth as well as leveraging the PSP network code, which comes with a couple of incremental costs. But ultimately, looking to drive EBITDA margin dollars while we focus on cost containment and efficiencies, but overall driving EBITDA dollars out of the U.S.
Operator, Operator
Our next question is from Tim Casey from BMO Capital Markets.
Tim Casey, Analyst
Mirko, when you think about the growth in AI-powered solutions, doubling the $750 million by 2028, will it be a balanced contribution between Ateko, Cyber, and AI Fabric? Or is one of those more likely to be an outsized contributor? As a follow-on, at AI Fabric, do you not have to onboard the additional data centers? I think it takes 9 months to build one. Is there a cadence to that, that would be more back half or loaded more towards the back half of that timeframe as you onboard these data centers? Or am I thinking about that incorrectly?
Mirko Bibic, CEO
Thank you, Tim. To break down the $1.5 billion forecast, we anticipate around $400 million from Cyber, approximately $700 million from Ateko, and about $400 million from AI Fabric. You can refer to John Watson's Investor Day presentation for the expected annual growth rates for each segment of our AI-powered solutions. Regarding AI Fabric and the timeline for building data centers, I want to keep this general. We expect around two to three launches in 2026. While we will indeed sign contracts in the first half of 2026, the revenue won't be fully realized until 2027. However, we anticipate some activity in 2026 due to our current pipeline and our capability to open a few data centers next year.
Operator, Operator
Our next question is from Aravinda Galappatthige from Canaccord Genuity.
Aravinda Galappatthige, Analyst
I just wanted to go back to the free cash flow question, Curtis. Obviously, you benefited a little bit from working capital movements. I think that's what you're referring to in terms of timing. So should we expect a more unusual or higher-than-usual working capital outflow in Q4? The larger question is, when I look at the '25 to '28 free cash flow projections you provided, it suggests that you still anticipate fairly meaningful working capital outflows right through that period. I wanted to understand a little bit more why that needs to be consistently fairly high. Maybe I'll just start there.
Curtis Millen, CFO
Yes. Thanks for the question, Aravinda. Obviously, there are puts and takes within the free cash flow buckets I can't control all the timing on individual payments. I'd say we are continuing our focus on working capital management. We've seen benefits in inventory. We're managing receivables, especially. I do think we're seeing goodness there. Again, that fluctuates quarter-to-quarter. Overall free cash flow, I'm confident in our full year 2025 guidance. So no real change there. CapEx is a bigger driver of quarterly fluctuations. We do expect Q4 to be seasonally high relative to Q1 to Q2 and Q3 and higher than what Q4 was last year. So that's this year. Ultimately, as we talked about free cash flow from '25 to '28, there are a few things happening. One, CapEx dollars remain flat, but lease repayments will come down. As I think of the capital investments over time combined between what is accounted for as CapEx versus leases, that total bucket comes down over time. I do think the decrease in that combined bucket leads to payable decreases over time as well. You don’t capture all the free cash flow benefit in the same period your CapEx comes down. But over that period, we will capture the benefit. Beyond that, ultimately, you're just normalizing your way through. By the time you're at 2028, there’s no buildup or onetime. You're just a run-rate free cash flow generating business.
Aravinda Galappatthige, Analyst
Understood. And just my quick follow-up on the comments you made on the enterprise side, Mirko. As you kind of push ahead, talk to the CIOs with Ateko and Cyber, what kind of competition are you coming up against? Is it sort of the more fragmented independent players? Or is it the larger established players? To what extent do you have to unseat some of these existing incumbents in that area to win over the IT services and cybersecurity mandates?
Mirko Bibic, CEO
Yes. Thank you for the question, Aravinda. I'd say when it comes to the collection of the overall offering we have in AI-powered solutions business, the full stack AI, the AI platform, the integration of AI automation platforms through Ateko and Bell Cyber, when you put all that together, frankly, we're 1 of 1. I’m not exaggerating. When you look at all 3 together, we are 1 of 1. If you want to look at each one discretely, there are others who provide cybersecurity solutions, but nobody who can provide the integration of the Bell network security platform with the automated cloud-based AI-powered SOC that the former strategy, now the combined business is called Bell Cyber offers. On Ateko, there are a number of systems integrators out there. We all know their names. What stands Ateko apart are many things, but two, in particular, I'd call out. One, we are very focused. I've talked about this before. We have a clear lane of verticals where we’re targeting. We are hyper-focused on AI and the hyperscale platforms, whether it’s ServiceNow, Salesforce, or the 3 hyperscalers. We provide the benefit of really focused expertise. The other thing that stands Ateko apart is that we are both an operator and an integrator. Ateko does for our third-party customers what it does for Bell. We know the use cases that work and where customers waste their time. I think that's a value-add that the classic SIs don't provide. You see it in the growth numbers. We only started this in 2023 with Ateko. We've launched on May 29 of this year, AI Fabric, and it’s strong growth in its own right.
Operator, Operator
Our next question is from Batya Levi from UBS.
Batya Levi, Analyst
Question on the wireless side. Looking at your postpaid wireless base, is it possible to get a rough mix of what the Bell branded subs make? As that mix grows, I think we can expect wireless ARPU returning to growth? I believe the new MVNO also contributed in the quarter. Is it possible to quantify that?
Mirko Bibic, CEO
The network revenue from a wholesale relationship on wireless is insignificant. It's not a substantial figure, so it didn't significantly impact ARPU stability. There is considerable growth in the postpaid numbers you’re seeing. When looking at the details, the Bell brand postpaid net additions are quite large. The decline is mainly in the flanker brand net additions, which results in the consolidated total of 12,000. However, the Bell brand postpaid figures are quite impressive and align with our strategy. As mentioned in response to Batya's question, we're satisfied with the positive trend we're observing in ARPU as wireless pricing strengthens. It's also crucial that new and monthly rates exceed the existing base rates, which is helping to improve blended ARPU.
Operator, Operator
Our next question is from Lauren Bonham from Barclays.
Lauren Bonham, Analyst
You mentioned that you plan for the AST satellite service to launch in late 2026. Could you talk more about how that service will be marketed and included in the mobile plans, or what sort of price points you're targeting, as well as how you're thinking about the size of the potential market there?
Mirko Bibic, CEO
Thanks, Lauren. Look, the question is very on point. It's a little bit too early to give you information on that, given that we plan to launch in late 2026. We’ll come back to that at the appropriate time. But I get the question. It’s really important; it's just a bit too early. As we're running out of time, I want to thank everyone for the time you've given us this morning. We're on the right track here, and the stage is set for executing against '26, '27, and '28. We've improved the fiber net adds are growing in Canada and the U.S., improved fiber Internet churn, and improved product intensity. You see some stability in ARPU. The Bell brand postpaid loadings are very strong. Ziply is better than the investment case. You can see our Crave growth is impressive. The digital revenue is progressing exactly as we said, and the AI-powered solutions business is delivering better than we expected. We're on our way to the $1.5 billion that we projected in 2028. All the key parameters of growth are there and are delivering. It's an exciting time as we've reset Bell, and the team across the board is energized. Thank you.
Curtis Millen, CFO
Thanks, everyone.
Operator, Operator
Thank you. The conference has now ended. Please disconnect your lines at this time, and we thank you for your participation.