BCE Inc. ($BCE)

Good morning, everyone. I'm Sebastiano Petti, and I cover the telecom, cable and satellite space for JPMorgan. I want to welcome BCE CFO, Curtis Millen. Curtis, thanks for being with us today.

Thanks for having me. Good morning, everyone.

Great. So just to kick us off, BC has undergone significant transformation over the past year. you reset the dividend, you've acquired Ziply launched Bell AI Fabric, which has a lot of momentum and I'm sure to touch on. And you also laid out a clear 3-year framework at the Investor Day. So now that you're 1 quarter into that plan and the growth vectors are becoming more visible. How should we -- how should you frame, I guess, the strategic opportunity from here for investors who still may think of BC as a traditional cable telecom. And I guess, where is the management team spending most of their time.

Yes. Ultimately, we're executing the plan that we laid out at Investor Day in October. So as you say, we're 3 months into a 36-month plan. And yes, you raised the AI Fabric. And ultimately, AI fabric might be a faster time line than we expected. We're about 40% of the way there. At Investor Day, we talked about 73 megawatts. We've monetized 300 on the side plus 28.7% out of the 773. So that opportunity seems to be coming faster than certainly we baked in October at our Investor Day, again, being fairly conservative in what we wanted to bake into the model, knowing that we're having conversations in terms of securing power where we secured up to 800 megawatts. Obviously, we're trying to secure even more than that. But that market, that industry is moving pretty quickly. But ultimately, we're executing the plan. As we can rip costs and spend away from legacy product, legacy network, legacy processes, and pivot into more and more kind of next-gen products and services, whether it's digital media whether it's AI fabric, whether it's fiber investments. I mean that's the plan. And again, we're allocating capital to where the best value is and the best returns are.

All right. Well, yes, so sticking with Bell AI Fabric and just Bell business for a minute. So I just want to touch on the comments around just the line of sight into the 800 megawatts. So on the call, you talked about monetizing the 800 over time. You just touched on already having 40% of the way through that. So maybe, I guess, I think also Merco kind of expressed a "very high degree of confidence, underscored by the conversations you were having with significant potential customers. So help us understand the demand funnel? What types of customers are driving that interest? And how do you think about the mix between hyperscalers, sovereign, or government workloads and maybe enterprise customers?

Yes. Interesting. So look, I think the most important thing is securing access to power and time to power. If you look around North America, for sure, there's not that much available power. So Canada ones will be in a fairly interesting in a market because there is a lot of power generation that's not already sold, not already taken up by the enterprise demand. So access to power, very important. Our role in the Canadian ecosystem for nearly 150 years really positions us well, especially in a world where the sovereign AI use cases are still ramping up, but the sovereign nature of assets winds up being important and important for decisions as to allocation of power and power in Canada largely owned and controlled by a government or quasi government organizations. So we are kind of right place, right time in terms of that, and we don't leverage during the winter, but it's a pretty cool place to go. So as a market for power generation, Canada is a pretty interesting place to be. And given our relationship and ability to build out an AI ecosystem as we've done in frankly, 11 months since we let folks know we're working on AI fabric pretty good spot to be.

So on Saskatchewan, you described it as a fully contracted take-or-pay 300-megawatt facility with an IRR of approximately 20%. And with additional upside as you think about the sovereign workloads. So construction is underway, equipment is ordered, I guess can you walk us through, I guess, the key execution milestones between now and the first revenue? And what are the gating factors? And how should we think about, I guess, perhaps the $400 million in prepayments setup fees, et cetera? .

Yes. So you're right. The revenue side of this, the EBITDA side of this is contractual. So we just need to deliver the facility. As you said, so 96% of the equipment has purchased orders on budget penalties if the equipment is late and being delivered. So feel pretty good about that. And then you start looking at just regular construction milestones. So we're prefabricating walls off-site. We'll move those on site. But think of this as get walls up, connect power. And ultimately, what we're doing is it's Shell, right? I mean we're building Shell and a flat floor connecting the HVAC chillers, coolers and the gen sets. Our tenants are doing the more detailed electrical work connecting their servers and Rags. And then you had talked about earlier, sorry, I didn't answer your question about who the other tenants would be, tenants like what we have. So Cerebos, core weave and then ultimately hyperscalers. So the sovereign AI of it that you asked about -- we do have the ability to claw back use cases for sovereign purposes. It would still be in the facility. It would just be offtake or would have to move to a different facility. So we're not taking the sovereign AI risk. But certainly, when the upside is there, we'll be in the right place to capitalize on it. So not quite the direct offtaker yet, but that's why we sold all 300 megawatts to Cerebos and Core from day 1.

Maybe for some of the folks in the audience and for myself have at least having covered the Canadian ecosystem, but just to try to understand the nuance in the market. There is power within Canada that's unique as you coming out of recent -- another session here, that's a gating factor for a lot of the AI data center builds and what have you. But do you think this model of -- because a lot of the power is government or quasi-government controlled in Canada, do you think this is a model that probably is replicable as you go forward of there's probably going -- if we think about U.S. or non-Canadian AI platform is moving NeoCloud moving -- maybe taking advantage of some of the power in Canada. Do you see this model of like, hey, this is your megawatt capacity, but perhaps leaving some for sovereign workloads or being able to perhaps claw back some of that power, some of those megawatts. Is that probably something like a model that probably gets replicated in the future as you think about some eating into or some of this demand?

I mean, ultimately, the demand is there. So I certainly think the model is scalable. The -- and look, there Again, we have a lot of sight into 800 megawatts of power. There's more than that in Canada. Obviously, we're having conversations to get more access to that. So I think the demand is scalable, the business model is scalable. TBD a little bit on how fast the sovereign AI use cases drives demand -- so whether or not you have to reserve some infrastructure to kind of change the chicken and egg problem historically of the infrastructure requirement to drive use cases and then back around -- so again, I know that sovereign AI use cases for government entities and heavily regulated entities will be there. I can't really underwrite is that 2 months, 6 months, a year, 2 years. It's certainly going to happen. We're in the right place. But until then, we'll keep scaling the other side of the business and driving infrastructure investment in Canada.

Got it. And then I think [indiscernible] touched on the possibility of securing additional power beyond the 800 megawatts. Can you give us a sense of where those incremental opportunities might come from geographically? And how are you thinking about the capital commitment required relative to your leverage targets?

Yes. Each province generates their own power. So not to say we're talking to everyone across the map, but we're going to talk to everyone across the map ultimately. And then capital requirements, Ultimately, we're not going to do anything that jeopardizes our 3.5x target by 2027. Saskatchewan is leverage neutral once it's up and running. So the economics are quite strong. If we're in a scenario where there's just too much demand that we can't afford it all ourselves, then we'll manage that with third parties. I mean we're obviously aware of partnerships that we've already signed in other business lines and be what's happening in the U.S. and in Europe, especially. So I do think there are opportunities. The ability to fund and come up with partners is not a gating factor for us, ultimately, gating factors, making sure that we land power and the revenue contracts at the same time because we're in the business of deploying capital, but de-risking the time line as we deploy capital, like in Saskatchewan, we're announcing construction plans and a contract when we already have the contracted revenue and EBITDA. So we're just overall in this environment looking to de-risk the overall CapEx spend.

Got it. And just thinking about the remaining megawatts, the remaining 427 megawatts that isn't yet contracted, should we think that this IRR threshold of 20%, similar to Saskatchewan? Is that what you're targeting? And I guess, how do you think about the risk return kind of trade-off?

It's yes, that is what we're targeting. And look, there's a dislocation between the risk/reward, where we're getting a risk-adjusted return. And at the same time, we already have contracts signed. So it's not a building and they will come. It's a -- we'll build it after we've already signed revenue contracts. And we're not spending money, as you know, on the actual chipset. So -- it is a very contractual business for us. And increasingly, as we build fiber and AI data centers, our business starts looking a lot more infrastructure heavy versus competitors in Canada and the U.S. .

And how should we think about the ancillary revenue opportunities from Bell AI fabric? I think Mercer described it as a full stack Canadian-controlled AI platform. So -- how do you think about that revenue opportunity above and beyond some of the core data center economics we've been discussing so far?

I think you described it exactly accurately. It's all upside for us. So whether it's fiber connectivity, which, again, is important in an AI world, obviously, lowest latency makes sense, and we are a fiber builder. We own fiber in Canada. So that's upside for us. Cybersecurity, again, wildly important in an AI world, Bell Cyber is a leading provider of cybersecurity in Canada. We'll wrap that around sales as well. And then our taco business is a managed service professional service, think consulting business for cloud migration, workflow management. So again, important for us to secure power and make things easy for the ultimate clients as we develop the full AI stack, some of it we own and control with our expertise. And then, of course, we're going to partner with the Canadian LLM Champion, Cohere, Coveo, Hyperscalers, which is what we've done for 150 years. We bring our own specific expertise product by product and then the overlap expertise is the fact that we can provide end-to-end solutions for our customers. So technology is different, but what we bring to the table with our 150 years of relationships and trust is exactly what's helping us secure power and deliver AI in whether sovereign or normal use cases in Canada.

And I think zooming out for a second here. As we think about Bell Business Markets stripping out AI fabric, I think a Tecan Bell cyber collectively grew north of 30%. And I mean how durable is that organic growth rate? And I guess, what's the pipeline or visibility like for those 2 businesses? .

Yes. Look, I expect it to continue to grow quite healthily. And if you stripped out the AI fabric revenue in Q1, enterprise business was still flat year-over-year, which on a relative basis is still quite strong. And fundamentally, that's because we're managing our core business, but we're also driving growth in cyber and in a taco. And if you look at the macro trends, I don't think cybersecurity is becoming any less important. I think if anything, the complicated nature of our systems and the kind of ability and desire for companies like Bell and other bigger companies in Canada and U.S. as well, to drive efficiencies off of platforms like Salesforce, ServiceNow and hyperscalers is doing nothing but getting more important. I mean, let's face it, we're all spending money on these platforms, and we need to drive the value out of those contracts and Teco Helps you do that.

And how should we think about the margin profile of Bell business markets as the mix shifts towards some of these higher growth solutions? .

Yes. I'm expecting it to be relatively flat. I mean AI fabric comes in at pretty high margins. The cybersecurity is solid margins ultimately, on the MSP side of the world, they offset each other. But ultimately, overall enterprise business, I expect to keep delivering margins.

So now shifting to the wireless business for a minute here. I think after 3 consecutive quarters of year-on-year improvements in churn, the postpaid churn reversed in the first quarter, which I think you guys attributed to just higher number, higher switching kind of given some of the aggressive competitive offers that were out there. But you've also talked about optimizing your offer strategy for customer lifetime value rather than just chasing volume. So help walk us through the competitive dynamics that maybe you're seeing today? And I guess what gives you confidence the improving trend will resume over time?

No, I do think I am still confident we're still confident that our churn profile improves over time. We all operate in the same market. If you look at our churn impact relative to our competitor churn impact, Year-over-year, our churn performance was much better year-over-year than there is. And look, that's a factor of lots and lots of blocking and tackling and to just overall drive customer experience, right? I mean number of billing systems from 18 down to 2, just making it right with the customer the first time, like making sure that the truck roll has the install and everything is working from day one. . Overall, blocking and tackling, customer experiences has just improved over time. Customer complaints have gone down. Our share of customer complaints has gone down. But overall, just a focus on driving a better experience for our customers. I mean driving more and more bundled households keeps improving churn. And ultimately, if you're selling fiber, we're selling other products also, where you're selling fiber, it is just a better technology, it's a better product. So over time, to be able to reduce our churn relative to cable operators. I mean, absolutely, I expect that.

And then on the fiber side, have you changed anything in your offer strategy to kind of help reduce the churn as well, whether it be, I think, contracts or something along those lines? .

Well, thanks for the prompt. It is the first time we actually offered wireline contracts. So if we're bundled wireline contracts are going to help us on the wireless churn side as well. And the other thing that helps us, we used to sell mobility in the Internet separately. Now it's mobility and Internet bundle, but content is also part of the bundle in Canada, a little less important in the U.S., I understand. But in Canada, especially on the streaming service and we own Crave, which has our own content, but HBO or Showtime content. So the -- we call it MISC, the mobility, Internet and Content Bundle also helps reduce churn on Internet and cell phones. And for us, we have a cost advantage because we're the owner operator of Bell Media, so the streaming service itself.

Now shifting to mobile phone ARPU, down 0.8% in the first quarter, an improvement relative to fourth quarter's decline. But on the call, I think Merco had touched on in the fourth quarter that you thought it would be possible to show some moderation in ARPU growth or show moderate ARPU growth rather by the time you exited 2026. But obviously, recent pricing environment probably makes that a little bit more difficult. I guess is that a 2027 story at this point in terms of ARPU inflecting back to positivity?

Yes, I think that's right. Again, April looks pretty good. It looks like Q3, Q4 of last year, but 1 month doesn't yet make a trend. So we said on the sidelines for part of the activity, but you can't sit out for the entire quarter. Ultimately, yes, you just get pushed out by next couple of quarters. And when you see total ARPU growth depends on what the industry does over the next few months. So it's hard to pinpoint a specific month, but I'd call it 2027. And ultimately, I'd say on the contracts that were signed in Q1, kind of fine. There was a bit more bring-your-own device activity than contracted activity, and there's a bit of an impact on your obviously APRU because that's just the math of it -- but the more BYOD you have, you don't actually spend the COA and our wireless margins actually went up. So there's a little bit of a disconnect between your ARPU performance and your and your margins. But again, long-term health of the industry, BYOD should remain relatively flat to what it had been second half of last year.

And is it a similar time line as we think about wireless service revenue inflecting back to positivity?

Yes, probably in and around the same time line. overage and usage tough to know if it's flattened out year-over-year, the trend is better. So we'll see. But I mean, the biggest impact, obviously, is the -- what you're signing folks up to on a monthly rate.

Makes sense -- as we think about industry growth, I think you and your peers have described a maturing market maturing Canadian wireless market, growing at low single-digit rate, obviously driven by lower population growth, but still steady penetration increases. How do you think the industry's volume -- how do you think about the industry's volume growth from here? And do we need to see policy changes out of Ottawa before seeing improvements in subscriber growth.

No. I think you'll continue to see subscriber growth. I don't think it's going to be what it was 5 years ago when the immigration new to category, new to Canada were at all-time highs. But our handset penetration has lagged the U.S. and Europe for the last 30 years. There's no real good explanation for it, but we've just been behind on the curve. So the increased penetration will continue to happen over time. And again, Canada in the medium term is going to be a net immigration positive country. If you look at our history, that's certainly driven a lot of population growth. It's just not over the next couple of years. So we're not baking that into the model, but over the medium term, for sure, it's going to be a driver.

Okay. And lots of focus on direct-to-device these days. And you mentioned excitement about the ASTS space mobile partnership from a direct-to-device connectivity perspective. I guess how do we think about the time line for commercial launch? And could this be a meaningful churn reduction tool as well over time?

Yes. I think it's all the valve type benefits. I mean I do think it's churn reduction. And I think it's overall just a positive for consumers. I mean I appreciate that in Canada, it's a little bit different than what telcos in the U.S. are thinking about when they're confronting LEO. But for us, it's a positive in terms of -- as you say, churn reduction and just an added value service, right? So whether you live in a bit more remote area, whether you live in urban and you have a cottage that's outside of wireless coverage, I mean it's -- we look at it as retention, acquisition, and ARPU increase as well. So ultimately, it is just a good product. And longer term, we'll have to see how that actually shapes our ability to reduce some of our legacy copper footprint along the way. But for the shorter term, I think it's just a benefit for consumers, which benefits us.

Got it. And shifting gears to the broadband business. You added close to 43,000 residential fiber-to-the-home subscribers in Canada in Q1, and combined with Ziply we're close to 50,000 net adds. At the same time, the Canadian fiber build is slowing, but how should we think about the trajectory of Internet loadings from here? I mean, how do you balance the strong penetration economics of your existing footprint against the slower industry growth backdrop?

Yes. Our fiber footprint continues to drive net adds, right? I mean it was 43% in Q1. It was 45% in the quarter before, 48% the quarter before that. So fiber continues to win market share, where we have fiber, we capture 50% plus of the market. And over the last 3, 4 years, there are a couple of million new homes that aren't fully penetrated. So I think you see just natural continued ramp-up. And ultimately, we have 75% of the wireline footprint in Canada. So it will pull forward wireless for us. But -- it certainly isn't fully penetrated. And frankly, I don't really see much of a ceiling. It's just a better technology.

And so how would you describe or what's the latest in terms of the competitive environment in your wireline footprint any changes.

I'd say fairly stable right? I mean our competitors came out with Internet revenue growth that seemed pretty good for them. Our Internet growth in Q1 was over 2%, 2.2%, so it's a pretty good market.

And then shifting gears to the U.S. fiber market. I think the expectation is that subscriber momentum will accelerate as the construction ramps from here. But you're focused more on the 2027 and 2028 financials. So maybe help us bridge the gap between the slower near-term subscriber trajectory and your confidence or your conviction in the 2027, 2028 ramp. I guess what specifically changes in the back half of this year? And I guess how should we think about the cadence of passings and penetration from here?

No, look, you're spot on, build engine is the biggest driver of value for us at this point. $1.4 million fiber locations passed looking to get to $3 million by the end of 2028. So the biggest driver of value for us in the U.S. is actually building out the fiber footprint. I mean it's early consistent in terms of being able to drive penetration of new fiber footprint. So the biggest driver for us is literally ramping up that build engine. And as you say, that will continue to ramp up through especially as we leverage the partnership with PSP continue to ramp up in '27 and hit our 3 million homes by the end of 2028. And again, as soon as you build out fiber, the data and Internet subscribers follow from there.

And any -- again, reiterating and you just did the $3 million by 2028, $8 million longer term. Any -- from a supply chain or logistics or obviously, permitting is always a big issue in the U.S. any snags there or if things are progressing as anticipated.

We, as you say, permitting, I don't mean it's blocking and tackling and it's permit community by community by community. And if you look at the permits they have, I mean, they're kind of double what they had last year. They've head count on the construction side. So it's just ramping up the build side of it. I mean we help bring them some scale in terms of access to suppliers and equipment. So no real issues on that side. It's just how do you ramp up even faster?

And now shifting gears to the 2026 guide on the financial guide. So you guided to a consolidated revenue growth of 1% to 5%, EBITDA growth of 0% to 4% and your CapEx was increased to take -- to contemplate the Saskatchewan build. But after 1 quarter, reiterated all of your guidance, I guess, where do you see the most though variability within the guidance range? Is it the wireless pricing environment, the pace of AI fabric revenue recognition, something else?

Yes. Look, we're confident in our ranges and our guidance at this point. So I don't really see a big impact either way. I do think AI fabric is probably accelerated a little bit faster than when we talked about it at our Investor Day. So that's a good news story. And frankly, the more that we can rip out spend on legacy networks, legacy product, legacy processes, and reallocate capital towards fiber AI fabric or more digital transformation. I mean we're going to take advantage of that opportunity, and that's what we're looking to do in '26. Now obviously, that has a bigger revenue EBITDA, free cash flow impact, '27, '28. But again, if you can drive value now and capture it to more absolutely, I'm looking to drive that value. So sensitivity of the model, wireless price is obviously not a surprise in market pricing. Would be a variable, but there's nothing that makes us back away from our current guide.

Does the -- as we think out the 2028 financial targets, I mean does the upsized AI-powered solutions target of $2 billion relative to $1.5 billion. Just from Saskatchewan. I mean does that incremental growth change how you're thinking about the multiyear EBITDA range? Or are there other kind of offsetting factors that have availed -- that have emerged since...

Yes. Look, Saskatchewan alone, we upped our guide after announcing 1 contract, right? Our free cash flow, 3-year CAGR. We upped it from 15% to 16.5% on the back of 1 contract. So look, as we continue to drive kind of capture of demand that's already out there and signed contracts, within the 73 megawatts that we've already talked about, that's just within our plan, we'll fund that. Look, if we can grow at a bit above that within the CapEx envelope, then we'll do that. And if there are other 100, 150-megawatt opportunities out there, we're certainly chasing them down. And let's say, it will be transparent when it happens, we'll announce it. And if we change the guidance from there, we will, but we'll certainly be transparent about the impact.

All right. Well, we'll be keeping an eye on. At the Investor Day, you outlined $1.5 billion cost savings target by 2028 with incremental efficiencies beyond that, driven in part, I think, by higher self installs, the 3G network shutdown, IT platform simplification. Can you give us a sense of, I guess, where you are on that journey toward the $1.5 billion? And how much has been realized to date, bigger -- some of the bigger buckets yet to come? And then just maybe help in terms of phasing over the 3-year period?

Yes, I'd say we're just over halfway. And we have pretty good line of sight into this. And it's the laundry list that you talked about, and it's a list of hundred other things. Ultimately, all within the how do we just rip out manual work, simplify our processes it's more cost efficient for us and it's better for the consumer, right? So as we think about how do you provide better customer experience, better customer service. Yes, it reduces return. But in this world, actually leveraging technology means it's more cost effective and cost efficient to deliver a better service to the customer. So it's kind of a win on all fronts there. And largely, it's leveraging technology ripping out manual work and meeting the customer where they actually want to meet with you, as you say, self-install digital capabilities.

And then with thinking about overall capital allocation, you guys have been pretty consistent in terms of your framework, which is to strengthen the balance sheet, fund strategic priorities, and return capital through a sustainable dividend. But we just talked about some of the simplification you're doing from a cost effort, but you also sold MLSE -- more recently, with the land mobile radio divestiture. But at the same time, you have growth initiatives in terms of $5 billion of CapEx this year because of Saskatchewan, but also Ziply, so maybe help us think about from your seat specifically as the CFO, how do you think about the tensions between accelerating investment in high-return growth opportunities while maintaining that pace of deleveraging towards 3.5 by the end of '27?

Yes.Well -- the good news is that there actually is tension here because we do have significant growth opportunities. So if there was no growth opportunity to be very different in discussion. But ultimately, like we drew a line in the sand, we should be deleveraging. We will be deleveraging 3.5 by 2027, we're going to hit that target. On the growth side, we still have the ability within our CapEx envelope to fund growth, whether it's fiber, digital transformation media or AI fabric. As you saw with the Saskatchewan announcement, it's net leverage neutral very quickly. and we increased the proportion of our revenue and EBITDA that's infra based and driven by contracts. So there is an ability to fund more AI fabric and still manage the balance sheet. But fundamentally, we'll hit the 3.5, and there are plenty of structures out there to fund incremental growth if the demand in contracts are right.

Are there any other noncore assets you'd consider divesting to accelerate that path to 3.5 by 2027? .

Yes. Look, the capital allocation is a tougher game, but it's a simpler game in a world where you have attractive opportunities, right? The capital allocation, as we talk about, it tends to be CapEx focused, but really, it's OpEx as well as assets. So like what we've seen, we announced $7 billion of asset sales, not that long ago, were $6.6 billion of the way there. So yes, there are a couple of other assets that we think we might not be the best owner of and that we can redeploy that capital, pay down debt and fund our growth opportunities.

Well, Curtis, I think that's a great place to end it. Thank you so much for joining us today, and thanks, everyone.

Right. Thanks for your time.