Jeff Johnston: Hi there and welcome to the All Day Digital podcast. I am your host, Jeff Johnston. For this episode of the All Day Digital podcast, we’re going to do something a little bit different. Normally I have a guest on and we talk about a variety of different things.
But for this episode, I’m going to provide my perspective and expectations around three key areas in the digital infrastructure market and they are: competition in broadband, M&A and valuations in broadband, and then lastly, AI infrastructure. And when I say AI infrastructure, what I’m really talking about here is, this whole bubble narrative. I mean, are we, are we in an A I bubble and is that bubble going to pop this year? And if it does, a whole bunch of bad things are going to happen. So, those are the three areas I want to provide some perspective on.
So let’s start with competition in broadband and I think what we can expect in ‘26 is competition to intensify this whole competition narrative really started to surface a couple of years ago and I personally think it’s only gotten louder over the last couple of years — and understandably so because of what’s happening in, well, a number of different areas, but one in particular is fixed wireless. That’s been an incredibly disruptive technology for fixed line broadband operators and I don’t think that’s going to necessarily change all that much in 2026.
As a matter of fact, it might even get a little bit more intense because when you think about this technology from a wireless operator perspective, they love it. They love it for a few reasons. One is they get a new customer, without having to subsidize an iPhone to get them. So that there’s a win right there. And secondly, the customers who take what I call the killer bundle, which is a combination of a broadband service and a wireless smartphone service. Those customers tend to be very sticky and have lower churn than customers that just take a smartphone plan. So that’s another reason why these folks like it.
And then lastly, it’s a great use of excess capacity in the network. So if there’s a way to load that up with new customers and for that to be a growth opportunity that, certainly bodes well for that particular product offering. So I think we’re going to continue to see those large wireless operators lean into that and we might even begin to see some dedicated capital or some dedicated network equipment that’s deployed this year to continue to grow that business. So that’s one area we got to watch out for that I think will add to the competitive headwinds for the industry.
I also think that we’ll probably start to see a little little bit more fiber overbuilding out there. We really didn’t hear much about this before. I mean, normally if you’re in a smaller rural town, if you were the first one to build a fiber network, chances of somebody coming in and overbuilding you were pretty low. But we have started to hear about some more over or some overbuilding activity happening in some of these markets and I suspect that’s going to continue and it could come at the hands of the large wireless operators who have entered into these joint ventures with infrastructure funds to buy and build out fiber assets, it could come from that. It could come from private equity, expanding existing assets into new markets. It could just come from traditional wireless operators is building out on their own. So I think there’s a number of different angles we can kind of look at and expect to see, but I suspect that’s going to become a bigger part of the broadband story in ‘26.
And then lastly, I also think that we’re going to start to hear more about LEO satellites and I think in particular we’re going to hear a lot more from Amazon and their LEO satellite broadband initiative. And when I think about the disruptive nature of this technology, I guess I less kind of worry about Starlink, even though they’re off to an incredible start and have, millions and millions of customers and people seem to really enjoy it who have it. I think Amazon’s Leo service has the potential to be far more disruptive and it really gets back to bundling and the different products and services and bats and balls that Amazon has to play with, and how they could be disruptive in the broadband industry with that strategy. So when you think about Amazon and you think about their overall strategic imperatives, I would argue that it really boils down to two things: They want to grow their Prime subscription base, Amazon Prime subscription base, and they want to grow, of course, their Amazon Web Services business.
Well, when you think about the first priority, Prime subscriptions, bundling a broadband service with Prime could offer some pretty compelling value and could entice people to switch over to the Amazon Leo service and look, we know that LEO satellite or internet service is not as good as fiber, but when you bundle a bunch of things together and create a value proposition for consumers, you know that technology could be just good enough for people to move over.
And then lastly, when you look at what Amazon announced in 2025 about rural America, they said that they were going to invest $4 billion into rural America to improve kind of Prime infrastructure, so deliveries and probably more kind of edge locations for deployment and hiring more people. So clearly rural America is a focus for Amazon in 2026 and it just feels like bundling a broadband service in with Prime would go a long way in growing that part of the business. So I think, I think we need to, I think we need to keep a close eye on what Amazon is up to. I mean they are a disruptor, they’ve disrupted a lot of industries and you know you look at their bundling of Amazon Prime video and what that did to the traditional media or at least help due to the traditional media space, I think that we should be well aware and paying attention to what Amazon is up to.
OK, let’s move on to M&A and valuations and and for 2026, I suspect that we’re going to see this kind of an ongoing drumbeat of M&A activity in the broadband space. I can’t sit here today and look at any one particular event or market player that’s going to really accelerate M&A or for this for there to be some kind of a tipping point. So I suspect that we’re going to kind of have this ongoing drumbeat of M&A in the market, but I also think that strategic acquirers so these are existing broadband operators buying other broadband operators. I think these, I think strategic acquirers are going to be much more active in the M&A market this year than they have over the last several year. And the reason for that is because things have kind of normalized and settled down and I think we’re now at a point where it makes sense for some of these strategic acquirers to take a real close look at expanding kind of inorganically through acquisition.
So when you think back kind of in the post-COVID time frame there was like this M&A frenzy going on, you know, institutional investors, infrastructure funds and all these folks wanted to get exposure to the digital economy to broadband operators. Because if you remember in that post-COVID time frame, there was uncertainty around, whether or not we’re going to go to the mall and shop anymore or we’re going to ever go back and into an office and work together. So, investing in digital assets was very attractive for these investors and so they got very aggressive and started buying up a lot of companies and there was bidding wars anytime a broadband operator was on the market, and as a result, valuations were driven up very, very high, I would say probably the highest they’ve ever been.
And so for these strategic acquirers during this time period, they were looking at this and going well, it doesn’t make sense for us to go out and buy and participate in this in this bidding war because we can build coverage and networks cheaper than we than would otherwise have to buy them based on these valuations. So they kind of took a step back. But again now that things have kind of normalized, meaning valuations have come down. I think it makes sense now for the strategic acquirers to take a very close look and at buying versus building but look, to be clear, nobody’s rushing into anything right now. People are being very disciplined, very careful on what decisions they’ve made.
And in terms of valuations, I feel like they’re kind of range bound for the foreseeable future. The growth isn’t as plentiful, if you will, now than it was five years ago. So that that’s putting some downward pressure on valuations. And I also think that what we’re starting to see is it’s difficult for broadband operators to be able to pass along some of the inflationary cost pressures that they’ve been feeling. I mean if you just look at how much your broadband bill has gone up over the last five years compared to your grocery bill — and I suspect that your grocery bill has gone up a lot more than your broadband bill — but it’s not as if broadband operators haven’t had to deal with increased labor costs and equipment costs and so forth. So that pricing power maybe is not as strong as maybe we thought five years ago.
So I think as a result of that we’ll probably see valuations being range bound, but look, these are great assets to own. There still is a lot of investor interest in these assets. These are critical assets and certainly with AI, I think that these assets become even more important. So this is not by any stretch a doom and gloom forecast. I just think it’s some normalizing and a different profile of who your typical acquirer is going to be in 2026.
All right, that’s a good segue to talk about AI to wrap things up. And again, what I want to talk about here is not really AI from a technology perspective per se. I really want to talk about AI infrastructure from a bubble narrative and how should we think about the valuations of these companies and the capital spend going forward in 2026 and are we getting way ahead of ourselves? Are we overbuilding capacity, and therefore a bad thing’s going to happen?
And I understand why that is because when you look at the increase in capital budgets from the four main hyperscalers, but when you look at the increase in capital budgets as a percent of their revenue, it’s been pretty remarkable. From November of 2022 when ChatGPT was first launched, we’ve seen Google increase its capex by 100%, Microsoft 91%, Amazon 47% and Meta 26%, that’s through the end of last year. So really remarkable increases in capital expenditures, and these capital expenditures are really underpinning this whole AI revolution that we’re dealing with.
So the natural question is, how sustainable are these levels of capital spend? And if things start to taper off or roll over, that could spell bad news for the greater AI economy. And so, I get that, I get those concerns. But I think you can really boil this down to a very basic question that can shed some light on what’s going on here right now. And that question is this: Are these capital expenditures, are these elevated or increased levels of capital expenditures, are they a good thing for these companies? Meaning, are they increasing shareholder value? Or are these capital expenditures decreasing shareholder value? Because we know that when management is you know, a little bit radical or irresponsible with capital expenditures, shareholder value is destroyed and that doesn’t typically go on for very long. So just asking that very basic question, are these management teams making good capital decisions when it comes to AI investments?
And the answer is yes. They absolutely are.
I pulled some numbers and I looked at the return on invested capital metric. Now this is a key metric that measures just what I just talked about: how effective is management in deploying the company’s resources, their capital? When you look at the increase in return on invested capital for the four hyperscalers since ChatGPT was launched in November in 2022, you see some pretty remarkable results. For example, Amazon increased their return on capital invested by 550% since ChatGPT was launched. Meta has seen their metric increase by 52% since ChatGPT was launched. And then Google is 20% and Microsoft is actually down 10%, but overall the direction for the hyperscalers in terms of creating shareholder value based on these elevated levels of an AI-induced capital expenditures are very positive. So that to me suggests that as long as that remains the case, we’re going to continue to see these, you know, these very aggressive capital budgets and they would argue that they might even need to go higher because the more they spend, the more value they’re creating. So that’s certainly a good sign for the future of AI infrastructure for 2026.
The other thing I looked at, which I thought was kind of another interesting way to think about all of this stuff is, how effective are those hyperscalers at integrating AI into their operations? And what kind of financial impact is that having on their business? Because you got to believe that these four companies, Amazon, Meta, Google and Microsoft, they’re probably the first companies on the planet of any significance to be deploying and integrating AI into their operations. They’re not just building AI to sell it; they’ve been building AI for the last few years to integrate it into their operations. So what I did is I looked at how much more profitable are these companies since ChatGPT was launched. And again you see some remarkable increases in operating income which measures profitability.
So for example, Amazon’s operating income is up 380% since November of 2022. Meta’s is up 96%, Google 23, Microsoft 11. So if you believe that these four companies were some of the first companies to deploy AI within their operations and realize all the productivity gains and so forth. Well, that these metrics suggest that for the greater economy we should also begin to see margin expansion and profitability improve. And that again would be another positive indicator around just overall spend and investment in the AI infrastructure market. So I there’s some pretty powerful tailwinds here to suggest that the good times are not going away anytime soon in this market.
Now, there’s always the other side, right? There’s always the counter argument to all of that, and I want to touch on that here briefly as well.
What gets people nervous as they think about 2026 and AI, it’s all these what they call circular deals, there’s growing interdependency between all of these major companies in the AI infrastructure market. So Oracle and Amazon and Nvidia and CoreWeave and Microsoft and Meta and Anthropic and all of these companies, they’ve done massive deals between all of them. And the concern is that, well, if one company’s not able to execute on that deal, is there a ripple effect across the overall ecosystem because of that? And I think that’s a fair concern, specifically as it relates to OpenAI.
So OpenAI has talked about spending over a trillion dollars in capex between 2025 and 2030, I think that’s the range and the number — it’s in that that ballpark. Which is, which is enormous, Which is like just an insane amount of capital to that’s going to be deployed. Specifically though, when you think about OpenAI’s revenue model, well, we don’t know for sure cause they’re private, but they probably did $30 billion, maybe upwards of $40 billion in revenue in 2025.
So it’s hard for investors to appreciate or build into their expectations a $1 trillion plus capex budget from a company that just did, you know, call it $30 billion-$40 billion in revenue last year. So if OpenAI is not able to execute on the deals they have with Oracle and others, what does that mean overall to the economy? And do we start to see some storm clouds brewing and numbers getting taken in and things like that. So I think that’s a fair risk.
I feel ultimately though it’s more of a localized risk because really the four main hyperscalers are underpinning this thing that we’ve talked about and they are seeing great results from the capital being deployed. So there could be some bumps along the road. There could be some localized risk, but I don’t think it’s sort of a systemic risk across the whole ecosystem for this coming year at least.
So I’m going to leave it there. Hopefully, that gives you all a few things to think about. I appreciate you listening and I wanted to give a special shout out to my CoBank associates Christina Pope and Tyler Herron, because without those two, there wouldn’t be an All Day Digital podcast. Thanks for listening, and watch out for the next episode.
