In Good Company with Nicolai Tangen
3 Dec 2025 · 45m · Nicolai Tangen, John Stankey · 25 quoted facts
AT&T CEO John Stankey argues that telecom is being reshaped by competition between wireless, fixed and satellite networks, and that AI will start a new growth cycle in data demand. In his view, fiber is the foundation of every network, and AI will need much more upstream (upload) capacity. He says AT&T should stay focused on connectivity in North America rather than build data centers or expand abroad.
The central tension is that a 150-year-old company must change fast while its leader is a lifelong insider. Stankey says the WarnerMedia deal taught him that reinventing two industries at once costs too much, so AT&T sold its content business. He now faces the harder task of cultural change across 130,000 people, using incentives, separate teams and an AI center of excellence.
Stankey says most consumer networks were built with more download than upload capacity. AI needs video and sensor data sent up to hyperscaler models and results sent back, which is why he favors symmetrical fiber and low-band TDD spectrum.
Stankey says repositioning in telecom and media at once needs more cash and patience than public markets will give. AT&T divested content to invest more heavily in its core connectivity business.
He says countries are growing more protective of foreign ownership of critical assets. He also says US telecom returns beat other markets, and AT&T has enough organic growth from fiber builds at home.
Stankey says only a handful of firms with their own chip designs will win at data centers. He sees some irrational exuberance and thinks power supply will limit how many announced projects get built.
AT&T uses pay incentives, separate teams measured differently, a central AI expert group, and partners who test new AI platforms for free. Stankey calls the broader shift a five-year journey with years still to go.
Stankey says AI will push far more data upstream, from sensors and cameras to cloud AI models. Fiber is symmetrical, meaning it carries as much upload as download, so it handles this better than networks built mainly for downloading.
Stankey says satellites are mostly complementary. He expects AT&T to sell satellite access as an add-on, such as day passes for remote trips, starting with real consumer services in 2026, though low-earth-orbit satellites may take over some Internet of Things uses.
Stankey says it needed to focus on connectivity and invest more in its core network. He argues that trying to reinvent two industries at once on a public balance sheet needs more cash and patience than markets will give.
He sees some irrational exuberance among data center investors and expects some carnage, with fewer winners than there are investors. He thinks limited electricity supply in the US will stop many announced projects from being built.
Stankey says AI already helps with customer sales, service and support, which lowers costs and reduces customers leaving. It is also being used for more dynamic pricing and for adjusting cell towers in real time, such as during a marathon.
Stated by the speakers. Each quote matches the transcript, which shows what was said, not that it is true.
AT&T (American Telephone and Telegraph Company) traces its origins to Alexander Graham Bell and about 1877, according to host Nicolai Tangen.
“AT&T American Telephone and Telegraph Company, you know, founded by Alexander Graham Bell, going back to, like, 1877. That's quite a history, right?”
John Stankey says AT&T is approaching its 150th anniversary the following year.
“Just about on year one fifty s next year, so it's a big deal.”
Stankey says technology has created real competition between wireless, fixed and satellite networks in the United States, causing a reordering of telecom assets after the Telecom Act's deregulation effort.
“There's real intermodal competition between wireless and fixed and satellite, and as a result of that, there's a reordering of assets going on in the United States.”
Stankey says the telecom industry, being highly competitive, tends to grow at roughly GDP or slightly above, rather than earning the dominant-position margins of the Magnificent Seven.
“we tend to grow at GDP or GDP, a little bit plus.”
Host Nicolai Tangen says AT&T bought WarnerMedia, dating the deal to 2016.
“Now, if we go back some years to, um, 2016, you you bought WarnerMedia.”
Stankey says AT&T divested its content business so it could focus on its core connectivity business.
“It was it was necessary, it was, uh, in order for us to be really good at our core business, which was connectivity.”
Stankey says he waited about five hours in line to get a Princess phone for his college dorm room, his first telephone.
“I had my very first telephone when I went to college, and I stood in line for about five hours to get a Princess phone to put in my dorm room.”
Stankey says his first wireless phone was an Oki car-mounted device with the unit in the trunk and the handset on the console.
“It's my first wireless phone, uh, was a, uh, it was an Oki car-mounted device that, uh, went in the trunk of the car and and sat on the console of the car.”
Stankey says that over the past couple of years countries have become more sensitive about foreign ownership of critical assets, which discouraged AT&T from expanding beyond North America.
“I see more of a balkanization going on, where many countries are becoming much more sensitive around foreign ownership of critical assets.”
Stankey says the United States delivers better competitive returns in telecommunications than other countries because of its policy approach.
“In telecommunications, the United States stands above every place else. They've gotten policy and approach to telecom right”
Stankey says parts of Europe have communications infrastructure he compares to the third world, though he says Norway may not be one of them.
“I think you have a developed, uh, set of economies that are basically on third world communications infrastructure in many instances.”
Stankey says he views all networks as fiber networks with different access technologies, such as cell towers or wall jacks, attached at the end.
“My belief is everything's a fiber network with different access technology hanging on the end of it.”
Stankey says the low-band TDD spectrum AT&T bought lets the company engineer more upstream capacity on its wireless network.
“the TDD spectrum that, uh, that low-band spectrum has allows you to engineer the upstream more aggressively.”
Stankey says AT&T often defends its networks against nation-state attackers who face no profit-and-loss constraints, so it collaborates with government and industry partners.
“We're a particular for-profit company that's oftentimes fighting nation-state actors that have no budget and they don't have a P&L they have to report on or a balance sheet they have to worry about.”
Stankey expects real end-user satellite connectivity services to begin appearing in 2026.
“I think you're going to see a lot in 2026, next year, I think you're going to start seeing real end user services coming about.”
Stankey says video accounts for over 80% of total traffic on AT&T's network, both downstream and upstream.
“Today, video in total, both downstream and upstream, drives over 80% of the total traffic on the network today.”
Stankey predicts 6G will be oriented toward enterprises rather than consumers, with better latency, security and device scalability.
“six G is going to be very enterprise-oriented as opposed to consumer-oriented.”
Stankey says the Trump administration did more work in about nine months on bringing licensed spectrum to market than was done in the previous four years.
“There's been probably more work done in, you know, a nine-month period of time than what was done in the previous four years.”
Stankey says AT&T uses AI to retune cell sites dynamically in real time to follow crowd demand, such as during a marathon.
“If a marathon is underway and and traffic, uh, people are snaking through a city with different, uh, demands at different parts of the city, we can tune the network to follow that demand”
Stankey says only a handful of companies, perhaps three or four, will build data centers at scale using their own chip designs, which is why AT&T does not invest in them.
“There aren't going to be eighteen people building data centers, there's going to be a handful, three, four, and they're all doing it on native hardware”
Stankey says available electrical power in the US will limit how many announced AI data center projects can be built.
“I think there's going to be a limiting factor, at least in the US, around how much power is available to turn it all up.”
Stankey says AT&T has about 130,000 employees and that getting them aligned is what makes culture change hard.
“somehow, you gotta get 130,000 people viewing something in a common fashion.”
Stankey says AT&T built an AI center of excellence whose experts act as internal consultants rather than spreading AI expertise thinly across the company.
“We've built a center of excellence where they're available, the expertise in AI is available to anybody in the company, and they work out as consultants in the company”
Stankey describes his transformation of AT&T, covering focus, cost cuts, fiber investment and customer-centric culture, as a five-year journey with a few years still to go.
“it's really been a bit of a five-year journey that has a few more years left to go before we get to where we need to be.”
Stankey says the internet took about 25 years to change business models and that AI will probably do the same in half the time.
“if we took us twenty five years to get to where we are on the dawn of the Internet and how it's changed our business models”
Stankey explains how wireless, fixed and satellite now compete, and why the content business was sold to focus on connectivity.
“you best be in a position to invest heavily to make that happen”
He explains why AT&T avoids foreign expansion and why he thinks European networks fall behind because of weak returns.
“Players aren't getting return on capital, therefore, they don't want to reinvest”
Stankey sets out why fiber and upstream capacity are central to AI, and how nation-state threats shape security.
“One of the things that AI is going to drive is more upstream traffic.”
He describes satellite as an add-on product and expects 6G to matter mostly for enterprises.
“I think you're going to see a lot in 2026, next year”
Stankey covers AI inside AT&T and explains why he avoids data centers and sees overheated investment.
“I think there's a little bit of irrational exuberance in some places.”
He lays out the tactics for building an AI culture and the challenge of changing it from within.
“changing culture is the hardest thing you do in a company, especially an established company.”
Stankey reflects on his leadership style, how he relaxes, and why people should treat learning as a lifelong discipline.
“you have to think about your career in chapters that are four or five years.”
Hi everybody, and welcome to In Good Company, and today we are live from New York. Now, the first thing I did when I arrived was to change my SIM card in my phone to an AT&T SIM card.
Thank you for that. Next question.
And so, therefore, it's kind of fun to be here with the CEO of the whole company, AT&T. It's good to be John, thank you.
with you, and thank you for changing your SIM card to an AT&T SIM card.
Now, AT&T American Telephone and Telegraph Company, you know, founded by Alexander Graham Bell, going back to, like, 1877. That's quite a history, right?
Just about on year one fifty s next year, so it's a big deal.
now, telecom industry has always been fascinating and interesting, but what's interesting, eh? for you in the industry right now?
I I think there's two things that are really important in the United States because of what came after the Telecom Act, which was a attempt to deregulate all the various forms of communication. We're now seeing technology actually achieve that. There's real intermodal competition between wireless and fixed and satellite, and as a result of that, there's a reordering of assets going on in the United States. And where we used to have companies that focus maybe in one technology branch or another, it's now just about can you deliver a network that gets customers on the internet?
And can you do it? Whether they need to be on wireless or a fixed connection or through satellite? And that reordering of assets is pretty interesting, cause I think it's gonna readjust the winners and losers in the market. And that's
But not, but but data traffic is exploding, right? We are using it more and more and more. Yet the telecom sector is kind of certainly lagging the, you know, the the mag seven and the big tech stocks. Why? why is that?
well, that's, uh, clearly the value add on the high margin products. When somebody can carve out, uh, a dominant market position, they tend to drive better margins than the Magnificent Seven have managed to do that in each of their segments. In our case, we're a pretty competitive industry, and as a result of that, we tend to grow at GDP or GDP, a little bit plus. And, um, our ability to kind of carve that same, uh, dominant position at large margins just hasn't arrived at this point.
Having said that, the next exciting part about the industry is the fact that AI is gonna drive another growth cycle in data and usage demand. And we think that's going to be a unique opportunity for growth in our business moving forward, and it's going to allow for that consistent, uh, steady improvement of cash flows. If you invest correctly during this change cycle that's occurring right So now.
so how should we look at the relationship between the explosive growth in data and the providers of the infrastructure?
Well, I think what, um, you know, for many years, people are saying, Are you going to be able to profitably grow as consumption continues to grow? And that's really been going on for decades. And I would say that the data points over the last, uh, number of years have shown that, despite the fact that consumers have an insatiable demand to use more and more data. By the virtue of technology, operating our business more effectively, the ability for us to add value into products and price differently, we are driving solid returns in our business.
And I do expect that's going to continue moving forward, especially as we see more convergence of customers buying services from one company across multiple networks. And your ability to be even more efficient and more effective at monetizing that household.
Now, if we go back some years to, um, 2016, you you bought WarnerMedia. At that stage, one thought, Okay, we need, uh, to combine content and distribution. And then, um, one realized that, uh, you know, that that wasn't a a particularly good idea, what? But when you look back on that, uh, at that, what? What was the learnings from that? In a way,
There's there's several, um, one is, I would tell you, uh, when you think about changing a traditional business model and having to reposition it, you best be in a position to invest heavily to make that happen. And and I think one of the learnings was media was going through its own, uh, repositioning and the reengineering that was necessary in the business model. And certainly, as we just discussed, telecommunications was seeing the same thing. And in today's competitive environment, if you're not investing aggressively to kinda win in a space, it's really hard to be successful.
And on anybody's, uh, public balance sheet, trying to reposition in two industries at the same time, uh, probably requires more cash and more patience than what the public markets are willing to offer.
you decided to divest the content business, uh, how did you come up with that?
It was it was necessary, it was, uh, in order for us to be really good at our core business, which was connectivity. Given what was going on in the industry that we just discussed, uh, reinvestment, repositioning, intermodal competition. We had to step up our investments, we had to get better at what we were doing. And I think today, one thing we know about customers is they have a low tolerance for less than a great product. And that's certainly something the internet has shown us all.
And as things have become much more dynamic and switching costs are lower, customers are gonna go where they get the best, best product and the best deal. And we had to reinvest in our business to make that happen.
How intolerant are customers now? Cause I don't know who into I, I know exactly how intolerant I am if I don't get in touch, you know, within whatever, just how, what's the what's, what's kind of the time lag that customers are accepting now?
Yeah, I think it depends on the product. I think our product, uh, you know, fortunately, is viewed as a staple in somebody's life. It's so critical to them, and, frankly, they invest a fair amount every month. So we're not the five dollar subscription, you know, we are the one where they have to give some very careful thought to having a relationship. So maybe they're a bit more tolerant of our our missteps if we demonstrate to them that we understand who they are and we give them a good product.
The vast majority of the time. But, uh, frankly, the you know, what used to be in this industry, where there's a lot of switching inertia, is not the case anymore. You started this out by talking about the fact that you showed up in town and you popped a new SIM card in your phone and you were on our network. And if you're upset about what happens tomorrow with our services, you may go get a SIM card from somebody else and pop it in your phone.
so when you when you have a telephone and you make a call, what, what is your acceptance? If you know, how long do you accept it to take before you're connected to the other end?
Well, in my case, I have very You? Okay, high standards, and if well, that's what I if
I that's what I assumed, it's if if but what is it's not happening it?
instantaneously, I'm on the phone with somebody. I was in a far part of New York this morning where I wasn't particularly happy with what I was seeing with some of my throughput characteristics. And I, I was back on the phone with somebody.
Now you've been, uh, at AT&T for forty years. What was your first telephone?
I had my very first telephone when I went to college, and I stood in line for about five hours to get a Princess phone to put in my dorm room. It's my first wireless phone, uh, was a, uh, it was an Oki car-mounted device that, uh, went in the trunk of the car and and sat on the console of the car. And, uh, I thought that was the best thing since sliced bread when it showed up in my life.
Well, it was right.
Yeah, it was.
Well, what a journey. Uh, do you think AT&T would be anything else but a pure communication utility?
Do I? I? First of all, I don't think about us as a utility, I think about us as a networking company, and I think we do some pretty sophisticated networking. It has to customize what we do for a lot of fairly complex customers, like large enterprises. And I do believe that we have an opportunity to move a little bit right and left of our core connectivity business with a combination of value-added services on top of those networks.
And, um, as we complete our migration into a software-driven network, our ability to begin building verticals and products that are software-driven on top of our network gets a lot better. And so those would be things like, can we begin to offer enhanced security services that are software-driven on top of the network? That allows us to move maybe laterally into security on top of connectivity? We have a lot of things. When you think about AI driving performance and workload, there are things that we can do to ensure that we're delivering quality of service at a bit level across different network fabrics.
That I think will be incremental opportunities for revenue that we can bring in. We're starting to work with players in the industry that allow us to do things like bring security and privacy services into the market. That allow parents to manage their relationship of with their children and how their children see content and have insights to those things. Those are all extensions that are built on top of the core network, and I do believe we can grow revenues and opportunities to grow the business a little bit faster as a result of that.
You've stayed in North America whilst you compete with some folks who are more kind of diversified globally. What are the what's the thinking behind staying here?
That's a really good question. I would tell you, at one point in my career, there was a desire to move beyond North America and did a lot of work around that. And I think that was at a time where maybe, um, in many countries, there was some flexibility around foreign ownership starting to to, uh, manifest itself. Um, I would say the last couple years that has reversed itself. I see more of a balkanization going on, where many countries are becoming much more sensitive around foreign ownership of critical assets.
Um, in addition, if you just kinda go around the world and say, where are competitive returns achieved? In telecommunications, the United States stands above every place else. They've gotten policy and approach to telecom right, and it's really hard to find another market to go into to say I can go and get as good a return in that particular country as I can in the United States. When, in fact, we have more organic growth opportunities in the US.
That's why we're investing so heavily in fiber and building our footprint and starting to build outside of some of the traditional markets that we've operated in because we think we can get better returns here.
Well, you are, it's, it's, it's a very big market, right? So you look at Europe, you have there's been a competition policy that you should have three players in each country, and you've got, like, uh, close to thirty countries, so that's, uh, that's a lot of operators.
Yeah, there's actually very few countries of scale in the in in Europe that have three players, many of them have four, moving to three. And, uh, you also have some dynamics there that are a little bit different on on how roaming gets done and two layers of regulation between the country and the EU.
So when you sit here, you know, we are in New York and you look to Europe, what are your what are your reflections when you look at the European telecom market?
I I think it's sad, uh, candidly. Uh, I think you have a developed, uh, set of economies that are basically on third world communications infrastructure in many instances. Maybe not the case in Norway, but there are parts of Europe where services are are clearly a step behind much of the rest of the developed world right now. My sense is in talking to policymakers in Europe, they they get that, and that's why there is this move to maybe go to a three-player structure that incents more investment in infrastructure.
And I believe if they move forward on that and they do it quickly, that you can, in fact, see a reinvestment cycle that will improve services in Europe, uh, pretty significantly.
So when you say you think it's, uh, it's not so great, that's because of the technical offering, or is it because of profitability of the players? Or
Players aren't getting return on capital, therefore, they don't want to reinvest in capital in their infrastructure and networks.
how how we? How do we see that European players are behind the US. players?
I would just go look at what, uh, customers get in terms of data quality, throughput and the advanced services, what's available in terms of, you know, is state-of-the-art five G and are is the performance of those networks what you would see in Asia or in the US. Inconsistently, as I move through Europe, that's not the case.
Tell me about your fiber build outs. Why? why is that so good? Look, there's Why is it better than putting up, uh, mobile towers?
fiber. When you think about running networks, I don't like to think about wireless or fixed. My belief is everything's a fiber network with different access technology hanging on the end of it. Sometimes it's a cell tower, as you just said, sometimes it's a jack on the wall that something plugs directly into, like a wireless router. Sometimes it's a hardwire connection into a hyperscaler's environment, but it all starts with fiber. And in particular, why fiber is so critical moving forward in the AI world, is most high bandwidth networks today, especially for the consumer space, have been built asymmetrical, more downstream bandwidth than upstream bandwidth.
One of the things that AI is going to drive is more upstream traffic. And the requirements to move into a hyperscaler's environment, to get to large language models, to do the processing of data and information in an AI world, is going to require a lot more upstream bandwidth, so more symmetrical networks. Fiber is symmetrical, and so the faster you get a bit of traffic onto fiber, the more capacity you have, not only down, but up, and the more ability you have to manage the performance of that bit on so when the network.
you look at five G and six G, you don't think that's a competition to fiber, you think it's just a feed into your fiber.
Yeah, I don't think it's competition. I think it's just another access technology, where in some cases, five G and six G are the best way in a wide area network to get somebody on a high bandwidth network. But there are times where other access technologies like Wi-Fi or being actually hardwired into the network are necessary, depending on the performance that's required.
You bought some, um, spectrum frequencies, uh, lately. How is that going to benefit you?
Well, it's going right to what we just talked about, where if you believe workloads are going to increase, these are going to dramatically improve our throughput on our wireless network. That will enable some of these applications to pop up. And, in particular, some of the spectrum we bought was low-band spectrum. That the way it can be engineered is effectively making sure there's a more robust upstream channel into the network. Um, so the difference between the two types of technology that's used on spectrum, TDD and FDD, the TDD spectrum that, uh, that low-band spectrum has allows you to engineer the upstream more aggressively.
And we think that plays very well into what the future of AI is going to require of mobile networks.
What about cyber? Is cyber cost or revenue opportunity? Is it good or bad news?
It's a it's a thing that keeps me up at night. There's not many things that keep me up at night. Um, so I I think the notion of building secure networks right now and doing it in the geopolitical environment we're operating in is a real challenge. And we spend a significant amount of money within our business to make sure we offer secure networks. Having said that, my f- my concerns about running a large public network are shared by every CEO and every CIO at every company, uh, around the globe.
And so there is an opportunity for us, as I said earlier, to extend our basic connectivity services to help folks manage their networks more effectively. Since we do things like monitoring our network, looking for patterns and changes in behavior, and and trying to hunt the bad guys actively, day in and day out, we can extend that onto some of our customers networks. And that's where us extending into other products and services that are software-based, I think, opens up an opportunity for us to grow new revenues as well.
So I see it as both. It's a cost because we've got to make sure we run our business effectively, but that cost is something that every business out there is enduring. And we can help them do it more efficiently and probably get remunerated for it.
So hunting the bad guys, what does that involve? What all the what all the weapons?
Depends on the day of the week. Uh, and you know, it's it's a little bit of, uh, an imbalanced fight, right? We're, uh, we're a particular for-profit company that's oftentimes fighting nation-state actors that have no budget and they don't have a P&L they have to report on or a balance sheet they have to worry about. And they'll invest, you know, untold amounts of money and creativity to try to figure out how they, uh, infiltrate and beat that down.
So it requires a tremendous amount of collaboration, requires collaboration not only within our company, but with government entities, our partners in the industry, where we're sharing our insights and knowledge. I think it's only through that collaboration that you can counter somebody who's going to spend literally billions and billions of dollars. Because it's in a nation-state interest to figure out how to crack the code on on, uh, compromising you. And, uh, if you don't have everybody helping each other, um, there you that are aligned in a common way, then you're gonna have a very difficult time.
Second thing is, it requires you to really work out with those that are on the leading edge on research and development. And there are, especially in the United States, some companies and some entities that are really, really good at this kind of thing. And staying in tune with them to understand what's next, what's coming around the corner, what the next type of vectors of attack are going to be, and how do you think about getting ahead of the curve on those things is pretty critical as well.
How do, um, satellites fit into the competitive landscape here?
Satellites, for the most part, I think, are, you know, they're first of all, it's going to be amazingly encouraging to see the innovation that Leo constellations bring. I think they're going to change a lot of things in our lives, and, uh, I'm excited about that innovation that's going to come forward. I think we're going to see truly the dawn of always-on connectivity, and that's going to be good for our business. I I think we're in a position, since we already have a preferred relationship with customers, to, uh, manage their connectivity.
We're a natural place to integrate satellite as an additional add-on to their connectivity relationships. So
So when, when will we see this kicking in?
I think you're going to see a lot in 2026, next year, I think you're going to start seeing real end user services coming about. And, uh,
So how am I going? how am I going to benefit from that?
well, you should expect, if you're a customer of AT&T, that you'll have an option to, for example, if you're g- uh, at a day where you're gonna drive off the network, uh, maybe into the Grand Canyon to go rafting. And you wanna stay connected, that you'll be able to buy a day pass to stay on the network as a result of that. Or if you're somebody who has a lifestyle that requires you to always have that kind of backup, maybe you boat on the weekends and you go offshore.
Um, maybe you have a job that requires you to be in rural places for oil drilling, you can buy a subscription and have it all the time. Uh, but I think satellite is more complementary than directly competitive. There'll be some overlap, Internet of Things, for example. In some cases, LEO will be a better way to handle IoT-type arrangements that may supplant what a terrestrial wireless network used to do. But in many instances, it's going to be complementary and add on to what are always already really robust, high bandwidth networks in the United States.
John, let's spend a bit more time on on AI. How do you think that will increase data volumes?
I think there's a couple things that will drive it. The one that I'm probably most excited about is if you kind of look around your house or your car or your workplace, there's sensors everywhere today, and oftentimes those sensors are cameras. Um, certainly, one of the things that drove volume on the network in the past was when we put phones, we put cameras on every phone. And it was all the content that people picked up on either video or taking pictures and wanting to ship it around.
That drove a lot of traffic on the network. When you start to think about all the sensors that are capturing video right now, and you think about how capable AI is in processing video and taking insights from it, I think that's one of the huge opportunities. So when we talk about the importance of upstream bandwidth, this is one of the reasons I believe we should be investing in that upstream bandwidth. Is because if all those sensors are out there picking up the video, it's got to get someplace to be processed and analyzed.
And that upstream capability to get it back up into the network, get it back up into the hyperscaler LLM to be processed is, I think, one of the places we're gonna see large upticks in traffic. And then ultimately, once it's a conclusion or an inference is driven from it, that has to be delivered back to the point of consumption quickly for somebody to make a decision or do something with it. I think that puts a tremendous amount of, uh, lean on high-performing networks, and that's really what we're attempting to build in this
Stupid question, but how much of the traffic through your networks would now be pictures or video?
in terms of the actual percentage. Today, video in total, both downstream and upstream, drives over 80% of the total traffic on the network today. And much of that's entertainment-based video. Uh, when you think about kind of in the wireless space, the four G revolution was really geared off of consumers consuming video for pleasure and entertainment. And that's what drove massive amounts of workloads onto the networks. I think in the six G and AI world, early on, we're probably gonna see enterprise applications drive more of that increase in demand on the network, but video's still gonna be a really important part.
What's the biggest, eh, uh, change from five G to six G?
You know, I think what you're going to see is that, uh, six G is going to be very enterprise-oriented as opposed to consumer-oriented. It's going to allow for the kind of management of networks in the in the types of round-trip latency and performance that are necessary in an AI world. I think it's going to add in, um, additional layers of security that have not typically been available in wireless networks that, um, are well-tuned to the threats that we see today.
And I think it will be scalable for the number of devices that are going to be connected. Um, we're
But are are we going to feel any different?
uh, as a consumer, I don't think you're going to necessarily feel the network per se being different, but the applications that you'll be able to operate and and achieve. If you're, for example, an enterprise using a wireless network, those applications will be different and better.
Can you deliver that increasing capacity?
Can we deliver it? Yes, I think we can. I think the United States, in particular, was running a little bit behind on getting the right amount of licensed spectrum available to operators early on in the Trump administration. There's been probably more work done in, you know, a nine-month period of time than what was done in the previous four years. To have a roadmap of how to bring more of that spectrum to market to ensure that it those workloads can be handled.
I feel much better about it today, based on the actions that have been taken and the path to, uh, spectrum auctions that are going to be here, that the US. can keep up with other key players around the globe. I wouldn't have said that probably eighteen months ago. Uh, so I feel much more confident we'll be in a better position.
When you, um, you have forward visibility into a lot of the things that's happening, right? Because you're a big corporation partner with a lot of these tech companies. What's, uh, when you look at the AI applications going forward, what's what are you most excited about?
Well, there's, uh, there's things we can do inside our business that I get really excited about that we're already proving in. Um, we've done a lot in customer sales, service and support that has, um, dramatically improved our customer service, lowered our operating costs, helping us reduce churn in places. Um, those kind of meat and potato things, I think most corporations are gonna have an opportunity to chase. I see some great opportunities for us in how we do pricing.
Um, when you think about what AI models can do and how they inform what's going on competitively in a particular area, much more dynamic models of pricing as compared to what have been more static models that we've used. So I'm really excited about that, how we engineer networks. Um, we were just I was going through a set of projects earlier this week about our ability now to literally look at how traffic is moving in any given moment in a city.
If a marathon is underway and and traffic, uh, people are snaking through a city with different, uh, demands at different parts of the city, we can tune the network to follow that demand and do it dynamically in real time. Uh, and AI is driving a lot of that, and that changes our engineering practices and our deployment practices.
So what do you do with the networks? So let's say now you have the marathon, right? You got, uh, 50,000 people running past there. What do you? So what do? What do you do with the networks?
So AI is knows exactly the route of the marathon. It takes it, looks and takes, uh, pictures of how many handsets or any particular places it evaluates where they're moving. It knows the route, it knows how it needs to tilt and move cell sites and redirect them. So when the loads get someplace, the cell towers are already tuned and adjusted to where those loads are going to be. And it's doing it dynamically, literally following things, and that's those kinds of things I think are gonna help service levels drmen- tremendously.
We used to do some of that stuff, it was labor intensive, we used to have people sitting around at keyboards adjusting. Now, if it dynamically occurs, we're more efficient at getting it done.
Have you seen any, uh, new cool devices? You know, I mean, you got the AI glasses, you got the well, openai is talking about some kind of new device. What's, uh, what are the kind of things you've seen?
I've seen the second generation of lenses, if you wanna call em glasses, that now start to have some utility. And you start to look at what they're doing, and they're maybe not perfect right now, but you can really start to see where, as an augment to your daily life, they can add some value. And I think, uh, another iteration of those, uh, could very clearly start to change some things. Form factor still needs to be worked on a little bit, ease of use still needs to be looked at.
But people who are conversant in using the technology can actually do some pretty powerful things with them. And there's an example of if, you know, some kind of a set of glasses or lenses or a supplement for you as a screen, uh, has to be connected. That's good for our business, right?
Yeah, absolutely. I have to say, when they first came out, or when the first generation came out, I was quite skeptical, but now I think it'd be pretty cool.
They're getting there and they augmented, you know, if it's smart enough, and and AI is smart enough now to listen, take context from speech, where you're not having to manipulate a user interface or give it direction. But it can intuitively know what you're attempting to do at any given point in time in a conversation. I do think there can be some very powerful applications.
Why are you not investing in data centers?
Look, we're if you think about anything in tech, scale matters, uh, if you build networks, scale matters, uh, putting fiber out, scale matters, building data centers. And when you think about how that's being done today, which is chip design, uh, it's your ability to, um, have intellectual property that allows you to network in that structure very, very differently. Um, and it's scale. There aren't going to be eighteen people building data centers, there's going to be a handful, three, four, and they're all doing it on native hardware, they're doing it on native chip design.
That's a business that, uh, requires a tremendous amount of investment and requires a lot of investment cycles to stay current. I have, uh, my hands full just making sure that the fabric that connects all those data centers is state-of-the-art, scaled and working properly.
Do you think data centers will be a good long-term investment for the people who do them?
Um, I think there's gonna be winners in the AI space. I think there probably are going to be fewer winners than the number that are investing in it right now. There's going to be some carnage.
Is there a bubble in that now?
Uh, yeah, I think there's, I think there's a little bit of irrational exuberance in some places. Um, certainly, when you look, you know, there's, there's the announcements, and then there's the reality. The announcements are pretty staggering. Um, those numbers,
so you don't, you don't think those will come through?
um, I think there's going to be a limiting factor, at least in the US, around how much power is available to turn it all up. And I think that's going to become a natural gating factor. And I. It's going to be very hard to achieve all the announcements that have been made unless we get on a different growth curve for the power infrastructure in this country.
But with the speed of, uh, technology development, you need to depreciate these assets over a relatively short timeframe, right?
I would think, um, and uh, that that's my view of when you look at the capital intensity of building those data centers, and you look at the capital intensity of my business, I'm I'm perfectly happy staying in one domain.
How do you, um, build an AI culture within AT&T? How do you get people to experiment and adopt?
Um, you know, I will, um, I'll weigh in on that three years from now, maybe after we get through it. But it's changing culture is the hardest thing you do in a company, especially an established company.
Why is it so difficult?
They're people, and I think, uh, you know, people generally have their own motivations, and they have, uh, it's what's in it for me and what's in my best interest, and how do I process? And this? And somehow, you gotta get 130,000 people viewing something in a common fashion. And, um, that's that's a hard thing to do. What what we're doing is trying a couple different approaches. We're using incentives, uh, and and giving people motivation to wanna embrace it and use it where it's in their self-interest to do that because of how they get paid or the objectives they have in front of them.
We're using structures where we're separating work groups from, uh, kind of the BAU construct, our business as usual. We'll pull people out and say, you're gonna run a site operation, you're going to be measured differently, you're going to try to make this better by inventing the new. And you're not encumbered by what we're doing today, and we're testing proc- processes that way. We've built a center of excellence where they're available, the expertise in AI is available to anybody in the company, and they work out as consultants in the company to adopt projects that people wanna carry forward.
Where we're not trying to build that expertise in too many places where it gets diluted, and we're finding partners to bring into the company that want to build the next platform that's an AI-based platform and need really hard test cases to do that. So they come into AT&T and say, let me build you the finance department of the future, and we'll do that collaboratively. We'll do that gratis with you, as long as the product that we built, we can go out and then use in other companies and and sell it further.
And so we're using a variety of different ways Which to get at it.
which part of the AT&T culture are you trying to shift or move?
You know, we're we have a lot of things that are really good in our culture. We don't, you don't stay for 150 years without doing something right. And we have a lot of employees that are incredibly committed to the mission of the business, which is to keep people connected. Uh, and we've we do that, uh, in ways that are, in fact, miraculous on some days, and that's that's a really important fabric of what we do.
We need to be a bit more agile, and we need to be a bit more market focused, and we need to make sure that everybody who works at our company is there. Because they're committed to the mission, they're not there because they feel like there's a blind sense of loyalty to a pension plan or or some other construct that, um, I f- I wanna work thirty years because there's something there for me to work thirty years.
It's I really wanna be here because I believe in the mission of the company.
But I mean, you've been there for nearly forty years, so you kind of you epitomize the AT&T culture. I mean, you are Mr. AT&T, right?
I, I, I don't. I've never thought of myself as Mr. AT&T, but I I think sometimes
But I mean, you brought up with the culture, you, but h- So how can you change a culture which you are part of in I a way?
think you can go and look at research, and some would tell you that if you're going to evolve and change a culture, sometimes it's the disruptor from within that can be the most effective as opposed to the disruptor from outside. And really, that's the test that's in front of me right now. Can I build a more market-based culture that's more focused on the customer and the customer need? And it works from the market back into the company, as opposed to one that has been really built on.
We'll tell consumers what they need, we'll tell consumers what the next round of technology is, and then we'll take that out to the market. As opposed to really understanding what the needs of the market are and then engineering the company to match those moving forward.
So when you took over as CEO, what were your priorities? What did you wanna change?
Well, first of all, it was obviously to get the company much more focused, which we talked about around making some decisions on the asset base and divesting. It was to make it a lot more cost effective, where we generated the kind of cash that we could be leading our industry in investment. That meant we had to get a lot of cost out of the business, we had to get out of a lot of the legacy products and services we have, and we've been aggressively moving through that.
We're not done, we've got another couple years of work to do there, but that's helped us get cash flows improved where we can reinvest at the levels we're doing. And doing that is to be the leader in building fiber and having the most robust and capable network infrastructure. And so we've invested at the top of the industry and then working this cultural change. That is a change in us being much more customer-centric and customer-focused and elevating our game in support of customers, including guaranteeing what kind of service we'll provide to our customers as the other aspect of differentiating our services moving forward.
And all those things have to come together at the same time. And it's really been a bit of a five-year journey that has a few more years left to go before we get to where we need to be.
How long time does it take to change corporate culture?
I think it's, I think it's never ending. I I think you, you never get to a place where you say, I can stay here. And even if you have a productive and successful company, you know, it's gonna have to evolve and change a little bit. Now, maybe the rate and pace that we have to achieve right now is higher than the average, it's maybe above average, and what we need to change in the period of time.
But if you ever say that I've got it right and I don't need to worry about it ever evolving again, I think that's probably a a recipe for extinction.
But don't you think the speed of required change is just going to accelerate going forward?
Could. Yeah, there's no doubt the level of change we're all seeing. If you think that AI, for example, if if we took us twenty five years to get to where we are on the dawn of the Internet and how it's changed our business models and how we operate, AI will probably do the same in half the amount of time. So I think that will require adjustment and change to everybody's personal work habits, how how work is arranged, how you think about attacking markets.
So I do believe it'll get faster and faster.
How would people characterize your leadership style?
I think what they would probably say is, I'm pretty direct, I'm pretty clinical. They would, people. I've had one boss who said, You're a bit like a fine wine. You want to buy it, hold it, and then open it and appreciate it for its complexity and what it's able to do, but it takes a little bit of time to get there. I think people see over time that, um, I can be a pretty reliable partner, um, I can be a trustworthy partner.
I can be somebody who, uh, is incredibly not only committed to the company, but committed to the individuals who wanna see themselves step up and raise the bar on their contributions and performance to the business. And if that,
Do you think you are? Do you? Are? Do you think you are respected more than loved?
um, I, I don't know that people use the term loved to describe me. I think, uh, generally speaking, my goal is to be respected. I think any CEO today, because of some of the things we're talking about, the rate and pace of change, the challenges that are in front of, especially established business models. There's a lot of hard decisions to be made,
Life life is not a popularity contest.
and those hard decisions are usually have some collateral damage in places. And over time, hopefully, investors, customers, employees on par, look at those things and say, I respect what was done and it was necessary. But you don't always know that at the moment when those decisions are being made. That's sometimes perspective that takes a few years to to develop.
Have you become a better leader with age?
Um, in ways, I've become a better leader with age. Certainly, seasoning is important. Um, I would tell you, I think anybody who doesn't understand that these jobs have a physical demand, that's necessary, uh, that physical demand. You can do a lot more when you're fifty than you can when you're sixty, even if you are serious about keeping yourself in good shape and you pay attention Are you to these
in good shape? things. How old are you?
I'm sixty two.
Yeah, well, you, uh, you come across as very young and, uh, you know, uh, I was agile,
even better when I was fifty.
do you think? It's kind of funny? Because, um, a lot of the people we interview, they they say they're better leaders, uh, you know, with age. So you have to start to think, uh, or wonder, you know, how many bad leaders there are when they're, you know, all these young people.
Yeah, I I think it
Well, everybody claims they get better.
it's situational, and I I think there's a balance. Because I also believe when you come into a job like this, the day you enter it, you begin distancing yourself from the reality of the business because of the demands of it, the time you have to spend outside the company, what you have to do with other stakeholders. So you're never gonna be as good as the day you walked into it in terms of your intimacy with the company.
And it's I'm not really a big believer in the imperial CEO model for that reason. Because I think, uh, getting somebody who's more, uh, closely aligned to what's going on in the business, um, has its merits. And I also think companies are living, breathing organisms. If there's one person at the top for a long period of time, people develop their big muscles to respond to that one person, and they let all the smaller muscles around them kind of atrophy.
And sometimes that change is what drives you to balance out the development of all your muscles and all your skills, and that's a healthy thing for a large organization.
do you trust your gut feel or are you very data driven?
It's both. I'm I use, uh, I have some guts that I usually use for hypotheses. Typically, I try to validate those hypotheses by bringing data to, um, determine whether or not I was right about them. You know, is the intuition correct? And it's a it's kind of a catchball process with it. But I think part of being a good leader is being able to evaluate what's going on around you and see unrelated but associated signals. And then test those hypotheses, go out, check them, bring data to them, and then act upon them.
And that's clearly an element of kind of my my style. I would say I try to be a little bit provocative at times.
I heard this thing this morning, uh, in God we trust, and the rest of you need to bring some data.
Yeah, 100%. It's trust and verify is another way to to look at it. And I, I get asked this a lot, Do you trust us? and I say, Of course, I trust you. But part of my job is to verify that there isn't a better way to do something. Or that that we're, you know, actually representing data in a clinical fashion, that we know exactly what's going on in our operation.
Do you enjoy pressure?
Um, I don't necessarily thrive on it. I don't go out and say, Gee, I would. I hope my day is really stressful today and it's all going to come. But I I think I understand it's an element and part of my job in any given day. Um, I do believe that if it's applied in the right fashion, it can be a motivating force for teams and for organizations. And so sometimes creating constructive pressure, uh, is a necessary tool to get an organization to move.
How do you relax?
I, uh, I have three grandchildren that are
Do you do you laugh because you don't relax so much?
I don't know. Or, uh, Some people probably debate. I have. I have three grandchildren that are fantastic. I'm in a different place when I watch them because you get the opportunity to see life through the most simplest of eyes. It doesn't take a lot to entertain them and keep them engaged. Uh, I love to play golf, I'm not very good at it, but I love to play golf. And time on the golf course, especially when I'm playing with some of my family members, is is a big decompression time.
I love to go to the mountains and be outdoors, whether I'm boating or hiking or skiing, that's a big release for me. Uh, and
why do you think? then, why do you think nature is soothing and relaxing?
um, because we're we're not, uh, my life, I'm not in nature day in and day out. I'm in, you know, I'm in buildings and I'm in meetings, and I'm in metropolitan areas, and I'm flying in and out of airports. And when you get into an environment where you're smelling pine trees and you're looking at water and it's not developed around you, I think it just resets, um, kind of the environment you're in. It's it stimulates the body differently.
What do you read?
Uh, I, there's a stack of books on the side of my bed that's about that high, and they come in. I I think three come into the office a week. And part of the art is to figure out of the three that come in, which one is the one that you should really hang on to and read. So, um, I'm very careful about what comes in. I I probably orient a little less to business books unless it's a emerging area of technology that's helpful to me.
Um, and I take recommendations from people on life stories, either inspiring leaders or autobiographies or kind of, uh, retrospectives on somebody's great accomplishments that I take a lot of satisfaction in. And then,
Which historical figure do you admire?
uh, I, I'm a big, uh, I like Abraham Lincoln. I think he he was in a
You have to say that as an American, though.
No, you don't have to, but I think you. What I like about it as a CEO is he was in a very lonely period of time for him to make some really hard decisions that I don't know that he could necessarily go and be confident that those decisions were gonna be have a certain outcome. That were could have been incredibly cataclysmic, depending on which way they went. Uh, and reading about how he went through that and and the process he went through, I find interesting and reassuring.
He hi, he hires some of his competitors, right? Are, are you? Are you doing that?
we we try to find people from all parts of the industry to come in, and we're doing a lot more of that as a company, yes, than what we were doing ten years ago.
What kind of music do you play?
Uh, I listen to everything. I I now. I never listened to country music before I moved to Texas, and my kids are big country music aficionados. So, um,
And is and is Dallas a good, uh, country music place?
the the it's good country music place, and I listen, I still Up there, up listen to a lot there with, of my
up there with Austin, and
uh-huh. I still listen to a lot of album-oriented rock from my youth. I listen to pop music today. I, when I need to calm down a little bit, I'll play the. I'll turn the piano on in the house and listen to classical music on our player piano, which when I retire, I hope to be able to play myself. But right now, I have to let it play for me. Um, so I, I'll listen to darn near anything,
sounds good, John. Last question what? Uh, what is your advice to young people?
especially now. My advice to young people is you need to think deliberately about what your process is to learn. Um, you know, this the notion that we can go to a university on a curriculum that's developed over five or six years and have relevant learning in a classroom environment, I think, is fading pretty quickly. Um, when you think about the resources that are available to all of us, including young individuals that are at our fingertips every day that we can go to school on.
And you think about how fast technology's moving, how fast business models are moving, as we've discussed, you have to think about your career in chapters that are four or five years. And every four or five years, you're gonna have to build a a new gig, a new foundation, a new set of skills, a new approach. And the only way you're gonna be able to do that over the course of a life that may be eighty or ninety years is if you are really, really good about being the dean of your own education.
And having a process, a discipline under which every day in your life, you think about what those learning moments are and the routines that you use to learn and carry forward on new skills. And if you develop those capabilities early in your life, in your career, they will serve you for your entire life and career. We have a unlimited amount of information available to us today. The winner is the person who figures out how to digest it.
And AI is only up the stakes on that now because it's allowed us to not only synthesize more information quickly, but it's always available to us to learn from. And people who master that are gonna probably be the ones who come out on top over time.
Absolutely well. We look forward to seeing you mastering the piano in your next gig.
Yeah, you won't be pay playing, paying any ticket prices to come see our And in the meantime, grandkids.
thanks for keeping me connected while in New York.
Yeah, thank you, Nicolai.
Thank you, great.
It's good to be with you today.