T-Mobile US’s CEO Mike Sievert on the Future of Connectivity

(Each year, TIME recognizes 100 companies that are revolutionizing the world. Apply now to have your company reviewed by our editors.)

At the end of 2012, T-Mobile US was not in a good place. With a market cap of less than $6 billion, it was in distant fourth place relative to competitors like AT&T and Verizon. The company’s trajectory changed in 2013, when it rebranded as the “un-carrier,” seeking to distinguish itself from competitors by offering simplified pricing structures, more flexible contracts, and a host of other customer perks. Thanks to these and other shrewd moves—notably its 2020 merger with Sprint, which gave it access to mid-band spectrum, crucial for 5G networks—today, T-Mobile has a market cap of over $200 billion and more than 100 million customers, making it one of the largest and most valuable mobile carriers in the world.

Despite this success, the company strives to retain the hunger—and the mindset of “punching up at giants”—that enabled its rapid growth in the first place, T-Mobile US CEO and president Mike Sievert tells TIME. Sievert took the helm in 2020, having previously served as COO since 2015. “We have a bigger opportunity in front of us than we’ve ever had, because we are a technology leader who’s transformed from a telecommunications company to a tech company, and that gives us a new canvas on which to serve customers,” he says.

Through partnerships with leading AI companies like OpenAI and Nvidia, T-Mobile, which TIME listed as one of 2024’s 100 Most Influential Companies, is looking to define the next decade of connectivity. TIME spoke with Sievert in November about his vision for the company.

This interview has been condensed and edited for clarity.

When you interviewed Nvidia CEO Jensen Huang earlier this year, you spoke about inventing the future together. What kind of future are you trying to invent?

Given the level of influence that T-Mobile now has, we have a huge opportunity to craft the future of connectivity. 

It’s amazing to think about how important connectivity has become in our lives. Last week, we had a big storm here in Seattle, and we lost our power in our house, but we didn’t lose our connectivity. The T-Mobile site nearest to us lost its power, but T-Mobile has this self-healing network, and it was able to reach us from a further away tower. 

I came to realize, if you made me choose between my power and heat or my connectivity, I’d choose connectivity. I felt so good being able to make sure my kids were OK, my mom was OK, even though our lights were off. It underscores the essential nature of the service that we provide. That’s why we’re working with the world’s best thinkers in the space to define what connectivity looks like a decade from now. 

Not surprisingly, it will be powered by AI technology: networks will optimize themselves in real-time to get you the very best possible signal, ensuring you’re connected everywhere you go.

With Jensen, we’re creating a future that we call AI-RAN [Radio Access Networks]. Right now, we have compute power at every one of our tens of thousands of network nodes all over the country, and we want to virtualize that into data centers—60 different data centers near the edge of our network—so we get the efficiencies of being able to create a virtualized data center that puts all that compute power in one place.

Network demands vary throughout the day. If we can use excess computing capacity to run AI workloads when network demands are lower, this opens up new potential for us to make the network better for users.

What does it mean to virtualize the network into data centers? How does this make things better for users?

In our core business, AI-RAN has the potential to allow the network to self-correct in real time, which is going to be more important as payloads rise. AI will radically increase the demands on our network, so it’s going to become increasingly important that we stay ahead of that rapidly rising demand.

For example, let’s say, in the future, you’re in a spot uploading high-def video to an AI cloud, to help you process the world around you, and a big truck moves between you and the tower you’re communicating with. You’re in motion, and the world’s in motion. The network will reconfigure itself to keep that stream uploading.

We believe AI workloads will create an important role for edge clouds. Right now you have two kinds of workloads: you have the massive cloud workloads, which are very helpful where massive compute is needed. They’re giving you textual responses, and sometimes photos and videos, but they’re mostly taking in simple voice or text inputs. There’s a lot of latency to that as it comes down from a central cloud. Secondly, there’s AI that runs locally on your device. And these are both very different from each other.

We think there will be a “Goldilocks demand” as, for example, LLMs will emerge to be much more about video and photos and audio going in both directions. To process what it’s seeing in real time, it needs to have the best of both worlds: a cloud-like compute capability, while remaining very close to the end user. So we think this will create an opportunity for edge clouds.

Now, we don’t have a business model around this yet. I rolled out a multi-year plan for the capital markets day that was financially exciting, and it includes zero for this. But we see it as a big opportunity, one of many now that T Mobile has a massive footprint and technology capabilities we didn’t have just a few years ago.

We’re making a bet that AI applications in general, as they become more immersive, will benefit from a Goldilocks architecture like I’m describing. That might be right, or it might be wrong. The great news for our financial picture is it doesn’t matter—we’re going to build it anyway, because it makes our core business so much better.

You’ve partnered with OpenAI to develop “IntentCX”, a platform that uses AI to figure out what customers need and act on it in real-time, which is set to launch in 2025. What do you think is the ideal customer experience?

Part of it is, we want to remove pain points. Being a great provider of wireless service means you shouldn’t have to interact with us much. We should prevent problems in the first place.

In 2023, we had the best year ever in terms of “churn”: people leaving T-Mobile. But because we’re so big, 8 million people left us. Think about that: 8 million people! I’m a pilot, so I’ll use a pilot metaphor: every one of them fell down an “accident chain”. With every one of them, we did something wrong, which wasn’t enough for them to quit, and then we did something else wrong, and then something else, until finally they threw up their hands and left. Every one of them left a breadcrumb trail of data behind on what exactly went wrong, and what we could have done to prevent it.

In the past, we’ve done nothing with this data. We’ve never even looked at it. But we’ve spent a lot of the last two years reorienting our data estate so that for the first time, AI can be trained to look at all this—network data, billing data, interaction and experience data, and other data about you—at one time, so that we can see the breadcrumbs.

We will have advanced bots that blow your mind and do great things. But it’s not about putting bots between us and our customers. We see this technology and ask ourselves: could we create an experience for our customers that’s so great that they love us more?

T-Mobile recently unveiled a plan to evolve from a “challenger” to a “champion.” What does that entail? 

Managing a team of people who have created so much success together over the last decade, one of my biggest leadership challenges is: how do you keep everybody hungry, humble, unsatisfied, and striving for the future? How we do that has changed.

Little old T-Mobile, when we were a fraction of the size of our principal competitors, created a culture of punching up at giants on behalf of customers, and created a turnaround that we called the “un-carrier”, because we’re unlike those giant companies. And we always told ourselves, one day, we’ll be as big or bigger than those other guys, but we must never become them: companies that basically hate customers, take them for granted, are motivated by old-school ideologies, and aren’t risk taking or interesting. How do we keep all that in a time when we’re successful? 

That’s what “challenger to champions” is about. Instead of just trying to right wrongs created by other companies, can we create new norms that nobody ever thought of before? Things like: could we be the first to provide direct-to-cell service, working with SpaceX, so that we can have the end of dead zones and your mobile phone could be directly connected to a custom low Earth orbit satellite system? We announced that intention two years ago, and we’re about to launch it. Another example is “T-Priority”: using advanced 5G technology to provide first responders across the nation with the world’s only dedicated 5G slice for first responders. What’s required of us in this challenger to champion era is to pioneer and innovate things that other companies couldn’t do.

In a recent earnings call, you mentioned T-Mobile’s “secret sauce.” How would you describe that?

Part of what has defined us over the last decade is that we do what we say, and in our industry, we’re the only ones. If you were to watch our Capital Markets Day in 2021, and then you watched everybody else’s from around that time, we’re the only company still working on the things we said. Investors and the public and customers love this about us. We’re reliable, and that takes a company that can execute.

You also have to be able to look around corners. And maybe we were lucky, maybe we were good. But a lot of our success has been that we called the ball on 5G. The whole world, except us and Sprint, thought that 5G would be about new gadgets, not your smartphone, and that they would be connected by a technology based on millimeter wave spectrum. As recently as 2018, everybody was building assets for that future.

And we looked at it, maybe because of our “love our customers” value, and said that’s not what customers want. They want their smartphone to work everywhere they go, at least 10 times faster and better than it does today. And the technology to get us there is mid-band technology in a multi layer 5G network. So we went after Sprint, which had this treasure trove of mid-band spectrum, but a stressed-out business model that wasn’t able to take advantage of it. Arguably, that’s been the most successful merger in telecommunications history, because it created this company.

We were right about what 5G would be. Everybody else has seen that and is scrambling to catch up, but they’re a few years behind. It’s important to try to see around corners again—we’re determined to get the next era right as well.

Does anything about the future of AI make you apprehensive? 

I’m worried about not moving fast enough. These technologies are rapidly changing how we engage with the world of computing and with each other. Companies that are able to see that unfolding, make some bets, and intercept it, are going to be the winners. One question people have about AI right now, at least financially, is “is it all a bubble? Will it turn out AI is better at poetry and humorous photos than creating value?” My belief is that no, this is no AI bubble, but it’ll be up to companies like T-Mobile to be customers of it, and to demonstrate at scale how it can transform our relationship with customers.

If the T-Mobiles of the world don’t show giant breakthroughs from the technology, then the benefits will be delayed. So it’s up to people like us. That’s one of the reasons companies like OpenAI and Nvidia want to partner with us. They see that we’re ahead, and they’re as interested as we are in demonstrating real value for end users that improves their lives, and creates financial value for everybody involved.

We can already see the potential. After two years of working on it, one of the aspirations that I unveiled in September was to reduce person-to-person problem-solving and customer service interactions by 75% within a three-year period.

And that’s going to happen—we’re well on our way. We have a detailed quarter-by-quarter plan that is absolutely on track, and a lot of that value is already dropping in. That’s one of many metrics that I unveiled. We’re demonstrating to ourselves that this is real, partly because we’ve laid all the track. We’ve got the data estate in order. We have our T-Life app out the door. We’re going to get the hardest part done in 2025. Once we’ve moved into big elements of delivery, I think we will have broken the back of this opportunity in the next 12 months, and we can taste that right now.

(To receive weekly emails of conversations with the world’s top CEOs and decisionmakers, click here.)