- T-Mobile's 5G-Advanced improves speeds, consistency and network slicing for enterprises and delivers faster and more reliable consumer networks, says T-Mobile's Ankur Kapoor
- AI-powered automation helps T-Mobile adapt its network in real time to disasters, remote coverage gaps and individual customer needs
- 5G-Advanced is "training wheels for 6G," with AI-native design as the defining feature of the next generation, Kapoor said.
I hate hearing about 6G. It's the same reaction I have when I see Christmas decorations in September. "Dear me, no!" I say. "Too soon!"
I'm still bruised from the unfulfilled hype of 5G. You remember: Self-driving cars, remote surgery, smart cities and factories, mixed reality, a new agricultural revolution, etc. Those advances have been slow to materialize, haven't required 5G when they do, and telcos have been left with staggering infrastructure costs that failed to deliver promised new revenue streams.
My reaction isn't quite as dramatic when I hear about 5G-Advanced. I just shrug. "Who cares?" It just seems like a new marketing buzzword.
So when T-Mobile offered an opportunity to discuss 5G-Advanced with Ankur Kapoor, the company's EVP and chief network officer, my main question was "Why should consumers or businesses care about 5G-Advanced?"
What 5G Advanced actually does
Kapoor has a 30-year career at T-Mobile, through the transitions from 2G to 5G. "I thought I didn't have any more Gs left in me," he told Fierce. But he's excited about 5G-Advanced and the outlook for 6G and its potential for AI.
For consumers, 5G-Advanced gets better quality on their video calls and gaming sessions, with reduced jitter and lag that come from network congestion. T-Mobile implements these improvements using L4S, an IETF network protocol and congestion control technology designed to reduce packet loss and latency while preserving network throughput. T-Mobile deployed L4S nationwide in 2025.
These advances are already available to T-Mobile subscribers, automatically and at no additional cost, Kapoor said.
Another feature of 5G-Advanced, RedCap (Reduced Capability), provides connectivity to connected devices that don't require a full smartphone radio, such as wearables, industrial sensors and mobile hotspots. RedCap modems are cheaper to build and draw less power than full smartphone radios, lowering the cost of always-connected devices for consumers and enterprises alike. T-Mobile deployed RedCap nationwide, Kapoor said.
For enterprise customers, the flagship feature is network slicing — a virtual private network carved out of shared physical infrastructure, Kapoor said.
For example, T-Mobile's T-Priority product gives first responders a dedicated slice, so their connections are prioritized automatically at a disaster scene. The NYPD was the first T-Priority customer. "Emergency connection is the most important connection that T-Mobile carries," Kapoor said.
The foundation for all of this is T-Mobile's 5G standalone (SA) core network, which went nationwide in 2020, years ahead of rivals.
AT&T made its nationwide 5G SA deployment official in October 2025 and Verizon followed suit around the same time — both carriers citing the 5G SA core as the pathway to 6G. Without a standalone core, carriers are limited to two-way carrier aggregation, bonding two wireless bands together for improved performance.
T-Mobile supports up to six-way, which results in iPhones on T-Mobile running 48% to 50% faster than on competing networks, Kapoor said. Both AT&T and Verizon have begun offering network slices now that their 5G SA cores are live, but T-Mobile has a two-to-three-year head start in deployments and operational learning.
That lead shows up in consumer perception. T-Mobile edged past Verizon on brand image for the first time in 13 years of TD Cowen surveys, conducted at year-end 2025, with T-Mobile scoring 1.96 versus Verizon at 2.01.
Competition is heating up for T-Mobile
But T-Mobile faces stiff competition. Now that AT&T and Verizon have achieved nationwide 5G standalone unlocks the same core capabilities — network slicing, RedCap, advanced carrier aggregation — that T-Mobile has been building on for five years. Verizon has been pushing C-band into suburban and rural markets as it works to close T-Mobile's mid-band coverage gap, while Comcast and Charter have been steadily accumulating wireless subscribers. — Charter had about 10.4 million and Comcast 8.1 million as of July 2025 — by bundling mobile service with existing home internet bills through MVNO arrangements, without building a single tower of their own.
AT&T and Verizon have both made fiber-plus-wireless convergence important to their strategies, bundling home broadband and mobile service to reduce churn and deepen customer relationships. AT&T closed 2025 with more than 32 million fiber passings and is targeting 60-plus million locations by 2030, with 42% of its fiber households also subscribing to AT&T wireless. Verizon closed its $20 billion acquisition of Frontier Communications in January 2026, giving it nearly 30 million fiber passings across 31 states and a platform for the same bundled convergence play.
T-Mobile has no fiber network of comparable scale and competes in home broadband exclusively through fixed wireless access — a service that draws on the same spectrum it uses for mobile customers, a capacity constraint its rivals don't face in the same way. T-Mobile does have a nascent fiber venture, but at roughly 130,000 annual additions, it remains a rounding error next to AT&T's and Verizon's ambitions.
AI is reshaping network operations
Beyond consumer features, Kapoor described how AI has transformed the way T-Mobile manages and expands its network.
The most vivid example is disaster response. During Hurricane Helene, T-Mobile's self-organizing network made 120,000 antenna configuration changes in three days — adjusting coverage physically and in software as populations shifted and disaster camps formed. AI measured traffic patterns every 15 minutes and redirected capacity accordingly. The result: 98% of customer issues resolved within 12 hours of an outage, a significant improvement over prior storms, Kapoor said.
T-Mobile uses customer-driven coverage (CDC) — a model that tracks where customers actually use their phones and expands the network to match, rather than building outward from population centers.
For an example of CDC benefits, Kapoor cited a case where performance complaints apparently coming from Sacramento customers were actually originating at Lake Tahoe, a popular vacation destination for Sacramento residents. T-Mobile's old network planning model would never have caught that, because it was built around where customers live, not where they actually use their phones. The CDC model uses real device data to close that gap, directing capital to where people actually need coverage.
In other words, the network now follows people, not just ZIP codes. T-Mobile plans to build roughly 4,000 new cell sites in 2026 using this model.
The AI integration extends to live translation, a consumer feature Kapoor finds particularly compelling. Using AI on the T-Mobile network, two people speaking different languages can have a real-time translated phone call — as long as one party is on T-Mobile. With 6 billion international calls made by T-Mobile customers annually and 60 million multilingual households in the United States, the addressable need is substantial.
Kapoor has spoken elsewhere about the strategic arc behind all this, including T-Mobile's move to run more intelligence in the cloud core network and the carrier's partnership with Nvidia to bring AI inference directly to the radio access network edge.
Training wheels for what comes next
Kapoor is explicit about the strategic purpose of all this investment. "5G-Advanced is training wheels for 6G," he said. Prior generations — 2G through 5G — treated AI as a bolt-on, something layered on top of the network rather than woven into it. 6G will be AI-native from the ground up, Kapoor said.
The shift he describes isn't about moving bits more quickly. It's about moving tokens — the language of machine intelligence — as a first-class network function. If 4G made the internet mobile, 6G will make AI native to the network itself, Kapoor said.