T-Mobile has delivered a strong third quarter and raised its full-year guidance. The US operator beat market expectations on earnings and revenue. Yet the most important message was not only financial. The company used its earnings call to outline bigger network plans.
Service revenue reached $19.4 billion, up 6% year over year. Adjusted EBITDA climbed 11% to $9.6 billion. Net income rose 27% to $3.5 billion. Postpaid service revenue also grew strongly, helped by premium plans and higher-value customers.
However, the bigger story sits in spectrum and AI infrastructure. T-Mobile expects to reserve more capital for upcoming FCC auctions. These include upper C-band spectrum in 2027 and 2.7 GHz spectrum in 2028.
Chief executive Srini Gopalan said the operator will stay disciplined. He noted that T-Mobile walked away from deals when pricing lacked value. That includes the EchoStar spectrum opportunity.
More spectrum could support wider coverage, faster service, and fixed wireless growth. It could also help enterprise services and future automation workloads. Still, spectrum spending can pressure capital plans. Operators must balance buying airwaves with building denser networks.
T-Mobile links this decision to its long-term leadership in 5G. Gopalan said the company’s early 5G push transformed its business. It helped boost rural share and expand fixed wireless access.
The company now sees a similar opening around physical AI. This means connected robots, drones, vehicles, and machines using mobile networks. Gopalan described “millions of robots, drones, [and] all sorts” as “kind of a given”.
To support that world, networks need fast response times and edge computing. Edge computing moves processing closer to users or devices. That reduces delays and can lower hardware demands on devices.
T-Mobile says its Kinetic Edge work explores this model. The operator is working with partners including AWS, Figure AI, and Serve Robotics. The focus includes robot intelligence, fleet management, and distributed cloud services.
This approach could create new revenue beyond basic connectivity. T-Mobile says customers may pay for complete solutions. If edge processing reduces battery size or onboard AI hardware, device costs can fall.
Yet the market remains early. Business models are still forming. Operators must fund compute capacity, manage traffic growth, and prove service reliability.
Chief technology officer John Saw also highlighted changing traffic patterns. He expects networks to remain downlink heavy by 2030. But uplink traffic will grow as AI devices send more data.
Saw pointed to AI-RAN as another promising area. T-Mobile has tested AI-based radio planning with Ericsson in New York. “Candidly… the gains were actually more than we expected,” he said.
Meanwhile, T-Mobile rejected any mobile virtual network operator deal with SpaceX. The company remains open to commercial resale or edge compute arrangements. But it sees no need for SpaceX as an MVNO partner.
Gopalan was direct on that point. “We see ourselves in the lead in terms of those capabilities, and we see no space for an MVNO here because what additional capabilities does it give us?”
The message is clear. T-Mobile wants to lead in spectrum, AI infrastructure, and next-generation mobility. But it will avoid partnerships or purchases that fail its value test.
