T-Mobile US has yet to see AI-derived mobile traffic

During T-Mobile US's second quarter earnings call in July 2026, CTO John Saw delivered a striking admission: the carrier has detected no surge in AI-derived traffic on its mobile network. Current AI compute, he explained, is concentrated on "model training, large-scale agentic automations, and heavy backend automations," nearly all of it running on "wireline transport networks and massive data centers." This type of workload, Saw noted, has "not seen putting a material strain on mobile networks."

Yet T-Mobile US management is not waiting passively. Saw emphasized that the carrier has "sufficient runway to continue to invest in our capacity" and is preparing specifically for mobile AI traffic through infrastructure already deployed. The carrier recently became the first operator to roll out 5G-Advance technology, a move Saw tied directly to anticipated AI demand. 5G-Advance improves uplink performance—the data flow from device to network—through uplink carrier aggregation, uplink MIMO (multiple-input, multiple-output systems), and transmit switching. "All that is actually in preparation," Saw said, "for not just giving our customers with phones a better experience but also better uplink for future AI traffic that we expect to be seeing soon."

CEO Srini Gopalan amplified this message, stating the company looks forward to AI-driven traffic growth because "our network's more prepared than anyone else." Gopalan has previously emphasized physical AI—systems designed to power autonomous vehicles, robots, and factory automation—as a major driver of future mobile demand. These systems require low-latency connections and edge processing capabilities. He framed this as a shift away from ChatGPT-style models toward "automation and robotics," which demand 6G networks that can deliver both low-latency connections and edge infrastructure capable of "processing tokens," not just bits and bytes.

Meanwhile, T-Mobile US continues managing strong demand for 5G-based fixed-wireless access (FWA), a service the carrier sells based on excess network capacity. Average FWA customers now consume 580 GB per month of data—up from 450 GB two years prior. Some analysts have raised concerns about whether the carrier can sustain the service without compromising network quality, particularly as FWA revenue per customer remains lower than traditional handset connections. Gopalan dismissed these concerns at the latest earnings call, arguing that while FWA consumes "a fair amount of our capacity," the carrier's total capacity is "several times multiple" of current traffic, making the current load manageable. The carrier recently reshuffled senior leadership to address growth opportunities: chief business and product officer Mike Katz is departing, replaced by former AT&T veteran Chris Sambar, who will take the title of chief enterprise officer. Analysts noted Sambar's background as a lead architect of FirstNet—AT&T's complex government wireless deployment—positions him to expand T-Mobile US's 10% market share in federal and mid-market enterprise segments.

T-Mobile US faces a paradox: while AI adoption is accelerating across the tech industry, the carrier has detected no measurable impact on its mobile network yet. This gap between hype and reality reflects where AI compute is concentrated—backend training, large-scale automations, and data center operations remain tethered to wireline and fiber infrastructure rather than wireless. T-Mobile US management has positioned this lag as an opportunity rather than a concern, using it to justify continued network investment and to position the company as uniquely prepared when demand does arrive.

The carrier's confidence rests on spectrum depth, recent 5G-Advance deployment, and an expanding fiber footprint that CEO Gopalan frames as a "numerator-denominator" advantage: FWA (fixed-wireless access) traffic consumes substantial capacity, but capacity itself is "several times multiple" the current load, leaving room for future growth. This same logic extends to AI—the company has invested in infrastructure capacity before demand peaks, particularly through uplink improvements designed for the low-latency, token-processing demands of physical AI systems like robots and autonomous vehicles. Gopalan's emphasis on 6G and edge processing suggests T-Mobile US is betting that the next wave of AI will move beyond language models into embodied automation, a shift that would fundamentally alter how network traffic flows and what speeds matter most.

No comments yet. Be the first to share your thoughts!