On T-Mobile's third-quarter earnings call, CEO Srini Gopalan flatly rejected an MVNO partnership with SpaceX's Starlink Mobile, saying: "I see no room for an MVNO with SpaceX either on the consumer side or on the physical and edge AI side."

He explained that T-Mobile's MVNO criteria require a deal to deliver a new addressable market, new distribution channels, or new capabilities, and satellite service — which the carrier views as purely complementary — meets none of those thresholds.

He remains open to a simple resale arrangement like the existing T-Satellite setup, or to SpaceX becoming a T-Mobile customer, but called an MVNO "a wholly different thing."

SpaceX has been aggressively positioning Starlink Mobile as a full-fledged US carrier, and an MVNO would have offered the fastest route to market by renting capacity on an existing network.

According to gsmarena.com, SpaceX bought more spectrum earlier this month to expand beyond direct-to-cell satellite service and is reportedly in talks for access to at least 80,000 cell sites nationwide.

Notably, Verizon's CEO Dan Schulman made a similar rejection of any satellite MVNO days earlier, meaning both of the nation's largest carriers have now slammed the door — leaving SpaceX to build its own costly terrestrial infrastructure.

The rejections came alongside strong results: Light Reading reports service revenue up 6% year-on-year to $19.4 billion, postpaid service revenue up 9% to $16.3 billion, and adjusted EBITDA up 11% to $9.6 billion, with 274,000 net postpaid accounts added.

Gopalan also dismissed satellite broadband for home internet, calling it — per gsmarena.com — "an inferior version of FWA" due to beam-size constraints, noting two-thirds of T-Mobile's home internet customers live in dense urban areas where fixed wireless wins.

The stakes extend far beyond SpaceX. On the same call, CTO John Saw claimed T-Mobile will be the first US carrier to launch 6G around the industry's expected 2030 timeframe, with new applications like integrated sensing and communications, where base stations behave as radars.

He expects 2030 traffic to remain downlink-heavy but increasingly uplink-rich, driven by AI wearables, agents and physical AI devices such as robots, drones and vehicles.

Gopalan admitted physical AI is in "very early innings" commercially, but argued the world of 2035 — with millions of robots and connected moving objects — is a given, and the real question is what infrastructure will support it.

SpaceX's path to a full US wireless network now runs through slower, more expensive self-deployment of terrestrial infrastructure, even as its spectrum purchases and cell-site negotiations signal a long-term build-out.

T-Mobile, meanwhile, is pressing ahead with its own physical AI vision — Light Reading details partnerships with Figure AI's humanoid robots, Serve Robotics' delivery robots, and AWS on a "kinetic edge" combining its 5G-Advanced network with cloud compute.

Upcoming spectrum auctions for upper C-band and 2.7GHz will give the carrier capacity for 6G and richer uplink bandwidth.

The backdrop is dramatic: former CEO Mike Sievert resigned from the board effective November 1, and Bloomberg reported Deutsche Telekom has paused plans to bid for a full takeover.