AT&T CEO John Stankey: Starlink Mobile Route Not Feasible
AT&T Chairman and CEO John Stankey expressed skepticism about SpaceX's mobile solution for phones, stating that relying solely on satellites to service mobile operations does not align with existing network architecture, and randomly deployed micro base stations cannot fill the hard gaps in terrestrial cellular coverage.
At an investment banking conference, he stated that one cannot expect to service mobile directly "from the outside," nor can one rely on "very small cells, randomly deployed" to solve the problem. He estimated that AT&T covers about 98% of domestic wireless demand, with satellite direct-connect devices only addressing the remaining 2% blind spots. The company's goal is to bring that 2% online, not to redo the 98%.
He rejected the idea of wholesale resale as a primary path: wholesale is for entering markets that one cannot reach; since AT&T can handle about 98% of traffic, it will not sign a dependency agreement for an mvno relying on a single constellation for "non-existent problems." The satellite side is more willing to connect with multiple constellations, including AST SpaceMobile, Amazon Kuiper, and SpaceX, only purchasing that small segment of off-network traffic.
On a physical level, he pointed out that satellite uplinks are more fragile, and AI scenarios require more uplink capacity; ground stations can quickly connect to fiber optics, while satellites cannot achieve the same low latency and capacity. He compared that each of the three major operators has about 300 MHz of available spectrum per base station, while the satellite point beam radius is much larger than the two to three kilometers of terrestrial cells, yet only overlaps about 80 MHz, making it difficult to replace indoor, hospital, campus, and venue coverage one-to-one.
SpaceX's public vision is to build fewer large towers and mount cellular base stations on Starlink antenna supports, using a large number of small stations and Starlink backhaul to cover the ground layer, and purchasing part of EchoStar's spectrum to operate both direct connections and ground services. President Gwynne Shotwell has expressed optimism about the ground layer and competing for customers from the three major operators; analysts have questioned power supply, site selection, return on investment, and the compatibility of basebands, which will only be rolled out in the new device cycle.
In market mechanisms, this is a struggle for pricing power between existing operators and a potential fourth network: those selling are dealing with sunk costs in fiber and sites, while those buying are shareholders worried about their market value being discounted by Starlink narratives. Funds remain tied to ground packages and fiber bundles; the beneficiary is AT&T, which does not have to relinquish 98% of control for 2% of traffic, while SpaceX, which wants to leap over the network construction timeline with small stations and satellites, is under pressure. The event-driven narrative comes from the competition following Starlink's mobile and spectrum transactions, not from new user migration data.
In July of this year, the combined market value of the three major operators evaporated by about $46 billion due to rumors of "space entering the mobile business." Stankey positioned competitors as tools for filling gaps who are "late to the game and need to catch up on decades of infrastructure," effectively relegating the mobile strategy from an alternative to an accessory.
Source: Public Information
ABAB AI Insight
This is not the first time Stankey has made these statements. During the summer earnings season, he already mentioned on television that new players are late, do not sign Starlink's main wholesale, and that 98% of traffic is handled internally; in September, he targeted small station architecture again at the investment banking event. AT&T recently purchased another package of spectrum from EchoStar, while SpaceX bought a different package, with both companies replenishing their stocks in the same spectrum relocation. Historically, traditional operators have defined satellites as backup and wilderness solutions before refusing to lease core networks to potential fourth competitors—similar to how they approached the transition from cable to wireless.
The capital path is to protect sunk costs on the ground. Fiber to the station, low-frequency coverage, indoor distribution, and mobile subsidy cycles represent decades of AT&T's billing; if they acknowledge that rooftop small stations plus Starlink backhaul can replace these assets, they would need to be revalued. Therefore, money continues to be invested in finding partners to fill that 2% gap rather than purchasing nationwide wholesale from a single constellation. SpaceX is mobilizing launch rhythms, user antenna retention, and EchoStar spectrum to turn broadband dishes into cellular stations, avoiding the billion-dollar checks required to buy operators or build macro stations. Both sides are competing for the entry point of "who defines the default mobile connection," not for wilderness messaging.
Comparatively, in the 2000s, Wi-Fi was once said to replace cellular, and cable television used Wi-Fi hotspots to challenge mobile: ultimately, they became complementary layers until someone simultaneously acquired spectrum, device cycles, and subsidy channels. T-Mobile once viewed Starlink as a direct-connect partner, while AT&T chose multi-constellation procurement. The industry phase is still in the control period of the ground network, with satellites on the eve of expansion, proving they can fill gaps but not yet proving indoor capacity. The slowdown in device replacement cycles and the need for compatible modems to wait for new devices lock SpaceX's ground layer into hardware upgrades, which is precisely the defensive trench most familiar to operators.
The structural judgment is that the transfer of pricing power has not succeeded. The mechanism is that physical and architectural factors are used to freeze valuations: with large point beam areas, weak uplinks, and poor indoor penetration, national package prices are still determined by the ground network. Only when small stations can truly provide stable uplinks in densely populated areas will pricing power shift from site rental to spectrum and terminals. The current function of these statements is to prevent the capital market from prematurely writing the "mobile strategy" as a replacement equation.
ABAB News · Cognitive Law
Covering 98% of people, the most fearful is that the 2% is written as a replacement
Networks that do not make physical sense are first rationalized in valuation
Not leasing out the core network is equivalent to not acknowledging the other party as a peer