SpaceX is moving hard and fast into the US mobile market. Its 2025 deal to acquire EchoStar mid-band spectrum for approximately US$17 billion (plus another US$2.6 billion tranche) was followed by an agreement to buy Grain Management’s nationwide low-band spectrum, writes Stewart Baines. SpaceX is now in talks to acquire EchoStar’s MVNO Boost Mobile or to gain access to its wholesale network arrangements. The ingredients are coming together to create a nationwide mobile network that blends satellite direct-to-device coverage and terrestrial infrastructure.
The question is, why bother? Why does the world’s richest man have such an appetite for taking on incumbents who complain about flat revenues, complex estates and an even more complex regulatory environment? Even with deep pockets, is it a good investment of both money and attention? And the question operators outside the US might be asking is, why should I care?
Mind the gap
Across OECD countries, 4G reached 99% of the population by the end of 2024, leaving relatively few people permanently outside coverage. Serving these customers profitably has proved difficult for terrestrial operators, although satellites sharing capacity across much larger areas could change the economics.
Countries with more dispersed populations offer greater potential: CCS Insight predicts that by 2030, satellite direct-to-device will be the main rural mobile service in countries with more than 50% rural population.
However, the unconnected also includes the temporarily unconnected in densely populated countries, where a call or data session starts with good coverage but then drops out. Satellite services like T-Satellite in the US, Rogers Satellite in Canada, O2 Satellite in the UK and One NZ in New Zealand are addressing this by offering infill for messaging, maps or calls. Combining permanent coverage gaps with intermittent demand makes a broader business case.
Starlink, along with AST Mobile, GlobalStar and others, serve as operators’ backstops for reaching the hard-to-reach. But who wants to be a bridesmaid? “Connecting rural Scotland and Wales will not be enough to justify the investment. We expect SpaceX to seek a much bigger role in mobile,” says Luke Pearce, a senior analyst at CCS Insight.
Appetite for more
SpaceX could assemble a nationwide proposition through wholesale access, targeted infrastructure purchases, and its own satellite network. A mobile core and billing system are achievable; cell sites, permissions and a field force are more difficult but not impossible. Nonetheless, this is not an overnight play.
“I’m confident they are in this for the long term. They don’t want a small seat at the table, but I’m less sure they want the responsibilities of a traditional mobile operator,” says CCS Insight’s Pearce.
Look ten years ahead and the rationale is clearer. Alongside smartphone subscriptions, a potentially more profitable segment could be connected devices – AI-enabled vehicles, industrial robots, glasses and pebbles running AI inference locally and in the cloud.
“Across Tesla, robotics and AI, connectivity becomes strategically valuable. Musk may want greater control over it rather than depending entirely on operators to extend their networks,” says CCS Insight’s Pearce.
Imagine connectivity bundled into a vehicle or robot subscription, independent of its owner’s smartphone contract. Starlink could provide some, if not all, of that connection without persuading the customer to abandon their existing mobile provider.
Wholesale partnerships can already deliver connectivity. Greater control could allow SpaceX to coordinate coverage, service priorities and pricing across its devices and AI services while retaining the customer relationship. SpaceX (or Tesla) wouldn’t need to own every terrestrial connection, but it could mean less dependence on partners’ investment decisions.
This could be a significant addressable market, both in the US and internationally. For instance, CCS Insight forecasts that satellite connectivity will be standard in more than half of new connected cars by 2035. As well as accommodating larger antennas, connected cars spend significantly more time outdoors than the average smartphone or wearable.
Affluent markets in Europe, the Gulf and Asia-Pacific could therefore be attractive even where mobile coverage is extensive. Consumers who can pay for connected vehicles and AI services offer a different opportunity than remote populations needing basic access. But these are the same affluent customers operators hoped would lift their flagging fortunes.
The race for space
SpaceX’s “acquisition” of xAI in February brings Musk’s connectivity and computing ambitions under the same roof. SemiAnalysis reports that SpaceX is targeting approximately 13GW of earthly data-centre capacity in 2027 and believes that this is achievable. In addition, SpaceX has also sought permission for up to a million satellites operating as orbital data centres. Musk is not alone in seeing the potential of edge AI being at the edge of space. Google launched a Project Suncatcher prototype in October, while Jeff Bezos’s Blue Origin has applied for a 51,600-satellite system for orbital compute. The idea being that solar power could help relieve terrestrial energy constraints, although rejecting heat in a vacuum requires radiators: the cold of space does not provide easy cooling.
These projects suggest how Musk’s interests might fit together: connected devices using AI services supported by infrastructure he controls. Orbital computing might not prove economical or the best location for time-sensitive inference, but does reveal a vision, a divisive one. After spending US$474 million unsuccessfully on Republican campaigns, he said, “Guess my election-interfering space lasers didn’t work this time,” following up with “Data centres in space looking pretty appealing right now,” alluding to the Democrat politicians objecting to data centre plans in their backyards.
Direction of travel
Better satellites and tighter integration through 6G should improve performance, but cannot eliminate the capacity advantage of densely deployed terrestrial networks. Indoor reception remains an issue, so Wi-Fi offload and terrestrial access will be required. SpaceX’s own terrestrial plans acknowledge the need for a hybrid approach.
SpaceX’s advantage over satellite competitors, even Blue Origin, is control over satellite production and launch. But capital committed to building or acquiring a terrestrial network competes with constellation expansion, satellite replacement (they don’t have a long life in space), and other ambitions. Deep pockets do not mean bottomless pockets, and Musk has recently acknowledged that the cost of running AI inference does not match AI subscription fees.
“SpaceX has the resources to pursue a much bigger role. But being able to build or buy a mobile network does not explain why it would choose to do so,” says CCS Insight’s Pearce.
And even if SpaceX can create a compelling proposition, does success in the US mean it’s an exportable model?
All over the world, from Africa to Asia, governments want control over domestic licensing and satellite spectrum. Faced with years of permacrisis, tariffs and technology being used as a geopolitical weapon, governments want to build national resilience and sovereign infrastructure that can be governed by organisations subject to local regulations.
The EU is a prime example. The EU has proposed favouring European providers for much of its 2GHz satellite allocation in 2027, while leaving little room for non-EU competitors. Deutsche Telekom, Orange, Telefónica and Vodafone announced plans to explore a joint European satellite-spectrum bid. Their proposal combines European governance and ground infrastructure with a satellite partner still to be selected. The US threatened retaliation if SpaceX is excluded.
The recent incident in which Ukrainian drones lost connectivity with SpaceX’s Starshield military communications service because of decisions made in Washington illustrates the consequences of external control over essential connectivity.
Indeed, satellite connectivity is appealing to European operators such as Vodafone and AST SpaceMobile’s Satellite Connect Europe venture, which offers an operator-facing model intended to preserve customer relationships and European control of operations and data handling. Operators that can find suitable satellite partners could gain coverage and resilience while retaining enterprise integration, service management and the customer relationship.
The issue for operators will be whether SpaceX is friend or foe. If Musk can bundle connectivity into vehicles, robots and AI services, he could reach their most valuable customers by stealth. Operators would still carry some traffic but could find that someone else owns the service and captures much of its value. The question is how much of the pie Musk is willing to share.
Stewart Baines