SpaceX weighs 80,000 US sites for a terrestrial Starlink network

By Rebecca Stone

SpaceX is discussing access to 80,000 US mobile sites to build a terrestrial wireless network, according to Bloomberg, putting a potential Starlink footprint on roughly the same scale as the networks of AT&T, T-Mobile and Verizon. Crown Castle and other tower companies are involved.

Those discussions do not amount to a disclosed rollout, and Starlink has not explained how its hybrid satellite-terrestrial model would make money or why it wants to challenge the three national carriers.

What would the hybrid network do?

On Tuesday's earnings call, SpaceX president Gwynne Shotwell said Starlink Mobile will mostly use S-band spectrum from the EchoStar acquisition. She said 14 MHz of low-band 800 MHz spectrum being acquired from Grain Management would help penetrate buildings, supported by Starlink rooftop dishes and other towers.

Shotwell also said Starlink Broadband would be in the mix because most mobile connectivity is actually Wi-Fi.

Starlink has FCC approval to deploy up to 15,000 V2 satellites for direct-to-device service. Shotwell said the combination of that authority and the low-band spectrum puts Starlink on a path to becoming a major US mobile carrier.

Critics say SpaceX's spectrum holdings, including the low-band holdings, are insufficient to match the big three. No launch timetable, tower contract value or subscriber target has been disclosed, so regulatory permission does not yet establish a commercial network.

Are the tower talks real?

Crown Castle CEO Christian Hillabrant said this week the company was “talking to all of them” when asked whether discussions were under way with SpaceX and other satellite players. He said SpaceX's mobile ambitions could be a catalyst for the wireless infrastructure company.

His wording indicates discussions, not an agreement to provide site access. The emerging route otherwise looks conventional: leased towers and an MVNO arrangement.

New Street Research analyst David Barden said SpaceX still wants an MVNO agreement with one US carrier, based on his conversations with the company. He said a tower buildout and an MVNO were not mutually exclusive, but questioned what end they serve.

Barden said he would like to see Musk's presentation on the mobile strategy before fully accepting the company's claims.

Is Dish the route to scale?

TMF Associates analyst Tim Farrar posted on X that the obvious starting point would be the acquisition of Ergen's existing network, with its leases folded into the tower-company discussions. His suggestion remains speculation; no acquisition talks have been confirmed.

Why is Musk asking investors to wait?

At this week's equity analyst call, Musk said SpaceX should hold a longer presentation about its Starlink strategy and why he thinks the connectivity business will generate $100 billion in annual revenue at some unspecified point. He also said outside investors were missing pieces of information known internally.

Musk wrote on X: “We definitely need to explain this in detail. Veterans of the cellular industry look at the chess board of existing pieces and don't see a solution. They don't understand that SpaceX is developing entirely new pieces that fundamentally alter the strategic landscape.”

That response leaves the central commercial questions unanswered. No timeline has been given for the fuller presentation, the $100 billion revenue milestone or a terrestrial rollout.

Can the bandwidth forecasts support the plan?

Musk put three figures behind his argument:

• Bandwidth demand: 1,000 times higher than today within three to five years, according to Musk.
• Constellation bandwidth: 100 times today's level, probably within the next three years.
• Satellite target: on the order of 100,000 spacecraft, each carrying more than 10 times the bandwidth of the current Starlink system.

Shotwell said connecting cars, robots and the adoption of superintelligence would raise communication demand by orders of magnitude rather than percentages. Musk said AI or superintelligence agents would use more bandwidth per agent than people, while the number of agents would exceed the number of people.

Musk said Starlink revenue would be at least somewhat proportional to total bandwidth delivered. He also described high-bandwidth direct-to-cell capability and terrestrial buildout as the route to broad-based connectivity worldwide wherever SpaceX and Starlink may operate.

These remain management projections rather than measured output, and they do not disclose per-user capacity or capital costs. The 15,000-satellite V2 authorization also does not itself cover a constellation on the order of 100,000, and no timetable for further regulatory approvals has been disclosed.

Has Starlink's posture toward carriers changed?

At the Satellite 2020 conference in March 2020, Musk presented Starlink broadband as a way for telcos to reach hard-to-reach areas and insisted that it was not a huge threat. He wanted to be clear on that point.

SpaceX has since changed posture and now plans to compete with traditional US carriers. The available details do not include a site-access agreement, capital budget, launch schedule, subscriber target or retail pricing, making the scale of the threat difficult to assess.

What is the related AI-RAN development?

T-Mobile made the first AI-RAN call with Nokia and Nvidia, while also reporting performance gains with Ericsson's AI beamforming. No test configuration, throughput measurement, deployment scale, commercial timetable or subscriber effect was disclosed, so the result shows integration progress rather than a quantified network upgrade.

Until SpaceX publishes a detailed Starlink carrier plan, secures tower and MVNO agreements and sets a regulatory timetable for expansion beyond the cited 15,000-satellite approval, its 80,000-site plan remains a negotiating target rather than a scheduled rollout.