80,000 cell sites: SpaceX hunts the ground for a fourth phone network, and the market splits in two

Eighty thousand cell sites. That is the shopping list, and on Friday it split the American telecoms market clean in two. Bloomberg News reported late on Thursday that SpaceX is in talks to access at least 80,000 cell sites to build a fourth US wireless carrier from scratch. By Friday the landlords were celebrating and the tenants were not. Crown Castle, named in the report as one of the infrastructure firms in talks, surged 11.6 percent. American Tower rose 4.9 percent and SBA Communications added 4.6 percent. On the other side of the same trade, AT&T fell 5.6 percent to a low of more than three months, T-Mobile shed 5 percent, Verizon lost 1.6 percent and Deutsche Telekom, which owns more than half of T-Mobile US, lost 3.2 percent in Europe. SpaceX itself rose 1.7 percent.

The split makes sense once you see who owns what. Crown Castle owns, operates and leases about 40,000 cell towers in the United States, and the big three carriers supplied roughly 90 percent of its site rental revenue in 2025. A tower company does not care which carrier wins the customer. It rents the same steel to whoever shows up, on contracts that run for years. A fourth national carrier is not a threat to that business. It is a fourth tenant, and tenants are the whole business.

SpaceX starts from the opposite corner. Starlink already earns money selling internet in the countryside and in places the cables never reached, and a few weeks ago the company acquired low band spectrum to advance its carrier plans, a move that at the time sent AT&T and Verizon to their worst one day falls in more than two decades. What SpaceX does not have is a network on the ground. Phones in cities, offices and stadiums do not talk to satellites for long. They talk to towers, street by street, and 80,000 sites is roughly the footprint a real national carrier needs before anyone takes its coverage map seriously.

Nothing is signed. The tower firms and SpaceX did not comment on the report, and Bernstein, the research house quoted by Reuters, put the honest caveat in one sentence: the competitive impact may still be years away, but anticipation alone, reinforced by each successive spectrum and infrastructure move, is likely to remain an overhang on sector sentiment. In plain terms, the carriers may not lose a single customer to SpaceX for years, and their shares can still be taxed for the threat every quarter until they do.

Our call is that the tower move is the tell. Spectrum can be bought in an auction and announced in a press release. Eighty thousand site leases cannot be faked, hurried or quietly dropped without the market noticing, so if agreements start appearing, this stops being a rumour with a share price attached and becomes a building programme with a deadline. Watch for the first named lease, not the next headline. Until one signs, the landlords collect the premium and the carriers pay the fear tax. The day one signs, both sides of Friday's split stop being a trade and start being a forecast.