SpaceX could require between $50 billion and $130 billion to construct a standalone terrestrial mobile network for its direct-to-device services, Bernstein estimated. The range includes spectrum acquisition costs, with the brokerage's base-case scenario—incorporating low-band spectrum from Grain—projecting approximately $70 billion in total expenditures.
Excluding spectrum, Bernstein's cost projections for the network range from $15 billion to $80 billion, depending on design choices. The firm noted that securing as little as 10 MHz of low-band spectrum, such as Grain's holdings, could reduce the number of required macro sites by roughly 30%, improving efficiency. The base-case National build is estimated to cover about 70% of the U.S. population over an eight-year period, requiring around 57,000 macro sites.
Bernstein highlighted that the direct-to-device mobile business represents the most challenging segment of SpaceX's broader opportunity, which also includes launch services, orbital data centers, and Starlink broadband. The firm maintained a positive outlook on SpaceX stock, citing the potential across these ventures. Bernstein also identified Grain's spectrum as the only readily available low-band option, though Elon Musk disputed reports of SpaceX pursuing the spectrum on social media platform X.
The timeline for deployment includes mid-2027 for the commencement of Starship launches under SpaceX's Mobile V2 satellite constellation. AST SpaceMobile, a partner in the initiative, secured a 30-day special temporary approval from the Federal Communications Commission in August 2026 to test supplementary coverage using Grain's spectrum.
