Elon Musk: Starlink Mobile U.S. carrier network will be built in 2 to 3 years
On October 26, 2026, Elon Musk replied on X that SpaceX’s Starlink Mobile U.S. carrier network would “take 2 or 3 years.” The reply was in response to Verizon CEO Dan Schulman’s assessment during the company’s third-quarter earnings call that same day. Schulman said that if SpaceX were to build a competitive U.S. mobile network, it would take at least 4 to 6 years and would involve 65,000 to 80,000 large base stations, backhaul, RAN, and a large amount of spectrum, with cost estimated at $300 billion to $500 billion over the next 7 to 8 years.
Schulman also pointed out on the call that satellite service itself cannot compete in suburban and urban areas, which account for about 98% of Verizon’s revenue, and said Verizon is 100 to 1,000 times more efficient in these areas. He said that assuming SpaceX executes its plan perfectly, it would still need retail, customer service, and back-end systems after completion, and could then become the market’s sixth-largest rather than fourth-largest competitor, and he ruled out providing an MVNO agreement to a satellite provider.
SpaceX had previously advanced the Starlink Mobile-related layout, including acquiring spectrum assets and obtaining FCC approval for next-generation satellites, with the goal of providing services by combining satellites and ground facilities.
Original sources:
Title: Will take 2 or 3 years Publisher: Elon Musk (@elonmusk) Date: 2026-10-26 URL: https://x.com/elonmusk/status/2114774224015663608 Source type: direct post by the party involved Verification status: Full text of the original post was actually read in this check, verification date 2026-10-26 Supporting content: Elon Musk directly replied, “Will take 2 or 3 years”
Title: Verizon: SpaceX Needs 4-6 Years to Build US Carrier Business Publisher: PCMag (Michael Kan) Date: 2026-10-26 URL: https://www.pcmag.com/news/verizon-spacex-needs-4-6-years-to-build-us-carrier-business Source type: media report Verification status: Relevant passages were actually read in this check, verification date 2026-10-26 Supporting content: Schulman’s statements about 4 to 6 years, $300 billion to $500 billion, 65,000–80,000 sites, and the “sixth-largest competitor” as well as Musk’s reply
Title: Verizon says SpaceX needs up to 8 years and a half trillion dollars to compete in wireless Publisher: Business Insider (Dominick Reuter) Date: 2026-10-26 URL: https://www.businessinsider.com/verizon-starlink-needs-huge-investment-to-be-competitive-in-us-2026-10 Source type: media report Verification status: Relevant passages were actually read in this check, verification date 2026-10-26 Supporting content: Schulman’s statements about efficiency comparisons, the 98% revenue areas, and the cost/time estimates
ABAB AI Insight
SpaceX’s Starlink business began in the late 2010s and has grown into a global low-Earth-orbit satellite broadband system. After the company completed its IPO in June 2026, it announced on October 8 that it had agreed to acquire Grain Management’s nationwide 800 MHz low-band spectrum portfolio (up to 14 MHz paired spectrum, with media reports putting the price at about $8 billion); it had also previously agreed to acquire about 65 MHz of mid-band spectrum from EchoStar (involving roughly $17 billion to $20 billion). These assets are intended to support Starlink Mobile’s indoor penetration and capacity, and the FCC has approved the deployment of up to 15,000 next-generation satellites.
On the capital side, the spectrum acquisitions and satellite deployment are supported by the company’s existing cash, IPO proceeds, and operating cash flow. SpaceX’s total revenue in the third quarter of 2026 reached $14.6 billion, of which Connectivity revenue (mainly Starlink) was $5.25 billion, up 69% year over year; during the same period, AI-related business revenue had already exceeded Starlink’s. The company plans to advance with a hybrid architecture, including satellite-to-phone direct connectivity and small base stations deployed at existing Starlink installation sites, rather than a full rollout of traditional large macro base stations.
Compared with traditional carriers, Verizon and others rely on tens of thousands of macro sites and long-term spectrum accumulation, and Schulman’s estimate is based on that assumption. SpaceX, by contrast, emphasizes covering blank areas with satellites and small distributed nodes; the difference lies in not pursuing a traditional tower network with the same density. Historically, similar satellite direct-to-device attempts have often been limited by spectrum and terminal compatibility; this time, the additional low-band spectrum and the existing Starlink user base (more than 13 million consumer subscribers) provide a different starting point.
Musk’s statement about 2 to 3 years was directed at Schulman’s assumption of a full-scale traditional terrestrial network; if the path is satellite-led plus distributed small nodes, the time and capital requirements could be significantly lower than the $300 billion to $500 billion range. Key indicators to watch are the progress of spectrum transaction closings, the pace of next-generation satellite launches, and the actual scale of ground-node deployment, rather than simply the number of macro sites.
ABAB News · Cognition Law
For the same network goal, the time and cost assumptions can differ by a factor of two between a traditional tower-site path and a satellite-plus-distributed-node path.
After spectrum acquisitions solve coverage penetration, carrier competition shifts from capital density to architectural choice.