Verizon Communications (VZ) reported its strongest third-quarter wireless subscriber growth in five years on Oct. 26 and raised its full-year outlook.

Adjusted earnings beat the analyst estimate. Revenue missed. Shares still fell 3.49% to close at $43.10.

The question that dominated the earnings call was not about handset upgrades or guidance ranges. It was about Starlink.

Chief Executive Dan Schulman said Verizon is open to working with satellite providers, including Elon Musk’s SpaceX, to extend coverage in areas where terrestrial networks cannot reach.

Still, he drew a clear boundary around what that cooperation would not include.

The one deal Verizon would do with Starlink

Schulman said satellite connectivity has a legitimate role in rural and hard-to-reach areas where building cell towers is not economical. He said Verizon sees potential to pursue coverage arrangements through direct partnerships or joint ventures with satellite providers.

Also read: AT&T CEO holds firm stance on Elon Musk’s SpaceX phone strategy

That is a different arrangement than an MVNO deal. A mobile virtual network operator agreement lets a company sell wireless service over another carrier’s infrastructure without owning the underlying network.

Schulman said Verizon carries no backdoors in its existing MVNO contracts and has no plans to create them for satellite companies.

“I see absolutely no benefit whatsoever to extending an MVNO deal to any satellite provider,” Schulman said on the earnings call.

The line Schulman drew separates two outcomes. In one, a satellite provider fills coverage gaps for Verizon customers in remote locations, adding value within Verizon’s network. In the other, a satellite provider uses Verizon’s infrastructure to reach those customers with a competing mobile offer.

Verizon is open to the first. Schulman said the second is not under consideration.

Verizon has already entered into satellite coverage arrangements through a partnership with AST SpaceMobile, which uses low-earth-orbit satellites to extend coverage to standard smartphones.

That deal represents the approach Schulman outlined: satellite-based service that fills gaps in Verizon’s terrestrial network without giving a satellite company direct access to Verizon’s core customer base.

Why Starlink can’t replace Verizon yet

Schulman said Verizon takes SpaceX’s plans seriously and assumes the company will execute on them. His case against treating satellite as an imminent competitive threat rested on the scale of infrastructure SpaceX would need to build to reach Verizon’s level of terrestrial coverage.

To replicate Verizon’s network, Schulman estimated SpaceX would need 65,000 to 80,000 macro cell sites, along with backhaul infrastructure, radio-access equipment, and meaningful spectrum holdings. He put the cost at $300 billion to $500 billion, with the physical buildout taking four to six years.

He also addressed the argument that Verizon should offer Starlink network access as a defensive move to prevent SpaceX from finding another path to its customers.

“We will punch back hard when that happens,” Schulman said.

How Verizon hit its best subscriber quarter in five years

Verizon’s third-quarter results showed adjusted earnings of $1.31 per share against a $1.27 consensus estimate. Revenue came in at $34.31 billion, below the $34.74 billion estimate. The company said the revenue shortfall reflected fewer device upgrades after it pulled back on handset subsidies.

Wireless subscriber additions were the standout figure. Verizon called it the strongest third-quarter performance in five years on that metric, and the company raised its full-year financial outlook alongside the results.

More Verizon:

• Verizon faces trouble ending an old phone service customers use
• Verizon takes on T-Mobile with new phone plan for customers
• Verizon scales back a perk that keeps prices low for customers

Postpaid phone net additions came in at 406,000, up more than 820% from the same quarter a year earlier. Broadband additions totaled 314,000 for the quarter, split between fixed wireless and fiber, bringing total broadband connections to 17.4 million.

Postpaid phone churn improved to 0.94% from 0.98% a year earlier, a sign that fewer existing subscribers are leaving the network, despite Verizon’s pullback on handset subsidies.

Investors weighed the earnings beat and subscriber momentum against the revenue miss. Shares closed down 3.49% at $43.10 on Oct. 26, the day the results were released.

Why Elon Musk disagrees with Verizon’s math

Schulman’s timeline has already drawn a public response. Musk disputed the projection, saying SpaceX could build a competing terrestrial network in two to three years, roughly half the time Schulman outlined on the call.

Whether SpaceX pursues that path depends on more than timelines. Bloomberg reported days before Verizon’s earnings that SpaceX was in active talks with cellular infrastructure companies to secure access to at least 80,000 cell sites, close to the upper end of Schulman’s own estimate of what a competing network would require.

Those talks suggest SpaceX is moving beyond satellite coverage into direct terrestrial competition.

Schulman’s cost estimate assumes SpaceX follows the capital structure of traditional wireless buildouts: acquiring spectrum, building macro sites from the ground up, running backhaul and radio equipment.

If SpaceX can secure site access through third-party agreements instead of building infrastructure outright, both the cost and the timeline compress significantly.

That is the scenario Verizon is preparing for. Starlink’s growth into mobile has already put pressure on all three major U.S. carriers. Schulman’s case is that the economics make a full terrestrial buildout far harder than Musk projects.

How fast SpaceX moves on those 80,000 site talks is the number to watch.

SpaceX currently operates Starlink as a satellite broadband service and has run limited direct-to-device tests in partnership with select carriers. Moving into full retail wireless would require more than cell site access.

SpaceX would also need consumer distribution, customer service infrastructure, regulatory spectrum licenses across multiple bands, and the back-office systems that support a nationwide subscriber base.

Schulman said his $300 billion to $500 billion estimate excluded those costs entirely. Retail operations, customer service, and back-office systems were not included in the figure.

The investment required to reach Verizon’s scale as a consumer carrier would be higher than the number he put on the call.

Related: Verizon takes on T-Mobile with new phone plan for customers