from the and-a-pony dept

Mon, Aug 24th 2026 05:32am - Karl Bode

Earlier this month I noted that Elon Musk’s Starlink is giving very unsubtle indications that the company wants to jump into the wireless business. I also noted that there’s a very long list of reasons why this isn’t likely to go well for him, ranging from the extremely high cost of network build-outs, to the fact that entrenched giants like AT&T and Verizon are very good at crushing insurgents.

As a launch gets closer we’ve seen more details into how Starlink actually hopes to try and make a wireless phone service work. Experts have already noted that the low-Earth-orbit Starlink satellite network is generally too congested to scale in the way the SpaceX IPO claims it can. The IPO projects a jump from 10 million to 300+ million in just a few years; an impossible feat.

Some had speculated that Starlink would accomplish this by buying a company like T-Mobile. The company also is poised to buy around $17 billion in AWS-4 and H-block wireless spectrum licenses from Echostar after Brendan Carr specifically launched an “investigation” into Echostar making it possible.

But instead of buying T-Mobile, Musk seems convinced that he can launch a nationwide wireless phone service by simply plunking down thousands of meshed femtocells installed on customer rooftops alongside existing LEO satellite dishes. Such femtocells would eat up backhaul capacity shared with the already capacity-constrained satellite-delivered broadband.

Even normally staid analysts at industry trade mags have called the idea incredibly stupid:

“This ranks as one of the top three dumbest ideas in my four decades of being in this industry,” said Earl Lum, the founder of analyst company EJL Wireless Research. What’s currently unclear is SpaceX’s precise definition of a small cell, but a typical outdoor small cell would come with power output of 5 watts per channel, in Lum’s book, and be difficult to install at residential properties.

“You need a real antenna in three sectors. To deploy this, it means you have to have a pole on a roof that’s good enough to hang three radios and three antennas, which is going to be hard,” he said. “You would have to go through the standard permitting for any macro cell site, and at that point why do you want a small cell?”

Keep in mind that Musk’s companies (especially Starlink and Tesla solar) don’t really have, or believe in, functional customer service. So the idea that existing Starlink customers are going to make all this work without coherent support is another wrinkle. There’s very little indication this would work; and it’s near impossible to make it scale up in urban areas where they’d compete with AT&T and Verizon.

Musk and friends may belatedly realize the unworkable nature of the idea later, at which point they just gobble up T-Mobile, which has steadily become shittier and shittier in the wake of the Sprint merger (precisely as deal critics predicted). Though even that would be very expensive and include lengthy, cumbersome integration, with no guarantee of meaningful success.

Filed Under: elon musk, mvno, phone service, spectrum
Companies: spacex, starlink, xai