For years, subscribers to satellite services have installed dishes at their homes for watching TV or connecting to broadband. Smartphone customers, by contrast, aren’t normally required to set up a basestation and cover the related property fees. Elon Musk wants to change all that through a widely ridiculed plan of Starlink, the satellite comms subsidiary of his overvalued SpaceX business. Experts with anything positive to say about it seem to be rarer than Mars landings.
The small cell technology Musk wants to combine with Starlink’s satellite receivers would naturally sit on your rooftop. But it wouldn’t be for your exclusive benefit, instead providing coverage for passersby on Starlink plans. To obtain a signal on such a short-range technology, they may have to be camped in the garden like guests of some outdoor Internet café, if not unnervingly pressed up against the walls like a zombie herd. Their usage will degrade your satellite broadband service because the backhaul is shared. But don’t expect Musk to share revenues.
“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?”
The alternative of an indoor small cell with a power envelope of 1 watt or less would limit coverage to the area of the home, according to Lum, and therefore would be useless as a mobile network technology. SpaceX President Gwynne Shotwell used the word “femtocell” to describe the technology that her company plans to integrate with its Starlink receivers. Despite sounding more like a feminine hygiene product, these femtocells were initially conceived as small cells for providing a 3G boost indoors. “You would need five of them to cover a major house,” said Lum.
Starry, Starry night
A similarly dubious assessment comes from Paul Rhodes, a wireless network expert with advisory group Real Wireless. He is dismissive of suggestions that SpaceX could develop some kind of advanced 4G and 5G technology for installation at customer homes. “You need 360 degrees, so you need three or four antennas if you want to keep the box the same size,” he said. “Then you’ve got far more RF [radiofrequency] capacity than you will have backhaul capacity.” Coverage would also be limited by the power supply of a Starlink unit, which measures only about 60 watts.
Whatever form the small cells take, heavy use of small cell services could problematically eat up backhaul capacity shared with the satellite-delivered broadband. “Who’s accepting that?” said Rhodes. “It’s a very tough thing to take a big part of the relatively modest bandwidth and break it off into third-party services.”
Given the laws-of-physics limitations, a Starlink network of small cells would probably have more holes in it than a hiker’s socks. Neither Lum nor Rhodes sees much likelihood of coverage in urban areas that enjoy mainstream and lower-cost alternatives. “It works best in rural because you’ve got fewer choices for fiber and so most customers are there. Where are your customers to repeat it in the city?” said Rhodes. “Well, they don’t exist because they’re not that stupid.”
If Starlink did target residents of high-rise multi-dwelling units (MDUs), Lum thinks it would potentially encounter similar challenges to those once faced by Starry, which tried using rooftop antennas to provide terrestrial broadband services. “It becomes a Starry issue all over again, where you have to hit the rooftop and then see through cable into the building,” he said.
Starry filed for bankruptcy protection in 2023 before its assets were acquired by Verizon earlier this year. Physical constraints are likely to prevent Starlink from directly serving an individual home within an MDU, Lum reckons. “No MDUs are going to allow this because you would have to hang a dish off the balcony.”
This absence of many Starlink customers in urban areas of the US is going to make competing against AT&T, T-Mobile and Verizon extremely difficult for SpaceX, according to Amit Nagpal, a spectrum expert and partner with Aetha Consulting. “Starlink dishes are mostly in rural and suburban areas,” he said. “Yes, there are a few in urban areas, but not that many, and that’s the complete opposite to the traffic profile on a mobile network.”
Even if Starlink can develop a relatively sophisticated small cell, and attract millions of customers, the likely escalation in subscriber costs could badly hurt its popularity. “If you’re starting to effectively use your customer terminals as basestations, what happens when they get their electricity bill for this and suddenly it’s gone up by four?” said Nagpal. “Can you imagine the outcry?”
Not such a genius after all
If Musk’s scheme had been concocted by a relative unknown, it would undoubtedly have received even harsher criticism. Musk’s entrepreneurial reputation earns him greater leniency, and there are observers who seem to think he can perform some small cell sorcery and do for that technology what his low-Earth orbit constellations have done for satellite. Don’t write off the engineering talent within SpaceX, said analysts at MoffettNathanson in research previously covered by Light Reading.
Yet the small cell plan itself seems to mark a rare Musk climbdown, a recognition that direct-to-device (D2D) satellite technology is not as good as he previously claimed it would be. On a podcast last September, he notably talked about using AWS-4 spectrum bought from EchoStar to pipe high-speed connectivity from Starlink satellites to smartphones. If Musk still expected this to be doable and competitive, why would he need small cells in, of all places, rural areas – supposedly the sweet spot for satellite?
“The physics is that your phone wants to talk to the strongest signal it sees, and the strongest signal is always going to be the one on Earth a mile away or a kilometer away rather than 400 miles up in the sky,” said Lum. “Nothing will ever make the signal from the satellite stronger, and the signal from the satellite will never get into your home.”
Nevertheless, SpaceX’s Shotwell might not be wrong when she says she expects “to acquire quite a few of their customers” in reference to AT&T, T-Mobile and Verizon. Of Starlink’s 12 million international customers at the end of June, some 3 to 4 million were estimated to be in the US. Any that rarely leave their homes and are unconcerned about coverage outside their own community might be attracted to a Starlink cellular service, especially if SpaceX absorbs the additional network equipment costs. Even a target audience of just a few hundred thousand subscribers would represent an unwelcome potential loss for the main terrestrial networks.
The idea, then, may be to exert enough pressure to force those companies into a mobile virtual network operator (MVNO) deal that they have so far appeared to resist. “If they had a possibility to get an MVNO, then it would be an interesting Phase 1 because it is easy to get those existing Starlink consumers to become Starlink mobile consumers,” said Lum. “They’re already buying your stuff, so it’s not hard to give them that. The problem is how are you going to market that? You’re going to market it as it only works in your house, and you can never leave it.”
Rhodes agrees that SpaceX may be trying to convince the telcos it poses a credible threat. “I would say they’re going to try to do some kind of deal with one of the carriers, and they just want to say we’ll go it alone if you don’t do it,” he said. “The point of this is to prove that they can do it.”
A catalyst might have been the dwindling interest of AT&T, T-Mobile and Verizon in Musk’s wholesale offer of D2D services. “There isn’t that much revenue there, so fundamentally it’s not worth the operators paying a huge amount to have exclusive access,” said Nagpal. “They’re probably pushing quite hard on price and his way of pushing back is to say if you don’t play ball I’ll launch a network against you.”
But the senior executives of the big telcos appear equally unpersuaded of the Musk challenge. The CEO of Deutsche Telekom, which owns T-Mobile, sounded full of contempt for the technology identified by Shotwell when he recently addressed analysts about the German operator’s second-quarter results.
“Yesterday, we heard something about femto,” said Timotheus Höttges, according to Deutsche Telekom’s own transcript of the call. “Guys, we had a femto discussion years ago, remember that?” It was dismissed “for good reasons from a technology perspective,” he added, saying that “some of these aspirations are very far-fetched and very conceptual.”
If Musk can make it work, a mass-market phone service could feasibly generate billions of dollars in new revenues for SpaceX. Despite its market cap of $1.85 trillion, which makes it worth more than nine times as much as T-Mobile, it generated sales of just $12.5 billion and suffered a net loss of $5.5 billion for the first half, versus the telco’s revenues of $46 billion and net profit of $5.7 billion over the same period. Money would be useful in funding travel to Mars, something Musk previously expected to happen by 2021 at best and 2031 at worst. Without a better telecom plan, the red planet will look like an even more distant speck than it already does.
