Starlink’s disruption of the space industryby Jeff Foust
|
| “That would be a big step, to have more than zero in the ‘not bankrupt’ category,” Musk said of Starlink in 2020. |
Starlink was an ambitious bet for SpaceX. The company believed it could leverage its experience in launch in telecommunications, using vertical integration and mass production to create a satellite constellation that would offer broadband services for a competitive price. If successful, it would allow SpaceX to enter a far larger market than launch, generating significantly more revenue to satisfy investors and fuel founder Elon Musk’s visions of sending humans to Mars.
It was also a dangerous bet based on the experience of past ventures that tried to develop similar systems, which either went through bankruptcy reorganization or never got off the ground. A quarter-century ago, for example, Teledesic proposed a constellation of as many as 900 satellites—a mind-boggling number at the time—and had the backing of Bill Gates and Craig McCaw, but went out of business before launching a single operational satellite.
Musk was keenly aware of that when he spoke at the Satellite 2020 conference in March 2020, after the company had launched about 300 satellites. “That would be a big step, to have more than zero in the ‘not bankrupt’ category,” he said then. (A few weeks after he spoke, OneWeb, in the process of deploying its own constellation, filed for Chapter 11 bankruptcy protection.)
So far, that bet has paid off. SpaceX has launched more than 6,500 Starlink satellites to date, according to statistics maintained by Jonathan McDowell. Of those, more than 6,050 are still in orbit (the 60 satellites launched in May 2019, called v0.9 Starlink satellites, are among the 450 that have deorbited.) Starlink is not only far and away the largest satellite constellation in service, it also constitutes the majority of all operational satellites in orbit today.
Starlink is arguably one of the biggest and most disruptive developments in the space industry in the last five years. That is an honor often reserved for SpaceX’s launch achievements, including reuse of boosters and payload fairings. That reuse is essential to Starlink’s business model: the satellite constellation would be far more expensive if the company had to build a completely new Falcon 9 for each batch of Starlink satellites.
But the reverse is also true: SpaceX’s reusability quest depends on Starlink. The company’s ability demonstrate rapid reuse of boosters requires a demand for launches of them that has come primarily from Starlink. Last year, when SpaceX performed 96 Falcon launches, 63 of them were devoted to Starlink.
| Without Starlink, SpaceX would be a very different launch company than it is today, lacking the forcing function provided by the constellation. |
That high demand from Starlink has allowed the launch side of SpaceX to refine its processes, reducing the time between launches and increasing the number of reflights of boosters (the current record is 21 flights.) That gives SpaceX the ability to accommodate not just more Starlink missions but also missions from other customers, an essential capability in a launch market where supply has been constrained by development delays, launch failures, and geopolitics.
In short, without Starlink, SpaceX would be a very different launch company than it is today, lacking the forcing function provided by the constellation. That will likely also hold true as SpaceX seeks to scale up launches of Starship vehicles, which will be, on many of its early missions, carrying larger Starlink satellites.
Competing against Starlink
Starlink has affected much of the rest of the satellite industry even more. Traditional geostationary orbit (GEO) satellite operators were seeing declining demand for their services before Starlink, as demand dropped for television services faster than it rose for data services. The rise of Starlink, though, disrupted plans by those operators to rely more on broadband services as SpaceX attracted customers with high-bandwidth offerings at latencies far lower than what GEO systems could achieve.
A few years ago, GEO operators started adopting a new buzzword: “multi-orbit.” They said they would combine their GEO satellite fleets with satellites in LEO and medium Earth orbit (MEO) in an effort to combine the best of both worlds. Those operators took different approaches: Eutelsat bought OneWeb, SES continued to build out its O3b constellation in MEO, and Telesat proposed its own LEO constellation, Lightspeed.
Intelsat, the other major GEO operator, was slower to develop its own multi-orbit strategy as it went through a Chapter 11 reorganization, but as recently as last fall said it was looking into developing its own MEO constellation.
On April 30, though, SES announced it reached an agreement to acquire Intelsat for $3.1 billion after more than a year of on-again, off-again discussions about a potential merger. The combined company would have $4.1 billion in revenues, based on projections for this year, though the deal is not expected to close until the second half of 2025.
One factor in that deal, executives said, was reducing costs for their multi-orbit strategy. “It’s about optimizing the future of multi-orbit satellite investments and fleets,” SES CEO Adel Al-Saleh said in a call with analysts. “We just don’t need to spend as much money as we were spending separately. The combination will give us the opportunity to reduce that.”
Costs have been an issue for other satellite operators seeking their own constellations to respond to Starlink. Telesat shifted manufacturers for its Lightspeed satellites last year from Thales Alenia Space to MDA, seeking smaller and cheaper satellites after both delays by Thales Alenia and struggles by Telesat to raise the several billion dollars needed to build and launch it.
Eutelsat has a completed constellation with the OneWeb system it acquired, but is already planning a second-generation system. It has disclosed few details about that system and how it plans to finance it, though, other than it will select a manufacturer for it this summer. Company executives added in an earnings call in May that the company doesn’t plan to launch another GEO satellite until 2026.
 An illustration of a set of Starlink satellites prior to their deployment. (credit: SpaceX) |
Starlink and the military
Starlink’s influence goes beyond business competitors to geopolitical ones. The system has become a key element of Ukraine’s defense against Russia, providing communications services when terrestrial or other satellite systems were unavailable. But the unusual approach of SpaceX working directly with Ukraine’s military has led to debates and disputes about the use of Starlink.
| When SpaceX started deploying a constellation of thousands of satellites, many raised concerns about space safety. With that many satellites in narrow orbital bands, collisions with debris or active satellites seemed inevitable. |
“We know the military is using them for comms and that’s OK. But our intent was never to have them use it for offensive purposes,” Gwynne Shotwell, president of SpaceX, said last February, adding that the company had taken steps to limit offensive usage. Water Isaacson’s biography of Musk, published last fall, claimed that SpaceX had not enabled Starlink coverage off the coast of Russian-occupied Crimea, disabling Ukrainian drones sent to attack Russian ships docked there, although Isaacson later attempt to “clarify” that claim with mixed success.
It shows, though, that power of Starlink, and of companies like SpaceX, to affect military operations. SpaceX has offered a military version of Starlink satellites, dubbed Starshield, that could offer imaging and other capabilities in addition to communications. A Falcon 9 launched May 22 from Vandenberg Space Force Base in California carrying the first set of satellites for what the National Reconnaissance Office called “proliferated systems featuring responsive collection and rapid data delivery.” The NRO did not release other details about the launch, although it is rumored to be part of the Starshield program with contributions from Northrop Grumman.
NRO did not disclose how many satellites were on the launch, but later tracking data showed at least 21 satellites were placed into orbit, similar for a typical Starlink mission.
Avoiding apocalypse
Perhaps just as important as what Starlink has done is what it has not. When SpaceX started deploying a constellation of thousands of satellites, many raised concerns about space safety. With that many satellites in narrow orbital bands, collisions with debris or active satellites seemed inevitable. Astronomers also warned of the impact the constellation would have on their observations when that initial batch of 60 satellites launched five years ago created a brilliant “string of pearls” display in night skies in the weeks after launch.
Five years later, though, the worst-case scenarios for space sustainability have been avoided, at least for now. Starlink satellites routinely and autonomously maneuver to avoid conjunctions, and SpaceX has worked with other satellite operators to better manage operations. One example is an agreement between NASA and SpaceX to coordinate maneuvers, with SpaceX agreeing to move its satellites should any make close approaches to NASA spacecraft.
The company has also worked with astronomers on ways to reduce the brightness of its satellites, developing technologies ranging from visors to mirrors to keep sunlight from reflecting off parts of the spacecraft. Those efforts have not completely resolved the problem—Starlink satellites are still brighter than what many astronomers prefer—but they have helped at least mitigate the problem.
“To be clear, industry doesn’t have to play nice with us,” said Kelsey Johnson, an astronomer at the University of Virginia and president of the American Astronomical Society, during a session on satellite constellation interference with astronomy at the organization’s conference in New Orleans in January. “They have invested real time and real money and effort to working with us that they don’t have to do.”
 SpaceX has emphasized the ability of Starlink to connect people around the world, but analysts note that it remains primarily a “rich-world service”. (credit: SpaceX) |
Generating cash and growing fast
Starlink has had a clear impact on companies, governments, and others. But can SpaceX do all that without going bankrupt?
SpaceX, as a privately held company, does not release financial details about Starlink. The company recently noted more than three million people worldwide are using Starlink (which is not necessarily the same as three million paying subscribers) with the service now available in nearly 100 markets. Executives has suggested that the system is profitable, but have disclosed no specifics.
That’s left it up to outside analysts to make their best guesses on the economics of Starlink. The latest assessment came from Quilty Space, which rolled out its analysis of system earlier this month using a bottoms-up financial model it developed.
“We started out this whole exercise wanting to know whether it’s profitable. We think we’ve come to that answer pretty clearly, which it this business does generate cash and it is growing fast,” said Chris Quilty, president of Quilty Space, during a May 9 webinar to discuss the study.
| “It looks like Starlink will be the first megaconstellation to get to a cash flow positive place on its own accord,” Cadman said. |
The analysis estimated that Starlink will bring in $6.6 billion in revenue in 2024, with EBITDA (earnings before interest, taxes, depreciation, and amortization) of $3.8 billion. When subtracting capital expenditures, like manufacturing and launching satellites, Starlink will produce free cash of about $600 million this year.
“With this, the business is now self-sustaining. That’s the first major test of whether Starlink will be around here for the long haul,” said Justin Cadman, chief financial officer of Quilty Space.
The analysis credited that to a couple factors. “You’ve got to put up some crazy subscriber growth,” Quilty said. The company achieved that, seeing growth far greater than other satellite Internet companies like Hughes and Viasat, whose subscriber numbers have gone down somewhat since Starlink’s introduction.
“They’ve hit a spot in the market where there is, in fact, a willingness to pay,” Cadman said, calling Starlink a “rich-world service” for customers in wealthier countries that have limited connectivity options. “This is not connecting the unconnected. That’s a gap that Starlink probably will not be serving in a substantial sort of way any time in the near future.”
Another factor is keeping costs low. “We estimate that Starlink has been able to do an incredible job at keeping their spacecraft costs lower than basically any industry precedent,” said Caleb Henry, director of research at Quilty Space. The initial “V1” satellites cost about $200,000 each, he estimates, with the larger “V2 mini” satellites now being launched coming in at around $800,000.
“It looks like Starlink will be the first megaconstellation to get to a cash flow positive place on its own accord,” Cadman concluded.
Competitive pressures
Will others follow Starlink into profitability? Other constellations are in development like Telesat’s Lightspeed and a future OneWeb constellation, as well as Rivada Space Networks and its planned 600-satellite constellation. (Both Rivada and Telesat have contracts with SpaceX to launch their satellites.)
Perhaps the biggest competitive threat comes from Amazon’s Project Kuiper, which plans to launch more than 3,000 satellites to provide broadband services. It launched two prototype satellites in October and, last week, the company announced it completed testing and would deorbit them ahead of launching its first operational satellites later this year.
Amazon, like SpaceX, is building its satellites and terminals in-house, leveraging vertical integration to lower costs. Amazon can also bring its major marketing and distribution capabilities to bear to promote Kuiper. However, it is reliant on outside launch providers, including vehicles that either have just made their first launches—ULA’s Vulcan Centaur—or have yet to fly at all: Arianespace’s Ariane 6 and Blue Origin’s New Glenn. (Amazon does have some of the last Atlas 5 rockets from ULA to launch some initial satellites, and even bought a few Falcon 9 rockets from SpaceX as well.)
| “If you want to be competitive in there, then you have to own your own rocket and build your own satellites,” Beck said of Rocket Lab’s constellation ambitions. “We’re just marching very methodically towards that, step after step.” |
Starlink faces other threats. Its use in Ukraine led to threats from Russia of attacking those satellites. Earlier this month, the New York Times reported that Russia had, after two years, found ways to interfere with Starlink services in Ukraine, with Ukrainian forces reporting severely degraded services.
The long-term economics of Starlink remain uncertain, the Quilty Space analysis noted, given questions about whether the company’s pricing is sustainable. “Starlink is on a path to delivering attractive returns on invested capital,” Cadman said, “but it’s not there yet.”
The Starlink approach itself, though, seems attractive to other companies. During an earnings call in February, Rocket Lab CEO Peter Beck started musing about operating a constellation, rather than just building and launching satellites for others, citing the much larger market for providing space services like communications.
Analysts on the call quizzed Beck about this proposed constellation, which the company had not discussed in detail before, although he offered few details about what it would do or when it would be developed.
However, he made clear that, if and when his company pursued a constellation, it would take a page from SpaceX’s Starlink playbook. “If you want to be competitive in there, then you have to own your own rocket and build your own satellites,” he said. “We’re just marching very methodically towards that, step after step.”
Note: we are now moderating comments. There will be a delay in posting comments and no guarantee that all submitted comments will be posted.