A closer look shows that Amazon and Starlink are chasing different markets: mass-market connectivity vs high-margin enterprise services
This week, Blue Origin hit a milestone — or thereabouts. The Jeff Bezos-founded rocket company had been noodling with the concept of a reusable booster for some time, to help lower spaceflight costs and up cadence.
On Sunday, the rocket company put the booster to test on the New Glenn rocket with an AST SpaceMobile satellite on board. 10 minutes after liftoff, the booster made an on-target touchdown, marking its first successful reuse — a feat that was met with great enthusiasm. However, the rocket later experienced upper stage failures, narrowly missing the correct orbit.
Although only partially successful, the mission was intriguing enough to reopen the Amazon Leo vs Starlink debate. With both being low-Earth orbit (LEO) players, over the past year, the theory that Leo is emerging as a direct competitor to Starlink has shaped into a convincing narrative, one that’s cemented by countless headlines and discussions in media about how the two titans are locked in a war to win the satellite race.
But, a closer look reveals some holes in that theory — and the argument starts to fall apart.
Apples to oranges
Starting with the basics, Starlink today commands a fleet of over 10,000 in-orbit satellites. Leo’s constellation, only some 235 satellites strong today, with 3,000 more planned in the first set, is no match to Starlink’s.
“If this is a symmetrical war for consumer broadband, Amazon is already mathematically dead,” Sebastian Barros, an independent analyst, who previously worked with companies like Google and Ericsson, wrote in an opinion piece on his Substack.
Let’s unpack that. As noted, with only a fraction of Starlink’s mega-constellation, Leo cannot possibly compete with Starlink on density. And without its own reusable rocket technology, the kind that Starlink’s parent company, SpaceX pioneered, Amazon is already on the back foot.
“[Amazon] executed a $10 billion commercial procurement for launches using expendable or partially reusable vehicles from ULA, Arianespace, and Blue Origin,” Barros notes. That’s a cost Starlink can dodge, all thanks to the 230-foot Falcon 9.
With its orbital-class reusability, the Falcon 9 has transformed rockets from single-use crafts to repeatable services, delivering tens of satellites in a single ride. It routinely launches Starlink satellites — and those for other companies for a price.
Barros observes that with the Falcon 9, Starlink has put a total of 1,000 satellites in orbit so far this year. That capacity will be amplified so much more with the upcoming Starship V3 that boasts capacity to haul 150 metric tons of satellite weight for only $10 million.
“By 2029, Starlink will be the “Android” of space, ubiquitous, massive, and built on a foundation of raw spectral scale,” Barros says.
Amazon on the other hand still does not have much to show for its launch capacity. “Amazon is held hostage by third-party vendors charging between $2,600 and $7,000 per kilogram. You cannot win a price war when your competitor rides for free, and you pay premium retail just to leave the atmosphere,” Barros argues.
Even in terms of earnings, the picture is deeply disparate. Although SpaceX’s annual revenues are not publicly known, people familiar with the matter told Reuters that the rocket company generated about $16 billion in revenue in 2025, with roughly an EBITDA (earnings before interest, taxes, depreciation, and amortization) profit of $8 billion. Starlink’s revenues account for 60 to 80% of this number.
Blue Origin’s numbers pale in comparison. Ars Technica reported that it is approximately “on the order of $1 billion a year.”
Busting the myth
So, knowing what its up against, what’s Amazon up to? “Amazon isn’t stupid enough to play SpaceX’s game,” Barros writes. According to him, Leo and Starlink are in very different plays: one in high-margin service, and the other, high-volume.
If you look at the companies’ low-altitude shells, the argument starts to make sense. In their third orbital shell, Leo’s satellites sit at 590 km to 630 km from Earth, whereas Starlink’s hover at 340 km to 550 km.
Being at a higher altitude, Leo’s satellites cover greater areas with fewer assets. Conversely, Starlink’s closer proximity to Earth requires it to maintain a much larger fleet to cover more areas and maintain line of sight at all times.
The catch is, less physical distance and higher density equals greater capacity from spectral overlap. This allows Starlink to support millions of users across rural, and urban areas in some cases. Not so much the case for Leo.
“A single satellite at Amazon’s 600 km altitude illuminates a geographic footprint the size of a large state,” Barros writes, “If 50,000 rural households under one beam attempt to stream high-definition video simultaneously, demand would reach 1 Tbps, causing the network to collapse.” That naturally narrows down the market segment for Leo.
With physics not being on its side, the only way Leo can get on par with Starlink is by addressing a niche which offers higher margin, if not volume, and that’s what it increasingly seems to be doing.
As Barros writes, “the “head-to-head” narrative is a myth.” Starlink is in the B2C (business-to-consumer) game or as Barros calls it, the “consumer-volume” game. On the fixed broadband side, the company targets millions of rural households and enterprises with its $120 monthly subscription-based broadband service — a sizable portion of the market, left open by telcos. On the mobile side, its addressable market includes direct-to-device (D2D) consumers, such as hikers, emergency workers, etc., and commercial airlines. The company has contract with over 31 airline carriers around the world for in-flight Wi-Fi service.
Leo, on the other hand, is going for the B2B (business-to-business) market: high-end customers in the enterprise and cloud niche — a very familiar market because of AWS. These include enterprises, maritime companies, government agencies, and logistics firms that require guaranteed SLA-backed connectivity. Leo is also expanding footprint in the commercial airlines sector with deals with carriers like JetBlue — and most recently, Delta — to provide Wi-Fi services on flight.
“Having watched the 5G monetization failures of terrestrial operators, Amazon is avoiding the trap of squeezing $120 a month out of a suburban household. Instead, they are building a closed-loop, deterministic SLA router for corporate campuses, automated supply chains, and tier-1 operators looking for high-margin backhaul,” Barros explains.
His 2029 prediction for Leo is just as interesting. “By 2029, Amazon will not be an ISP; it will be an orbital extension of the AWS data center,” he writes.
The company is tapping into its techco privilege to curve its place in the market. Amazon owns AWS — the largest cloud infrastructure in the world — and its nifty satellite fleet is a physical extension of that, and path to the bigger orbital play. By integrating the AWS infrastructure with Leo’s network, Amazon is set to replicate its cloud-based transport network in space.
But, while Starlink evidently remains the bigger, stronger, and faster player in the race, some are willing to bet big on Leo. Ron Westfall, VP and practice lead of networking and infrastructure, HyperFRAME Research says, “The Amazon Leo proposition holds solid competitive prospects due to its deep integration with AWS and the massive existing retail logistics network. While Starlink has a significant first-mover advantage with thousands of satellites already in orbit, Amazon can undercut competitors by bundling satellite connectivity into broader enterprise and Prime consumer ecosystems.”
Westfall argues that Amazon’s deep tech foundation provides Leo an unexpected edge over Starlink. “The competition will likely shift from basic connectivity to integrated services, where Amazon’s ability to offer low-latency links directly into its global cloud infrastructure provides a distinct value proposition that Starlink can be hard-pressed to directly counter.”
Additionally, the deal with Apple which came tagged with Amazon’s recent $12 billion acquisition of Globalstar gains it a straight entry into the D2D space, giving its upcoming next-gen services, coming in 2028, an on-ramp and a soft landing.
However, to settle the Leo vs Starlink debate for now, it’s safe to say that the companies are going in different directions, focusing on bringing differentiated services to markets where the economies are most compelling. As Barros puts it, “They are no longer competing for the same dollar; they are defining two different layers of the modern telecommunications stack.”
.
