How Amazon Leo Plans to Steal Starlink’s Crown

Credit: Shutterstock

Amazon eyes 2026 entry to a satellite internet market dominated by Starlink

For the better part of a decade, satellite internet and Starlink have been synonymous. SpaceX got there first, scaled fast, and built a network that now blankets most of the planet. But that monopoly is about to get tested.

Amazon’s long-rumored satellite project, recently rebranded from Project Kuiper to Amazon Leo, is not far. It’s flying. And Amazon is trying to launch it as soon as possible.

Here’s what that race actually looks like in 2026, and why Amazon’s approach is more interesting than that of its competitors.

Let’s start with the uncomfortable truth for Amazon: Starlink is miles ahead.

Starlink is the largest satellite constellation ever deployed. Amazon, meanwhile, is staring down an FCC deadline to launch half its constellation by July 2026.

So why is anyone betting on Amazon? Because the satellite internet market in 2030 won’t be decided by who launched first. It’ll be decided by who owns the customer relationship, the spectrum, and the device. And on those three fronts, Amazon just made some very loud moves.

Move 1: Amazon bought a competitor to skip a generation

In April 2026, Amazon announced it was acquiring Globalstar, a 30-year-old satellite operator with globally harmonized spectrum and an existing direct-to-device (D2D) network.

This matters more than it sounds. Starlink’s “direct-to-cell” service, which lets unmodified phones connect to satellites, has been the company’s most ambitious bet for mainstream relevance. According to industry reporting, the service has been seeing less usage than initially expected. Amazon, by contrast, just bought its way into the D2D business rather than building it from scratch, and it inherited the spectrum rights that make it work.

Amazon plans to launch its own next-generation D2D satellite system in 2028, integrated with the Leo broadband network. It’s the same playbook Starlink is running, except Amazon’s is bolted onto an existing operator with paying customers, regulatory approvals, and decades of operational experience.

Move 2: Apple switched teams

Apple signed a deal with Amazon Leo to power satellite features on iPhone and Apple Watch. Currently, Apple Emergency SOS via satellite (the feature that’s been quietly saving lives for more than three years) runs on Globalstar. With Amazon now buying Globalstar, Apple’s satellite layer is effectively becoming part of the Amazon Leo network.

Move 3: The aviation play

While Starlink has been racking up airline deals (Air France, Virgin Atlantic, Alaska Airlines, Air Canada, and Southwest are among the carriers signed on), Amazon Leo just landed one of the largest single-airline commitments in satellite internet history: Delta Air Lines, 500 aircraft, starting 2028.

Delta’s reasoning is interesting. The airline already runs much of its tech stack on AWS, and adding Amazon Leo will be just an integration. Amazon isn’t pitching Leo as “Starlink but cheaper.” It’s pitching it as a part of the AWS ecosystem.

For enterprise buyers like airlines, shipping fleets, oil and gas, and governments, that pitch has more influence than Musk’s “buy a dish, plug it in” approach.

What Amazon still has to prove

None of the above means Amazon wins. There are three real risks.

1. Amazon has 302 production satellites in orbit and needs roughly 1,600 to meet the FCC’s July 2026 milestone. It has filed for an extension, but missing the deadline could result in the loss of licensed spectrum.
2. Amazon contracted 92 launches across ULA, Arianespace, and Blue Origin, worth over $10 billion, and has since added Falcon 9 and additional New Glenn flights. But Blue Origin is still ramping up New Glenn cadence, and Vulcan has had its own delays. Starlink, meanwhile, owns its rockets. SpaceX can launch on its own schedule.
3. Starlink at 10 million subscribers isn’t just bigger but also learning faster. Every customer provides feedback on coverage gaps, terminal performance, and service quality. Amazon will spend its first two years building the experience that Starlink already has.

The real bet

Starlink wins on time-to-market, but the satellite internet market is still in inning two of nine. The actual prize isn’t 10 million broadband subscribers paying $120/month for a dish in their backyard. It’s:

• Direct-to-device connectivity for billions of phones
• Enterprise infrastructure integrated with cloud platforms
• Aviation, maritime, and automotive as default-connected categories
• Government and defense contracts that prefer non-Musk-aligned vendors

On three of those four, Amazon has a stronger position than SpaceX. However, the crown won’t change hands in 2026. Probably not in 2027 either. But by 2030, the situation will probably be different.