Amazon's $11.6 Billion Globalstar Gambit: Inside the Satellite Acquisition That Reshapes the Space Internet War

When a hyperscaler that already runs its own rocket company, its own satellite factory, and its own constellation writes a roughly $11.6 billion check for a much smaller incumbent with only two dozen working spacecraft, it is not buying satellites. It is buying a shortcut. Amazon's April 14 agreement to acquire Globalstar, announced on the company's own news site, does that twice over — collapsing years of spectrum licensing, Apple's emergency-SOS backbone, and a global mobile-satellite-services authorization portfolio into a single transaction that immediately pulls Amazon Leo out of a distant third place in the direct-to-device race and parks it, on paper, right behind Starlink.

What Was Actually Sold

The headline number is large but the structure is more revealing than the sticker. Globalstar shareholders will get a choice for each share held: either $90.00 in cash, or 0.3210 shares of Amazon common stock with a value capped at $90 per share, per Amazon's announcement. Cash elections are capped at 40% of total shares, with any excess automatically converted to Amazon equity. That is the mechanical shape of a $11.57 billion purchase price, as TechCrunch summarized it, and it is the reason the deal looks different from a textbook strategic acquisition.

An all-cash bid at this scale would have transferred the entire risk of Globalstar's deployment and regulatory timelines to Amazon on day one. A capped-cash, forced-rollover structure does something more interesting: it makes a large fraction of former Globalstar holders into Amazon shareholders by default. They now have an incentive to want the combined platform to succeed, not just the standalone operator. On a deal where the regulatory review and satellite milestones stretch well into 2027, that alignment matters.

The second tell is the earnout. Amazon disclosed a potential $110 million downward adjustment tied to what it calls certain HIBLEO-4 replacement satellite milestones. That is the program that is supposed to refresh Globalstar's aging low-Earth-orbit fleet. Amazon is, in essence, writing a price protection into the contract that pays Globalstar more if it delivers the new satellites on time and less if it does not. A buyer that cared primarily about the operating company would not bother; a buyer that cared primarily about the spectrum and the new birds going up on schedule absolutely would.

Why This Is a Spectrum Deal, Not a Satellite Deal

Globalstar's operating fleet is small. The Register counted 24 operational satellites included in the acquisition, and TechCrunch noted roughly fifty more pending from a launch program that had already signed SpaceX as a partner for 2026 replacement launches. Twenty-four is a rounding error next to Starlink's deployed fleet, which The Register placed at approximately ten thousand satellites. On raw spacecraft, this is not a fair fight and the transaction does not make it one.

What the transaction does change is what Amazon owns on the ground. Globalstar holds exclusive access to what is known as Band 53, a sliver of S-band spectrum from 2483.5 to 2495 megahertz that the company describes as optimized for high-performance, low-latency, interference-free connectivity. Amazon's own announcement frames the acquisition in broader terms, describing MSS spectrum licenses with global authorizations as part of what it is buying. Those two framings — a narrow boutique band on one side, and a worldwide mobile-satellite-services regulatory footprint on the other — describe the same asset from different angles. Either way, the unique value is on the frequency chart, not in the operating bus.

That is because spectrum is the scarcest resource in direct-to-cell. Satellites can be manufactured, launched, and replaced on a rolling basis; anyone with enough capital can put thousands of birds into orbit. Spectrum in which a handheld phone will reliably talk to something moving overhead at seven and a half kilometers per second is not manufacturable. It has to be licensed, usually once per country, often decades in advance. Band 53 gives Amazon Leo a narrow but globally harmonized slice of spectrum that works with existing and future smartphone chipsets, and the broader MSS licenses give it a pre-built regulatory footprint in jurisdictions where a newcomer would have spent years lobbying for access.

The Apple Equation

The second reason this deal is not about satellites is a company with its logo etched on the back of a billion consumer devices. Apple is the largest single outside stakeholder in Globalstar, holding a roughly 20 percent stake it took in 2024 as part of a $1.5 billion investment. Globalstar today is the satellite backbone behind the emergency-SOS feature that activates when an iPhone or Apple Watch loses cellular signal; TechCrunch notes the current service menu includes emergency text messaging, requesting roadside assistance, and location sharing on iPhone 14 and later devices. That service is now about to run on Amazon infrastructure.

Apple is not receiving an Amazon ownership stake in exchange for its Globalstar shares. It is simply cashed out alongside other holders at the $90-per-share reference price, as The Register reported. The equity bridge between the two companies dissolves. What replaces it is a separately negotiated commercial agreement: Amazon's press release describes Amazon Leo as positioned to power satellite services for iPhone and Apple Watch, including continued support for Emergency SOS via satellite.

This is an underappreciated strategic shift. For two years, Apple's satellite roadmap ran through a public company in which Apple could functionally block any change of control by withholding its stake. By selling at $90 a share rather than rejecting or counter-bidding, Apple has effectively chosen to accept a supplier relationship with Amazon over continued ownership of the operator. Several reading lines follow from that choice. Apple presumably negotiated long-term service, capacity, and performance terms as part of the deal architecture; otherwise a 20 percent holder does not quietly fold. Amazon, for its part, is no longer selling spare satellite capacity in a market where its main potential anchor tenant also sits on its supplier's cap table. The supplier relationship is cleaner, and it is now on Amazon's balance sheet.

From Kuiper to Leo, Before the Deal

To understand why Amazon would spend this much on a spectrum and anchor-tenant play, it helps to see where Amazon Leo actually sat on the eve of the announcement. Amazon's broadband-satellite effort, rebranded from Kuiper, had already put somewhere north of 200 satellites into orbit, with The Register's more precise count putting the figure at 241. The full constellation is planned at 3,200 or more satellites, with Amazon's own phrasing describing thousands of advanced satellites in low Earth orbit sized to serve hundreds of millions of customer endpoints around the world. Those numbers are large in absolute terms. They are still small relative to Starlink.

More importantly, Amazon Leo was designed first as a broadband service — a bent-pipe replacement for Ka-band backhaul with aviation antennas hitting up to one gigabit per second down and four hundred megabits per second up. That is a market that sells to airlines, enterprises, and rural internet service providers. It is not the consumer direct-to-cell market, where the bet is that an unmodified phone in a user's pocket can punch a signal through the atmosphere to a satellite and back. Those are adjacent businesses that share a constellation, not the same business.

Amazon had publicly committed to entering direct-to-device but had not anchored that commitment to specific spectrum, a specific smartphone installed base, or specific launch dates. Buying Globalstar collapses all three into a single afternoon. TechCrunch notes that Amazon expects to begin offering direct-to-device services later in 2026 on the existing Globalstar fleet and plans to launch its own next-generation direct-to-device system in 2028, which Amazon's release characterizes as offering faster speeds and better performance than legacy systems through improved spectrum efficiency. That is roughly the same timeline on which Delta Air Lines is committed to begin deploying Amazon satellite Wi-Fi across more than five hundred aircraft.

A Three-Player Race the FCC Wants

The deal lands in front of a regulator that seems to want it approved. FCC Chair Brendan Carr, speaking on CNBC the same day the acquisition was announced, said the commission is very open-minded to it, per Benzinga's recap of the interview. Carr also framed the transaction as consistent with U.S. leadership in direct-to-cell technology and added, "From my perspective, it's all about the consumer and the consumer here is a big, big winner." His structural view of the sector was just as direct: "We think that a healthy, three-player direct-to-cell market can be really good for the country."

The three players, per Benzinga's summary of the FCC chair's framing, are SpaceX with Starlink, Amazon with Leo, and AST SpaceMobile, the publicly traded operator that has spent the last several years stitching together partnerships with terrestrial carriers to extend mobile coverage into orbit. That list is notable both for who is on it and who is not. Iridium, Viasat, and several national-carrier-aligned entrants are absent. Globalstar, obviously, disappears as an independent competitor. What Carr is describing is a policy preference for a short list of well-capitalized operators, each paired with a different commercial partner set: Starlink with T-Mobile, AST with AT&T and Verizon among others, and now Amazon with Apple.

For Amazon, the political dynamics are unusually favorable. Reducing a field of roughly half a dozen viable D2C entrants to three is usually the kind of move that invites antitrust scrutiny. In this case, the regulator leading the review has publicly said the three-operator outcome is the desired one. Carr added on traditional carriers that he sees satellite technology as a complement to, rather than a replacement for, terrestrial mobile networks — a framing that tends to ease concerns from incumbent carriers that might otherwise oppose a new entrant aggregating spectrum and anchor tenants.

The Competitive Geometry

Starlink is not sitting still. SpaceX's constellation dwarfs every alternative by an order of magnitude, and its direct-to-cell partnership with T-Mobile has been live and commercially launched for months. Starlink also continues to win aviation business against Amazon, with Virgin Atlantic, United Airlines, and American Airlines on its side of the ledger against Delta's Amazon commitment. The deal does not change any of that in the short term.

Where it does change the geometry is three years out. Starlink's D2C service today is fundamentally bandwidth-limited by the spectrum it has been able to lease from terrestrial partners. AST has been working with allocated mobile-carrier spectrum and a handful of enormous satellites to compensate for sheer radio-link physics. Amazon Leo, after closing, inherits a globally authorized MSS band that was built for this exact use case, plus an installed base of iPhones already provisioned to talk to it. In the specific sliver of the market defined as unmodified smartphone, any country, emergency and then everyday low-bandwidth services, Amazon has leapfrogged into a structurally different position than where it stood a week ago.

The installed base matters more than it sounds. Direct-to-cell at scale is constrained by what phones are actually in consumers' pockets. Getting a new band and a new signaling mode into a shipping smartphone takes years of coordination with chipset vendors and phone makers. Through the Apple relationship, a very large existing iPhone fleet is already, in effect, an Amazon Leo endpoint for emergency services. The incremental cost of upgrading those endpoints to richer services as Amazon's next-generation satellites come online in 2028 is a software and firmware problem rather than a hardware replacement problem.

What the Deal Does Not Solve

None of this guarantees execution. Amazon still has to launch most of its constellation, and Globalstar still has to deliver its HIBLEO-4 replacement satellites on the milestone schedule that the $110 million price adjustment is pegged to. The closing is not expected until 2027, and a regulatory review that involves spectrum consolidation across multiple jurisdictions is rarely linear even with a supportive chairman.

There are also questions the available disclosures do not answer. Neither Amazon's press release nor the day-one news coverage specifies a per-channel capacity for direct-to-device on Band 53, the details of the long-term commercial terms Apple negotiated in exchange for agreeing to be cashed out, or the degree to which Amazon's 2028 next-generation birds will depend on spectrum beyond the Globalstar portfolio. Those are the variables that will determine whether the second half of this decade looks like a true three-player market or another one-company dominant model with two very capable seconds.

Implications for the Rest of the Industry

The clearest immediate loser is any privately held satellite operator that was hoping to sell itself in a bidding process to one of the three strategic acquirers. Amazon has just taken itself off the buyer's list by going vertical. Starlink is captive to SpaceX and has never been an acquirer in the traditional sense. AST is the only remaining strategic, and it is already trading at a valuation that would strain a meaningful M&A program. Secondary and tertiary satellite operators with niche spectrum or narrow coverage footprints will now have to justify their standalone strategies in a world where the scale players no longer need them.

For Apple, the transaction reads as a clean operational simplification. It surrenders a minority position in a public operator in exchange for an explicit supplier relationship with a company that now owns the constellation, the ground infrastructure, and a much larger customer base to amortize the cost across. Apple's own service quality depends less on Globalstar's ability to finance its next launch and more on Amazon's willingness to keep the Apple lane of service provisioned. That is a different risk profile — arguably a better one — for a consumer-electronics vendor that does not want to be in the balance-sheet-management business of satellite operators.

For the consumer, the practical answer is that nothing visibly changes for at least another year, and probably two. The emergency-SOS flow on an iPhone will behave identically. After 2028, if Amazon executes, the same flow should work in more places, carry more bits, and potentially extend into richer consumer services. Whether any of that becomes a subscription product or stays bundled inside the phone purchase is the next question — and the one that will ultimately determine how much of the $11.57 billion price tag the market ends up validating.

Key Takeaways

• The headline is the $11.6 billion price; the real asset is Band 53 spectrum plus global MSS authorizations. Globalstar's 24 operational satellites are a footnote next to what the company holds in licensing.
• The deal structure — capped cash, forced rollover, and a $110 million HIBLEO-4 earnout — tells you Amazon cares most about the spectrum and the new satellites arriving on time.
• Apple sold out of ownership but locked in supply. The 20 percent stake converts to cash at $90 a share, replaced by a long-term commercial agreement that keeps Apple's satellite features running on Amazon infrastructure.
• The FCC is signaling it wants exactly the three-player market this deal produces. Brendan Carr's "we're very open-minded" comment and his "three-player direct-to-cell market" framing amount to regulatory cover.
• The competitive consequences play out in 2027–2028. The Starlink lead is intact today; the question is whether Amazon's spectrum advantage and iPhone installed base compresses that lead by the time Amazon's next-generation direct-to-device satellites launch.