Amazon, known for its cut - throat tactics, has clashed with Elon Musk's Starlink.
Amazon has made another move.
This time, it acquired the satellite communication company Globalstar for tens of billions of dollars and continued the latter's partnership with Apple, directly entering the communication capability of "direct satellite connection for mobile phones".
This makes it hard not to recall the competition between Jeff Bezos and Elon Musk over space many years ago. It was thought that the story was almost over, but now it seems far from the end.
Instead of competing directly in the number of satellites, Amazon chose to start from the more fundamental D2D (direct device connection) capability, trying to break through from the communication mode itself.
If we look at this acquisition in a broader context, it is also in line with Amazon's consistent approach: not rushing to confront head - on, but entering from the periphery and gradually encircling. Its recent series of actions in the AI field are also repeating the same logic.
01
Amazon Acquires Globalstar
Let's first take a look at this acquisition.
Amazon's offer was very sincere. The acquisition price was $11.6 billion, 40 times Globalstar's revenue last year. Globalstar shareholders could choose to receive $90 in cash per share or 0.321 shares of Amazon stock. The upper limit of cash payment was 40% of Globalstar's shares.
On the surface, Amazon's acquisition of Globalstar is about buying a satellite company. However, if it is only understood as "buying dozens of satellites", the real value of this deal will actually be underestimated.
The most valuable thing about Globalstar is the hard - to - replicate spectrum resources it holds and the direct device connection capability that has been proven around this spectrum.
Globalstar doesn't have a large number of satellites, only a few dozen, which are continuously being updated. However, it owns a spectrum called S - band/L - band, which is a crucial resource for "direct satellite connection for mobile phones", scarce and with high regulatory thresholds.
Spectrum is allocated globally. It's not something you can buy just because you have money. US federal documents show that Globalstar is the only license holder of the 2483.5–2495 MHz spectrum for MSS (Mobile Satellite Service Spectrum). Earlier FCC documents also described it as the sole operator of the single MSS system in this frequency band.
Globalstar's satellite network is mainly for stable, low - data - volume communication that can directly connect to mobile devices, that is, Direct - to - Device (D2D).
This technology no longer requires devices to connect to ground cellular base stations, so it is crucial for supporting emergency services and providing connections in areas with insufficient cellular network coverage.
Globalstar's star customer is Apple.
For example, the satellite - based safety features such as "Emergency SOS" and "Find My" on Apple's iPhone and Apple Watch are supported by Globalstar's D2D technology.
Apple is not just a customer of Globalstar.
In 2024, Apple invested about $1.5 billion in Globalstar to support the expansion of its iPhone communication services. This deal also gave Apple about 20% of Globalstar's shares.
Globalstar also said at the end of last year that with Apple's support, its new network under development would expand its satellite fleet from the current approximately 24 to 54, including a small number of backup satellites.
While acquiring Globalstar, Amazon continued the cooperation with Apple.
In other words, through this deal, Amazon not only obtained valuable spectrum assets, opened up the D2D mobile phone direct connection scenario, but also had a major customer like Apple from the start.
02
Is the Space Battle between Bezos and Musk Not Over?
Amazon's goal is obvious: to strengthen its satellite business, which is still in its infancy, and catch up with Elon Musk's Starlink. Amazon's move comes at a time when SpaceX is advancing its IPO plan.
Musk's Starlink has tens of thousands of satellites and currently serves more than 9 million users globally. SpaceX has been deploying Starlink satellites at an extremely fast pace, launching dozens at a time, creating the world's largest satellite constellation. Starlink accounts for about 50% to 80% of SpaceX's revenue.
Amazon lags far behind, with only more than 200 satellites at present.
Amazon has been promoting its own low - orbit satellite project, formerly known as Project Kuiper and now more clearly referred to as Amazon Leo externally. It has obtained a license from the US Federal Communications Commission (FCC) to deploy 3,236 low - orbit satellites. According to the original regulatory schedule, Amazon needs to deploy half, that is, 1,618, by July 2026, and the rest by July 2029.
However, the actual progress is not that fast. The latest Reuters report mentioned that Amazon has only launched about 243 satellites so far, far from the mid - 2026 half - deployment target. A report from The Verge a few days ago had a similar figure, stating 241. There is a slight difference between these two figures, but they point to the same fact: it is very far from the original mid - term threshold set by the FCC.
So Amazon has applied to the FCC for an extension. Public materials show that it hopes to postpone the original mid - term deployment deadline of July 30, 2026, to July 30, 2028.
In other words, Amazon has admitted that it cannot meet the original plan at the current pace.
In terms of commercialization, Amazon's new schedule is to launch Leo commercial services in mid - 2026. Andy Jassy recently said this publicly, and some media also mentioned that Amazon has conducted a corporate preview and signed a number of customers or partners in advance, including aviation, operators, and government - related customers.
What really holds back the progress is mainly the launch capacity and the supply - chain rhythm. Amazon's problem is not only about building satellites but also includes the shortage of rocket launch resources, manufacturing delays, and the lack of a mature and high - frequency in - house launch system like SpaceX. Even Blue Origin cannot meet the launch demand. To make up for the progress, it has to rely on third - party rockets, including SpaceX's Falcon 9.
Although Amazon's founder Jeff Bezos has long retired from Amazon, his "space battle" with Musk has never ended. From the disputes between Blue Origin and SpaceX a few years ago to the current low - orbit satellite battle between Amazon and SpaceX, the rivalry will continue.
It's quite subtle that Amazon chose to enter from D2D. You know, Musk's Starlink mainly provides high - speed broadband services. Although it has cooperated with telecom operators such as T - Mobile to develop D2D services, it has not "dominated the market". Amazon has lagged behind Starlink in the satellite broadband field. Acquiring Globalstar will enable it to catch up in the D2D spectrum layout and achieve a leap in D2D deployment.
03
Amazon's Ambition
If the acquisition of Globalstar is more of a communication deal, looking further ahead, we can find that after 2026, Amazon has also been very active on another more - watched battlefield, which is AI.
Interestingly, just as Amazon's low - orbit satellite business seeks a breakthrough from D2D, Amazon's AI business also has its own characteristics.
A few days ago, an internal memo from OpenAI was exposed, and an interesting detail was that Amazon was repeatedly mentioned.
In the memo, Denise Dresser, OpenAI's new chief revenue officer, emphasized that the cooperation with Microsoft has laid a foundation for OpenAI. At the same time, she also said bluntly that this long - term binding also restricts OpenAI from serving enterprise customers "where they are", and for many enterprises, that place is Amazon Bedrock.
She also mentioned that since the cooperation was announced at the end of February, the relevant demand has been "astonishingly high". In other words, in OpenAI's own statement, Amazon has become an increasingly important enterprise distribution channel, far more than just a new "big sponsor".
Less than two months ago, Amazon and OpenAI just announced a multi - year strategic cooperation. Amazon will invest $50 billion in OpenAI. AWS will become the exclusive third - party cloud distribution provider for OpenAI's enterprise platform Frontier. Part of OpenAI's capabilities will be provided to enterprise customers through Amazon Bedrock. The two sides will also jointly develop a "runtime environment with state memory" to allow the model to continuously retain context, task status, and historical information during operation, rather than starting from scratch every time it is called.
Moreover, Amazon is not betting on a single line.
It invested in Anthropic earlier and more deeply.
In 2023, Amazon first announced an investment of up to $4 billion in Anthropic. By November 2024, this cooperation was further expanded. Anthropic officially confirmed that Amazon's total investment would increase to $8 billion. At the same time, AWS became its main cloud and training partner, and Amazon continued to maintain a minority - shareholder status. That is to say, in Anthropic, Amazon is betting on equity, cloud, and self - developed chip ecosystem; while in OpenAI, it is betting on larger - scale capital, enterprise distribution, and platform binding. Although the two lines seem different, the underlying logic is the same, which is to make AWS an infrastructure platform that leading model companies cannot bypass.
If this memo accidentally reminds the outside world of something, it is probably that Amazon's role in the AI track is no longer the "old - fashioned giant with low presence" in many people's impressions.
In the past, the outside world was more likely to focus on model companies such as OpenAI, Anthropic, Google, and Meta. Amazon always seemed to be standing in the back row.
But looking back now, it has actually quietly occupied several very key positions: it blocks the entrance of enterprise customers through AWS, undertakes model capabilities through Bedrock, extends to the chip and computing - power bottom layer through Trainium, and at the same time deeply binds itself to several of the most important foundation - model companies with capital.
It may not always be the company that makes the most noise, nor always at the center of public opinion.
While the outside world is still staring at whose model is stronger and whose product is more popular, Amazon has quietly strung together satellites, cloud, chips, enterprise distribution, and model cooperation into a larger map.
When this map is truly formed, people may realize that the contest between Bezos and Musk may never have been limited to space.
This article is from the WeChat official account "Zimu AI", author: Xiaojinya, editor: Wang Jing. Republished with permission from 36Kr.