Are LEO Networks Aiming to Become the Global Cloud and AI Backbone?

Low Earth Orbit (LEO) networks have continued to capture headlines as operators race toward space supremacy.

SpaceX: After acquiring EchoStar’s North American S-band spectrum recently, SpaceX has set its sights on a new 15,000-strong LEO network to provide Direct-to-Cellular smartphone and Internet of Things (IoT) connectivity. This is an entirely new network and separate from the company’s existing Federal Communications Commission (FCC) filings for up to 42,000 satellites in LEO, which is still not fully approved, and evolves SpaceX’s network deployment ambitions to exceed 57,000 satellites.

Globalstar: Not to be deterred by SpaceX’s posturing, Globalstar has announced that it will use another next-generation network, HIBLEO-XL-1, which is expected to operate 3,080 satellites in LEO. This is in addition to the C-3 network with which Apple will cooperate. With Apple having a 20% ownership stake in Globalstar, deploying this new network is undoubtedly a calculated move to ensure Globalstar/Apple’s spectrum moat and technology stacks provide a competitive advantage.

Amazon Kuiper: Amazon’s global cloud backbone network in space, Project Kuiper, has also ramped up deployment activity. As of September 2025, 129 satellites had been launched, and the company is rapidly chasing its FCC milestone target of 1,600 satellites by July 2026. The company has also partnered with Australian Communication Service Provider (CSP) National Broadband Network (NBN) Co to bring LEO services to more than 300,000 customers by mid-2026. 

China: Alongside continued deployments of LEO mega constellations Qianfan (SpaceSail) and Guowang, China has announced guidelines for promoting LEO satellite services. Following this was the announcement of one of China’s three telecommunications giants, China Unicom, to operate Mobile Satellite Services (MSS) for Direct-To-Cellular (D2C) services. In contrast, China Telecom and China Mobile already have a license for Device-to-Device/Direct-To-Cellular (D2D/D2C) messaging and call services with Tiantong-1 and Beidou.

Meanwhile, back on Earth, sovereign cloud, Artificial Intelligence (AI), Graphics Processing Units (GPUs), and chipsets are advancing at a dizzying pace. This rapid evolution across the entire technology stack suggests these new space networks may aim for more than just connecting the unconnected or rescuing those in need.

The surge in LEO network deployments is ushering in a new era in global connectivity. By 2030, over 429 million people in the Global South (Latin America, Africa, India, Southeast Asia, and Oceania) will remain “uncovered” by terrestrial connectivity services, representing over US$800 billion in lost annual Gross Domestic Product (GDP) contributions from the digital economy. Assuming, rather optimistically, these users could afford current entry-level satellite internet service Average Revenue Per Users (ARPUs) (US$50/month for Starlink as an example), this would amount to an annual revenue opportunity of US$50 billion for over 83 million households.

This reflects a moderate opportunity for any one provider, but with so many operators projected to compete in the LEO connectivity market in the next 5 to 10 years, fighting for a slice of an annual revenue opportunity of US$50 billion for networks that cost tens of billions to launch and maintain doesn’t make much business sense. In fact, the consumer segment remains the most interesting in the D2C and IoT markets, with an estimated installed base of 25 billion devices by 2025. Even getting a 1% slice of this pie and charging half price for satellite internet (US$20/month) would net billions more each year than connecting all the unconnected in the Global South. When comparing these opportunities, it’s no wonder that satellite operators are trying to plant their proverbial flags with as many carriers already tapping into these devices as possible.

What’s more, significant opportunities are emerging at the edge that lies beyond connectivity. Take the Amazon Kuiper network, for example. The network will offer a complete turnkey path for transmitting from the remote edge to Amazon Web Services (AWS) cloud regions and enable access to cloud resources on private and secure links, independent of public internet infrastructure routes. This positions the network as a connectivity provider and alternate global backbone for at least critical or priority AWS cloud and AI data services. ABI Research forecasts the AI software market to reach over US$467 billion by 2030, while the cloud services market is expected to reach US$2 trillion (30% of which is Infrastructure as a Service (IaaS)) by 2030. Amazon understood the emerging opportunity, as lower latencies in LEO are essential for edge cloud and AI adoption, especially for inference and real-time user applications such as AI assistants, live translation services, remote robotics operations, telemedicine, and more.

From this perspective, the potential for connectivity revenue drops compared to cloud and AI data transport services. Satellite networks offer a competitive alternative data pipe to traditional infrastructure altogether, bypassing subsea cables, regional data center buildouts, and fiber deployments. Moreover, these networks are fast evolving to transport these services directly to devices via flexible and software-defined architectures. Many of these operators also own key pieces of the technology stack. SpaceX has rockets, xAI, and Tesla, Amazon has Project Kuiper, the AWS stack, and key infrastructure, Globalstar has the Apple ecosystem, and Chinese operators have Alibaba, Huawei, and Tencent cloud and AI stacks.

With the launch of these networks, we are entering an era of ubiquitous global connectivity, one with multiple alternate and independent internets in space, or what the space community calls “the Outernet”. A network independent from legacy architecture enables global sovereignty over data and communications and directly connects to chipsets in everyday devices and more. It’s no wonder that every country is seeking to build a LEO network. With the lower latencies, faster upgrade cycles, and application flexibility unlocked by LEO networks, enabling cloud and AI workloads at the remote edge via satellite will become the flagship offering in LEO.   

LEO satellite networks represent a new frontier in data delivery. With the emergence and widespread use of cloud and AI workloads, the ability to transport these capabilities anywhere on Earth, regardless of terrestrial connectivity infrastructure, is compelling. This is not only from an inclusivity standpoint and drawing closer to universal access to AI and cloud resources, but also from a security perspective. It reduces overdependence on single points of failure and mitigates geopolitical bottlenecks that undermine national security, global harmony, and economic autonomy.

Opportunities exist for many stakeholders in this new frontier. Telcos, Cloud Service Providers (CSPs), and cloud providers can partner with satellite network operators to enable cross-border value-added services to consumers, businesses, and government customers. Partnering with different satellite operators will also unlock distinct technology stacks, where a partnership can bring a unique value proposition—from specialized AI models and cloud services, to vehicles, consumer devices, and more. All of these expand the opportunities to innovate and grow, from cloud to AI to Application Programming Interfaces (APIs) and beyond.