This article is based on reporting originally published by Livemint (October 2020) and covers announcements made at Microsoft Ignite 2020 and related industry developments.
Cloud Providers Enter the Space Race
On 22 September 2020, Microsoft CEO Satya Nadella announced the preview of Azure Orbital at Microsoft Ignite 2020 in New Orleans. Microsoft described Azure Orbital as a Ground Station as a Service (GSaaS) offering, designed to help customers communicate with satellites and spacecraft, and to process the resulting data—all on a subscription basis, without owning physical ground infrastructure.
Microsoft was not first to market. Roughly five months earlier, IBM had announced a beta of its IBM Cloud Satellite service. And Amazon Web Services (AWS) had an even longer head start: approximately two years prior, AWS launched AWS Ground Station, enabling customers to control satellite communications, process downlinked data, and scale operations without building or managing their own ground stations. On 30 June 2020, AWS went further, announcing a dedicated business unit called Aerospace and Satellite Solutions.
Taken together, these moves signal a clear industry trend: the world’s leading cloud providers are competing not just for enterprise workloads on Earth, but for the infrastructure that connects Earth to the growing constellation of satellites in orbit.
Why Satellites and Cloud Are a Natural Fit
The push into space is driven by a fundamental shift in how satellites are deployed and used. According to the International Telecommunication Union (ITU), non-geostationary satellite orbits (NGSOs)—including medium earth orbits (MEO) and low earth orbits (LEO)—are being used with increasing frequency worldwide. Unlike traditional geostationary satellites, which appear fixed relative to a point on Earth, NGSO satellites move continuously across the sky. This introduces significant engineering challenges around tracking, handoff, and data capture.
As launch costs have fallen and rockets have become more accessible, a new generation of private satellite operators has emerged. Their satellites serve a wide range of applications:
- Weather forecasting and atmospheric monitoring
- Earth surface imaging and remote sensing
- Broadband internet and communications
- Video broadcasting and media distribution
- Maritime and aviation tracking
All of these applications share a common requirement: the data captured in orbit must be transmitted to the ground, processed, and stored. That processing and storage happens in ground stations—terrestrial facilities equipped with large dish antennas and the computing infrastructure needed to ingest satellite data.
The Ground Station Cost Problem
While the per-kilogram cost of reaching orbit has dropped dramatically over the past decade, operating a satellite still requires significant ground-side investment. According to a blog post by Jeff Barr, chief evangelist for AWS, building and running a ground station can cost $1 million or more. For smaller satellite operators, startups, or research institutions, this capital expenditure is often prohibitive.
Beyond the antenna hardware itself, the processing demands are substantial. LEO satellites pass over any given ground station for only a short window—often a matter of minutes—so data must be ingested rapidly and at high bandwidth. Analysing imagery, telemetry, or signals data then requires significant compute capacity. Large-scale storage adds further cost.
Historically, only well-funded government agencies and large commercial satellite operators could absorb these costs. That is now changing.
Ground Station as a Service: The Cloud Model Applied to Space
Cloud providers are applying the same economic logic to satellite infrastructure that they applied to servers a decade ago: replace large upfront capital expenditure with a flexible, usage-based operating expenditure model.
Under a Ground Station as a Service (GSaaS) model, a satellite operator schedules contact time with a ground station managed by the cloud provider. When the satellite passes overhead, the provider’s antenna captures the signal, and the data is immediately available within the provider’s cloud network—already co-located with the compute and storage resources needed to process it.
Key advantages of this model for satellite operators include:
- No capital investment in antenna hardware or ground facility construction
- Global coverage through a network of geographically distributed ground stations, reducing latency between passes
- Elastic compute co-located with the ground station, so data processing scales with demand
- Pay-per-use pricing, which lowers the barrier to entry for smaller operators and academic researchers
- Faster time to orbit, since operators do not need to build ground infrastructure in parallel with their satellite programs
The Competitive Landscape in 2020
By late 2020, three major cloud providers had staked out positions in this market:
| Provider | Product | Model | Status (as of Oct 2020) |
|---|---|---|---|
| Amazon Web Services | AWS Ground Station | GSaaS, pay-per-minute contact | Generally available |
| IBM | IBM Cloud Satellite | Distributed cloud / hybrid | Beta |
| Microsoft Azure | Azure Orbital | GSaaS, subscription-based | Preview |
It is worth noting that IBM Cloud Satellite has a somewhat different scope than AWS Ground Station or Azure Orbital. IBM’s offering is primarily focused on running IBM Cloud services in distributed locations—including on-premises or at the network edge—rather than providing antenna-based satellite downlink services directly. The products are not perfectly analogous, and buyers should evaluate each on its specific capabilities.
Broader Implications for the Space Industry
The entry of hyperscale cloud providers into satellite infrastructure has implications beyond cost savings for existing operators. It lowers the barrier to entry for the entire sector. A startup with a single CubeSat can now access the same ground station network as a government agency, paying only for the contact time it uses.
It also accelerates the integration of space-derived data into conventional cloud workloads. Earth observation imagery, for example, can flow directly from a ground station contact into a machine learning pipeline running in the same cloud region—enabling near-real-time analysis of agricultural fields, disaster zones, or shipping lanes.
For cloud providers themselves, space represents a new frontier of differentiation. As core cloud services become increasingly commoditised, owning the infrastructure that connects satellites to the cloud creates a defensible position and opens access to high-value government and defence contracts.
Key Takeaways
- AWS, Microsoft Azure, and IBM all announced satellite-related cloud services in 2020, reflecting a broader industry push into space infrastructure.
- The driving force is the growth of LEO and MEO satellite constellations, combined with falling launch costs and rising demand for satellite data.
- Ground stations are expensive to build and operate; a GSaaS model shifts this from capital to operational expenditure for satellite operators.
- Co-locating compute with ground station infrastructure dramatically reduces the time and cost of processing satellite data.
- This trend is likely to accelerate as more commercial satellites reach orbit and demand for real-time Earth observation data grows.