Data Centers in Orbit: The Next Cloud Market?

The growing interest in orbital data centers shows that the cloud industry is already preparing to expand into space. Google is experimenting with artificial intelligence processors in orbit through its Suncatcher project, SpaceX is working on a constellation that could include up to one million satellites, and start-up Starcloud is attracting hundreds of millions of dollars from investors. In March 2026, the latter raised $170 million before announcing another $250 million funding round a few months later, bringing its valuation to $2.3 billion.

This momentum comes amid growing energy requirements driven by the emergence of increasingly powerful artificial intelligence models. According to the International Energy Agency (IEA), data centers consumed 415 TWh of electricity in 2024, accounting for approximately 1.5% of global electricity consumption, and the agency forecasts that this figure will reach 945 TWh by 2030. To meet these energy needs, space presents a unique opportunity. In low Sun-synchronous orbit (at an altitude of approximately 650 km), a solar panel can capture up to eight times more energy than on Earth, while benefiting from almost continuous sunlight.

The equation could therefore prove particularly attractive to the space industry: harnessing abundant solar energy, bringing computing capabilities closer to the satellites that generate data, and creating new capacity for AI workloads. While these prospects are promising, numerous technical challenges remain, and only a handful of players may ultimately be able to gain a competitive advantage.

Delivering Space Data in Real Time: The First Market for Orbital Data Centers?

Earth observation satellites generate considerable volumes of images and measurements that currently have to be transmitted to the ground before being processed. Although this approach is well established, it has one major drawback: the time required to process data on the ground.

To overcome this limitation, onboard processors equipped with AI capabilities could analyze data directly in orbit and transmit only the relevant information. This architecture would pave the way for near-real-time applications such as wildfire detection, maritime surveillance, mapping, meteorology, and more. Computing would thus become a form of space-based edge computing, with artificial intelligence capabilities located close to sensors or even embedded directly in satellites. Players such as Starlink already appear well positioned, as they could leverage the relay capabilities of their telecommunications satellites to transmit the considerable volumes of data generated by orbital clouds, both between satellites and from space to Earth. Google is also exploring this approach with Suncatcher, its program dedicated to deploying artificial intelligence aboard satellites. A first prototype was launched in October 2026 in partnership with satellite imagery company Planet.

Although achieving these objectives will not be without challenges, the operational benefits would undoubtedly be considerable. Space data could be exploited directly, with transmitted data processed by onboard artificial intelligence, paving the way for real-time applications of space-based observation technologies, including optical and radar systems.

Investors, Hyperscalers and Space Companies Are Betting on Orbital Cloud Computing

The amounts being invested show that the market is already beginning to attract venture capital. Starcloud is currently drawing much of the attention, having raised $170 million in March 2026, followed by a $250 million funding round that brought its valuation to $2.3 billion. This valuation is based on expectations of explosive growth in global demand for computing capacity. In 2025, Starcloud demonstrated the operation of an Nvidia H100 GPU in orbit [1], a chip used for artificial intelligence workloads. The company now aims to deploy much larger infrastructures. Google, meanwhile, brings a different kind of financial strength to the market: that of a hyperscaler capable of financing its own experiments and controlling part of the technology value chain. Its Suncatcher project [2] is directly aligned with its strategy around Tensor Processing Units (TPUs), cloud infrastructure and AI. Finally, SpaceX has a different advantage: control over the value chain (satellite manufacturing and launches) and operational experience with Starlink. The company has filed an application with the Federal Communications Commission (FCC) [3] for an architecture called “SpaceX Starmind”, which could comprise up to one million satellites dedicated, among other things, to orbital computing capabilities. This application is part of a regulatory process, and the constellation’s final size will depend on authorizations and industrial decisions.

This convergence of venture capital, hyperscalers and the space industry recalls the early days of cloud infrastructure: companies are investing heavily in physical capacity before the market has reached maturity. However, while this market appears promising in the medium to long term, short-term prospects remain limited due to the lack of immediate demand and the need to adapt space architectures. Initially, the business model could therefore be based primarily on selling computing capacity. Operators could rent out GPUs in orbit, charge for satellite data processing, or even offer space-based storage. Government and military applications could constitute an initial market, given their constant pursuit of operational superiority and resilience.

Operational Challenges and Constraints

Two major constraints explain why the first commercial projects are likely to prioritize applications offering high value per unit of computing capacity.

First, launch costs are currently the determining variable in the business model. Every kilogram sent into orbit represents an additional expense compared with a terrestrial data center. Orbital computing must therefore offset this cost through sufficiently significant gains in energy, data processing or infrastructure availability. In this field, SpaceX enjoys an advantageous position, as it currently offers the lowest-cost launch solutions on the market. As evidence of this dominant position, when Google launched its Suncatcher prototype satellite on October 1, it used Elon Musk’s Falcon 9 rocket.

Equipment lifespan is the second major financial factor. In a terrestrial data center, servers can be replaced gradually, whereas in an orbital facility, equipment failures, radiation effects, processor obsolescence and replacement operations must all be anticipated. Profitability will therefore depend heavily on how long each satellite can generate revenue. This revenue generation will depend on development and operating costs, with launch costs—the primary determining variable—playing a central role.

The Bet on a New Cloud

Given the growing integration of AI into space applications, the orbital data center market will ultimately be about computing capacity rather than data storage. The most credible scenario would therefore resemble a hybrid cloud distributed between Earth and space. Terrestrial data centers would continue to handle general-purpose workloads, while orbit could accommodate space-data processing, certain AI workloads and applications requiring specific energy or geographical infrastructure. For cybersecurity players, this development opens up a new frontier: a data center with part of its infrastructure located several hundred kilometers above the ground, whose availability depends simultaneously on software, networks, launch vehicles and the space environment. Faced with these constraints, players such as SpaceX and Google can capitalize on comparative advantages of an entirely different scale, raising questions about the ability of emerging companies to offer competitive solutions.