How data centres in space sustainably enable the AI revolution

Artificial intelligence (AI) is redefining every sector of our global economy, but this transformative technology comes with an unprecedented and growing appetite for energy and infrastructure.

As the world races toward artificial general intelligence (AGI), we are confronting a stark reality: Earth’s current energy and land capacity are struggling to keep pace sustainably.

This urgent challenge requires an out-of-the-box solution or, more accurately, an off-planet one. The emerging frontier of space-based data centres is no longer a concept confined to science fiction; it is a vital, technically feasible solution poised to unlock the next wave of AI progress while drastically minimizing our planetary footprint.

This is why orbital data centres are an essential topic for discussion at gatherings such as the World Economic Forum’s Annual Meeting in Davos, Switzerland, right now.

The AI revolution is overwhelmingly energy-hungry. Large data centres – the engine rooms of AI – already consume approximately 1.5% of global power, a figure projected to rise sharply as models such as GPT-6 and Llama 5 are trained, requiring multi-gigawatt (GW) clusters.

This surge in demand hits an already strained global energy grid. Utilities in the Western world, hampered by regulatory and planning restrictions, cannot adapt at the required pace and scale, leading to an impending energy crunch that threatens to hinder AI development.

Beyond energy supply, terrestrial data centres face profound sustainability challenges:

In short, the existing Earth-based model does not scale or sustain the future of AI.

Space offers an environment naturally suited for high-efficiency, sustainable computing, providing solutions to the very challenges that limit terrestrial development.

In a dawn-dusk sun-synchronous orbit (SSO), a data centre can leverage continuous, high-intensity solar power, unhindered by night-time, weather, or atmospheric attenuation.

According to a Starcloud whitepaper, this continuous illumination is crucial, as it allows orbital solar arrays to achieve a capacity factor of greater than 95%, compared to a median of just 24% for terrestrial solar farms in the United States. Furthermore, peak power generation in space is roughly 40% higher due to the absence of atmospheric losses.

This means a solar array in space can generate over five times the energy as the same array on Earth. By using this incredibly efficient, clean energy source, Starcloud estimates an equivalent energy cost of approximately $0.005 – up to 15 times lower than today's wholesale electricity prices.

This not only makes orbital data centres economically viable but fundamentally shifts the energy dynamics of AI development.

The deep vacuum of space serves as a gigantic, cold heatsink with an effective ambient temperature of around -270 Celsius.

Orbital data centres can leverage this by using simple, passive radiative cooling to achieve low coolant temperatures, eliminating the need for energy-intensive chillers and, critically, the consumption of fresh water for cooling. This results in significant operational cost savings and a massive reduction in environmental impact.

Perhaps most importantly, space provides a boundless domain for scaling compute. Orbital data centres can be linearly scaled almost indefinitely through modular design, assembled in a 3D architecture for ultra-low latency within the cluster.

A 5 GW cluster – necessary for next-generation AI models and exceeding the capacity of most of the world's largest power plants – is simply not possible on Earth with today's infrastructure but is achievable in orbit.

With the advent of reusable, heavy-lift launch vehicles, launching a 5 GW data centre could conceivably be accomplished in two to three months, dramatically faster than the years-long timeline for comparable terrestrial infrastructure development.

The orbital data centre is built on principles of modularity and maintainability.

While the economic and environmental benefits are compelling, realizing this vision requires responsible innovation that addresses the inherent challenges of operating in space.

Larger orbital structures necessitate being especially responsible users of low Earth orbit (LEO). This requires highly responsive spacecraft manoeuvrability, state-of-the-art space-object tracking and coordinating with all relevant bodies.

The modular design further contributes to circular sustainability, with old containers designed for re-entry and demising (burning up completely) in the atmosphere, or for salvaging hardware and materials at the end of the data centre's 10-15 year design life.

Operating in LEO requires radiation shielding for sensitive components. The challenge is balanced by two factors: the choice of a low-radiation orbit (dawn-dusk SSO) and the fact that the shielding mass per compute unit decreases linearly with container size.

This monumental shift – placing GW-scale data centres at the intersection of decreasing launch costs, an energy demand crunch and the proliferation of low-cost space connectivity – demands a global conversation.

We welcome international frameworks and multistakeholder collaboration to ensure space computing develops responsibly. This means balancing rapid innovation with ensuring that our orbital infrastructure champions sustainable stewardship of Earth.

By moving the most energy- and water-intensive parts of AI computation beyond Earth’s atmosphere, we turn an existential challenge, energy demand, into a profound opportunity for decarbonization, resilience and a truly sustainable digital future.

Space, once viewed only as a frontier for exploration, is now becoming a critical sustainability frontier for our planet's most transformative technology.