AI in extreme situations Why Shoot AI to the Moon When you can Sink it?

The idea of establishing data centers in Earth's orbit is not new and currently not economical. Tech elites indulge in the fantasy of AI from space while there is other uncharted territory to explore: the sea.

The increasing energy demand of data centers, particularly due to AI workloads, has led to increasingly radical infrastructure ideas in recent years. There is talk of locations in cooler regions, then of locations in areas with largely stable weather, such as deserts. Then there are discussions about data centers near renewable energy sources—or near decommissioned nuclear reactors that could potentially be reactivated.

A vision that is both fascinating and controversial regularly emerges in discussions and deliberations: data centers in space. The idea behind it initially seems plausible. In orbit, virtually unlimited solar energy is available. Space constraints, like those on Earth, are no longer an issue. Well, one might quickly believe that—if one chooses not to engage with the issue of space debris.

And yes, the waste heat could theoretically be released into space. Well... due to the vacuum, it’s not that simple. Anyway, let’s stay on the green playground for now. At the same time, data could be processed directly where it is generated—such as with Earth observation satellites. These are all valid ideas that are already being implemented.

Europe’s security landscape is changing rapidly

As defence budgets rise and EU programmes expand, civil technology providers are becoming vital contributors to Europe’s strategic autonomy. The event will act as a neutral platform for dialogue between technology suppliers, integrators, and decision-makers shaping the next generation of European defence capabilities and aims to open doors between civil industry and defence procurement, providing practical insights.

Find out more and register

Computing in space

However, when it comes to sending data from Earth into orbit for processing and then sending it back, this is where criticism arises, including from the analyst firm Gartner. In a recent analysis, experts concluded that orbital data centers will not meet terrestrial demands in the foreseeable future, and companies would be better off focusing their investments on infrastructure on Earth.

The reasoning is shaped less by vision and more by physics and economics. Transportation costs dominate every calculation, as IT hardware, power supply, cooling, and structure must first be brought into orbit. Even with sharply declining launch costs, mass and volume remain key cost drivers. Additionally, extreme environmental conditions such as radiation, vacuum, and temperature cycles require specially hardened components, further increasing costs. The challenges electronics would face have already been explained by our space enthusiast Hendrik Härter, when Jeff Bezos shared his vision of AI data centers in space at the Italy Tech Week 2025.

First the setup, then the maintenance

Another key point is maintenance. While servers in terrestrial data centers can be continuously replaced, any repair in orbit would require complex missions or advanced robotics. Without maintenance, systems would quickly contribute to additional space debris; an issue that is already critically discussed today due to its hazardous nature.

At the same time, data transmission remains a limiting factor, as large volumes of data would need to be reliably transferred between Earth and orbit. Against this backdrop, the idea of relocating a significant portion of global cloud capacity to space currently seems neither economically nor practically feasible. Even the often-cited advantage of cooling becomes relative upon closer examination, as heat in a vacuum can only be dissipated through radiation, requiring large radiator surfaces and additional mass.

Additionally, there is an aspect often overlooked in visionary discussions: space problems cannot truly be solved through orbital computing. Data centers require massive infrastructure—energy conversion, communication systems, structural modules—which also demand space and resources. The physical location merely shifts from the Earth's surface to orbit, without addressing the fundamental scaling issues.

Why look up... and not down?

A concept much closer to practical implementation, though still unusual, is data centers in the sea. Underwater data centers take advantage of the high heat capacity and stable temperatures of seawater for cooling, significantly reducing one of the largest energy consumers of traditional data centers. At the same time, many large urban areas are located near coasts, keeping latency low and allowing the use of existing fiber optic infrastructure.

That this approach is not merely theoretical is demonstrated by projects already carried out. Microsoft tested "Project Natick" over several years with a sealed data center module on the seabed off the Scottish coast. The results indicated, among other things, higher hardware reliability, attributed to the stable environment and closed design. Underwater data centers have also been tested or announced in China, including in connection with offshore energy projects. Such concepts enable direct coupling to renewable energy sources like offshore wind power, unlocking additional efficiency potential.

The other side of the coin

Of course, underwater data centers are not without challenges. Maintenance interventions are understandably more complex than on land. Long-term material durability in saltwater must be ensured, and ecological impacts need to be assessed. However, these disadvantages are significantly less severe compared to those of orbit. If maintenance in space is considered a solvable problem in many visions, this applies all the more to systems located a few hundred meters below the ocean surface, accessible with existing offshore technology.

Perhaps the future of data centers lies not above our heads but beneath the water's surface. While space-based data centers remain a long-term visionary topic with specialised niche applications, underwater data centers could become a realistic addition to existing infrastructure much sooner. They address specific issues such as cooling, energy efficiency, and space availability without having to overcome the fundamental physical and economic challenges of space exploration.

Whether we will actually need the data center boom in the next 5, 10, 15, or 50 years, or whether ways will be found by then to significantly reduce the energy and space demands of AI, is of course another matter. (sb)