Elon Musk Says Future AI Computing Could Move to Space

Elon Musk believes Artificial Intelligence computing could eventually move almost entirely beyond Earth, with future superintelligence relying on data centres operating in space. The SpaceX chief has suggested that the overwhelming majority of computing power required for advanced AI systems will be based in orbit, although he has not provided a timeline for the transition.

The prediction comes as major technology and space companies, including Google and SpaceX, explore ways to build AI computing infrastructure beyond Earth. However, the technology remains at an early stage, with high costs and engineering challenges making space-based data centres considerably more expensive than their terrestrial counterparts.

Elon Musk Predicts AI Computing Will Move Into Space

Musk made the prediction in a post on X while responding to investor Shaun Maguire and another account discussing SpaceX’s computing business.

Maguire suggested that SpaceX could become a low-cost provider of computing power, initially through data centres on Earth and eventually through orbital infrastructure. Musk went further, saying that future computing for what he described as superintelligence would be overwhelmingly located in space.

He clarified that by “overwhelmingly”, he meant a proportion that would round up to 100%. However, Musk did not explain when this transition could happen or provide details about how SpaceX would develop and operate such infrastructure.

Google Has Already Tested AI Computing in Space

While Musk’s prediction remains ambitious, Google is already testing the possibility of running AI workloads using computing infrastructure in orbit.

The company’s Project Suncatcher is a research initiative aimed at developing solar-powered satellites equipped with Tensor Processing Units (TPUs), Google’s specialised AI processors. The project explores whether interconnected satellites could eventually provide computing capacity for advanced AI applications.

Google launched its first prototype satellite on October 1 in partnership with Planet. On October 30, the company announced that it had successfully received its first AI-generated response from a data-centre-class TPU operating in orbit.

The satellite ran Google’s Gemma 4 model, marking an early demonstration of AI computing beyond Earth. Google plans to test two interconnected satellites in 2027 to explore how orbital computing systems could work together.

SpaceX Also Plans to Develop Orbital AI Data Centres

SpaceX is also exploring space-based computing infrastructure. According to a Reuters report cited in the supplied information, the company is targeting a demonstration of its orbital AI computing technology by late 2027, with potential deployments beginning as early as 2028.

SpaceX has also sought regulatory approval for a proposed constellation of up to one million satellites that could function as data centres. If developed, the network would represent a significant expansion of the company’s ambitions beyond satellite internet services.

However, the proposed constellation is not an operational network. Its development would depend on technological progress, regulatory approvals, launch capacity and the ability to make orbital computing economically viable.

Why Does Elon Musk Want AI Data Centres in Space?

One potential advantage of placing AI infrastructure in space is access to sunlight. Satellites operating in suitable orbits could receive near-constant solar energy, potentially providing a source of power for energy-intensive AI computing systems.

As AI models become more advanced, the computing resources required to train and operate them could increase significantly. Space-based infrastructure could offer an alternative way to supply energy and computing capacity without relying entirely on land-based facilities.

However, access to solar energy does not automatically make orbital data centres cheaper or easier to operate. Hardware would have to be launched into orbit, protected from radiation and maintained under extreme environmental conditions.

Space Data Centres Could Cost Up to 100 Times More

The biggest obstacle to orbital AI computing is currently its cost. Amanda Peterson Corio, Google’s global head of data centre energy, has said that building data centres in space is not economically feasible at present.

Speaking at The Information’s Women in Tech, Media and Finance Summit 2026, Corio said space-based data centres could be “incredibly, incredibly expensive, up to 100 times more expensive than building on Earth”.

“Explore all the terrestrial options before you go extraterrestrial,” Corio advised.

Her comments highlight the financial challenges facing companies considering orbital computing. Launching satellites and specialised hardware, replacing damaged equipment and maintaining the infrastructure could make space-based systems significantly more expensive than conventional data centres.

Technical Challenges of Running AI Systems in Orbit

Beyond the financial burden, operating AI processors in space would require solutions to several engineering problems.

• Radiation protection: Satellites and their processors must withstand radiation that can damage electronic components and affect computing operations.
• Temperature management: Hardware must operate in an extreme thermal environment, making reliable cooling and heat management essential.
• Launch costs: Large quantities of specialised equipment would need to be transported into orbit, adding significant upfront expenses.
• Maintenance and replacement: Repairing or replacing damaged equipment in space would be more difficult than servicing data centres on Earth.
• Satellite coordination: Interconnected orbital systems would need reliable communication and coordination to operate as a functioning computing network.

These challenges mean that space-based AI infrastructure would require substantial technological improvements before it could compete economically with conventional data centres.

Will AI Data Centres Really Move Entirely Into Space?

Musk’s prediction points towards a possible future in which orbital infrastructure plays a major role in powering advanced artificial intelligence. Google’s early experiments and SpaceX’s proposed plans show that companies are already investigating the concept.

However, the technology remains experimental, and the economics are not yet favourable. Google’s successful orbital AI test represents an early demonstration rather than proof that large-scale commercial data centres can operate profitably in space.

For now, Earth-based data centres remain the more economically practical option. Whether orbital computing eventually becomes a major part of the AI industry will depend on lower launch costs, advances in satellite technology, reliable hardware protection and improvements in the economics of operating infrastructure beyond Earth.