SpaceX: Space-Based Computing Could Transform AI

AI's appetite for energy and water has become one of technology's most pressing challenges.

Now, a solution that sounds more science fiction than reality could be taking shape.

The merger of SpaceX and xAI into a single entity valued at approximately US$1.25tn represents a bold attempt to relocate AI infrastructure beyond Earth's atmosphere, where solar power flows constantly and space constraints disappear.

Elon Musk, CEO of SpaceX and xAI, announced the combination as a solution to the fundamental resource limitations that currently restrict AI development.

The combined organisation plans to deploy AI satellites functioning as orbital data centres, powered entirely by solar energy harvested directly in space.

This approach could address the resource constraints on Earth, where data centres consume vast quantities of electricity and water for cooling systems.

Earth currently intercepts 173,000TW of solar energy, which is 10,000 times more than the planet uses.

However, capturing even a fraction of this requires infrastructure that terrestrial systems struggle to support.

"To harness even a millionth of our Sun's energy would require over a million times more energy than our civilisation currently uses," Elon says.

"The only logical solution therefore, is to transport these resource-intensive efforts to a location with vast power and space.

"By directly harnessing near-constant solar power with little operating or maintenance costs, these satellites will transform our ability to scale compute."

Musk estimates that by 2027, space-based systems could become the most cost-effective method for generating AI compute power.

Lunar infrastructure for satellite production

The technology strategy extends beyond orbital data centres to encompass manufacturing capabilities on the Moon.

Starship, SpaceX's heavy-lift spacecraft, has been designed to transport 200 tons of cargo per flight, delivering Starlink satellites to orbit before continuing to lunar destinations.

This payload capacity could enable the establishment of manufacturing facilities on the Moon, taking advantage of lunar resources to produce satellites that can be deployed deeper into space.

Musk explains that factories on the lunar surface could use an electromagnetic mass driver combined with local manufacturing to deploy between 500TW and 1,000TW per year of AI satellites into deep space.

This scale of deployment would represent a significant advancement on the Kardashev scale, a measurement system for civilisations based on their energy usage.

Humanity currently sits below the first level because it does not harness all available energy on Earth.

Reusability reduces space access costs

During his appearance at the World Economic Forum's Annual Meeting, Musk discusses how reusable spacecraft technology could dramatically reduce the cost barrier to space operations.

The economic model mirrors commercial aviation rather than traditional space launches.

"If you had to throw away an aircraft after every flight, that would be a very expensive flight. If you only have to refuel, then it's the cost of the fuel," Elon told the World Economic Forum.

This cost reduction could prove essential for the viability of space-based AI infrastructure, where multiple launches would be required to establish and maintain orbital data centres and lunar manufacturing facilities.

Musk also emphasises his belief that AI, robotics and solar power, if deployed more broadly, could unlock an era of unprecedented global abundance by overcoming existing energy constraints.

Technology consolidation accelerates AI ambitions

The merger continues a pattern of consolidation across Musk's portfolio of ten companies.

xAI began as a division of X before becoming valuable enough to acquire its parent company in an all-stock transaction.

Musk says the combination would "combine the data, models, compute, distribution and talent".

Only Neuralink and The Boring Company, an infrastructure venture, remain as separate entities outside the larger consolidated operations.

As demand for AI capabilities accelerates across business sectors, the infrastructure required to support advanced models has become a critical constraint.

Traditional data centres face limitations in power availability, cooling capacity and physical space.

Space-based alternatives could potentially circumvent these restrictions entirely, though significant technical and economic challenges remain before such systems become operational.

The technology being developed for space-based AI infrastructure could also support broader space exploration objectives across Elon's portfolio, creating shared capabilities that serve multiple purposes beyond AI compute alone.

Whether this vision of orbital data centres and lunar manufacturing proves technically and economically feasible remains to be seen, but the merger signals a serious commitment to pursuing technology solutions beyond Earth's surface.