Elon Musk’s SpaceX has acquired his AI startup xAI in a deal worth approximately $9 trillion Danish kroner, with plans to launch up to one million solar-powered satellites that will function as orbital data centers. According to the tech billionaire, space-based AI infrastructure could become reality within two to three years, driven by Earth’s inability to meet the enormous electricity demands of artificial intelligence.
The Merger Behind the Mission
SpaceX announced the acquisition of xAI on February 2, including the chatbot Grok and the social media platform X. The purchase price reportedly reached 1.583 billion Danish kroner, though the combined entity now holds a value of approximately 7.911 trillion kroner according to Bloomberg.
The merger brings together Musk’s rocket company with his AI ambitions under one roof. Interestingly, this consolidation addresses a critical challenge. xAI has been burning through capital at a rate of roughly one billion dollars monthly, while SpaceX generates as much as 80 percent of its revenue from launching Starlink satellites.
When the merger was announced, Musk explained his reasoning directly. On the long term, space-based AI is clearly the only way to scale, he wrote. Global electricity demand for AI simply cannot be met with ground-based solutions, even in the short term, without creating problems for society and the environment.
Why Space Makes Sense for Data Centers
The concept might sound like science fiction, but the technical advantages are substantial. John Leif Jørgensen, professor and space researcher at DTU Space, confirms that Musk’s timeline is feasible, even though “you never know with Elon Musk.”
According to Jørgensen, the energy benefits alone make the proposition compelling. The sun shines constantly in space, and solar panels generate approximately three times more energy there than on Earth. Without atmosphere to filter sunlight and no nighttime interruptions, the energy supply remains continuous.
Meanwhile, cooling presents one of the biggest challenges for terrestrial data centers. When millions of computer chips process user requests, they generate enormous amounts of heat. Traditional facilities require massive cooling systems that consume additional energy and often rely on water.
In contrast, space offers a natural solution. The vacuum of space sits at minus 270 degrees Celsius, making it remarkably easy to cool computer chips. SpaceX reportedly plans to use large, foldable radiators that deploy once satellites reach orbit, dissipating heat directly into space.
The Satellite Network Approach
Rather than launching massive structures, the orbital data centers would consist of smaller satellites working together. Jørgensen describes units roughly the size of washing machines with large solar panels extending from the sides.
This mirrors the architecture of Musk’s existing Starlink network, which currently provides internet service globally. SpaceX has recently applied for authorization to launch one million satellites specifically to support artificial intelligence operations, according to the BBC.
SpaceX’s Competitive Advantage
The Starlink project gave SpaceX invaluable experience in mass-producing and deploying thousands of satellites quickly. At the same time, the company has created something close to a monopoly on affordable, efficient space launches.
As Jørgensen points out, SpaceX maintains an extraordinary launch cadence. Just this week alone, Musk is sending up five batches of Starlink satellites, with 30 to 50 satellites in each batch.
This proven capability positions SpaceX uniquely in the emerging space-based AI race. Google has announced plans through Project Suncatcher to launch solar-powered satellites equipped with custom AI chips. OpenAI’s leadership reportedly recognizes the strategic importance and is willing to invest billions to secure an off-planet presence.
However, SpaceX’s combination of manufacturing capacity, launch capability, and existing satellite network experience creates substantial barriers for competitors.
Communication Through Laser Beams
The orbital system would rely on laser communication between satellites. These laser beams travel through the vacuum of space, achieving faster speeds than traditional fiber-optic cables without requiring extensive ground infrastructure.
This approach eliminates the need for building and maintaining terrestrial data transmission networks across diverse geographies and political boundaries.
The Economics and Timeline
Despite the ambitious vision, significant financial hurdles remain. xAI faces intense competition from established AI companies like Google, and the space infrastructure project requires enormous capital investment.
Naturally, Musk needs funding. According to reports from news agency AP, SpaceX plans an initial public offering later this year, possibly as early as June 2026. The IPO would provide access to public markets for raising capital.
Jørgensen suggests the timing connects directly to Musk’s competitive strategy. SpaceX generates substantial profits, creating a large capital pool that Musk can redirect toward other projects. At the same time, he needs to convince investors of his unique advantage.
As Jørgensen explains it, Musk excels at selling shares. He tells investors he can deliver cheaper AI than anyone else within three years, positioning them on an upward trajectory if they invest now.
Data center firms are projected to invest over $500 billion in expansion during 2026 alone, not including satellite and launch costs. Against this backdrop, Musk’s promise of a more economical alternative carries weight.
At the World Economic Forum in Davos, Musk elaborated on his timeline. Building AI data centers in space is a no-brainer, he stated. The lowest cost location for AI will be space, and that will be true within two years, maybe three at the latest.
Regulatory Strategy and Orbital Capacity
The one million satellite figure likely represents strategic positioning rather than immediate deployment plans. SpaceX used a similar approach with Starlink, requesting authorization for 42,000 satellites while currently operating around 9,600.
This strategy provides maximum design flexibility under Federal Communications Commission rules. Once authorized, SpaceX can scale deployment based on technical and economic factors without returning for additional permissions.
Currently, approximately 15,000 satellites already orbit Earth. Adding potentially millions more raises concerns about space debris and orbital pollution. FCC regulations require satellites to be de-orbited every five years, which creates both an environmental safeguard and an ongoing replacement cycle that ensures continuous revenue for SpaceX.
Engineering Challenges Ahead
While the concept holds promise, substantial technical obstacles remain. Maintaining and upgrading orbital facilities presents complications absent in terrestrial operations. Ground-based data centers simply replace failed or outdated hardware. Space-based facilities would require advanced robotics for repairs or launching entirely new satellites with every upgrade cycle.
The entire project’s economic viability depends heavily on Starship’s performance. After 11 test flights, Starship is scheduled to begin carrying operational payloads in 2026. The massive rocket is designed to transport millions of tons to orbit annually at dramatically reduced costs compared to previous launch systems.
Without Starship’s economics, orbital data centers would remain prohibitively expensive. With it, Musk may genuinely transform where and how humanity processes artificial intelligence.
As Jørgensen concludes, Musk occupies an exceptionally strong position. Nobody else can currently do what his companies can. Whether the timeline holds or extends, the technical foundation exists to make space-based AI infrastructure a reality.
Sources and References
The Danish Dream: Danish AI Software Challenges U.S. Military Tech
DR: Elon Musk har sammenlagt to af sine hjertebørn og vil nu lave datacentre i rummet
