Space Data Centers: The AI Infrastructure Race Heats Up | Brick & Bit
Artificial Intelligence
Space Data Centers: The AI Infrastructure Race Heats Up
SpaceX wants to launch up to 1 million orbital data centers by 2030. This could redefine terrestrial infrastructure value and create new investment opportunitie
B&B
Brick & Bit
April 3rd, 2026
7 min readMIT Technology Review
Key Takeaways
Space-based data centers could solve the water and energy problems plaguing terrestrial AI, but they must first overcome monumental technical challenges requiring innovation across multiple fronts simultaneously.
Elon Musk wants to put a million data centers in orbit. This isn't science fiction—it's the next frontier of AI infrastructure that's rapidl...
But SpaceX isn't alone in this race that's rewriting the rules of infrastructure economics. Jeff Bezos declared last year that the tech indu...
Elon Musk wants to put a million data centers in orbit. This isn't science fiction—it's the next frontier of AI infrastructure that's rapidly moving from concept to concrete business plan. SpaceX's January 2026 filing with the U.S. Federal Communications Commission represents more than just another ambitious project; it signals a fundamental shift in how we approach computational infrastructure at scale. This initiative addresses the core constraints threatening AI's exponential growth while opening entirely new markets and investment opportunities.
The Big Picture The tech industry is hitting Earth's physical limits with unprecedented force. The AI explosion consumes energy and water at rates unsustainable for many communities, creating environmental and geopolitical tensions that could throttle technological progress. In January 2026, SpaceX filed a landmark application with the U.S. Federal Communications Commission to launch up to one million data centers into Earth's orbit. The stated goal is unambiguous: fully unleash AI's potential without triggering environmental crises that could undermine the very societies AI aims to serve. This move comes directly in response to 2025 reports showing AI data centers consuming resources comparable to small countries, highlighting the urgency of finding sustainable solutions.
satellite constellation in orbit
But SpaceX isn't alone in this race that's rewriting the rules of infrastructure economics. Jeff Bezos declared last year that the tech industry will inevitably move toward large-scale computing in space, noting that terrestrial limitations are insurmountable for long-term growth. Google has accelerated its plans and now projects launching a test constellation of 80 satellites as early as next year, reflecting the urgency with which major tech players are addressing this challenge. Last November, startup Starcloud achieved a historic milestone by launching a satellite fitted with a high-performance Nvidia H100 GPU, proving that advanced AI chips can operate in the space environment. The company envisions orbiting data centers as large as terrestrial ones by 2030—a goal that seemed implausible just five years ago.
“Space-based data centers could solve the water and energy problems plaguing terrestrial AI, but they must first overcome monumental technical challenges requiring innovation across multiple fronts simultaneously.”
By the Numbers - **SpaceX's application:** Up to 1 million orbital data centers proposed by 2030 - **Google's test constellation:** 80 satellites planned for next year - **Minimum orbit temperature:** 80 °C in constantly illuminated orbits - **Thales feasibility study:** Possibility of putting gigawatt-scale data centers in orbit for Europe - **Current terrestrial data center consumption:** Approximately 3% of global electricity use - **AI growth projection:** AI energy consumption expected to triple by 2030 - **Terrestrial cooling costs:** Up to 40% of traditional data center operational expenditure - **SpaceX Starship target cost:** Aiming for under $100 per kilogram to orbit
By the Numbers - **SpaceX's application:** Up to 1 million orbital data centers proposed by 2030 - **Google's test constellation:** 80 satellites planned for next year - **Minimum orbit temperature:** 80 °C in constantly illuminated orbits - **Thales feasibility study:** Possibility of putting gigawatt-scale data centers in orbit for Europe - **Current terrestrial data center consumption:** Approximately 3% of global electricity use - **AI growth projection:** AI energy consumption expected to triple by 2030 - **Terrestrial cooling costs:** Up to 40% of traditional data center operational expenditure - **SpaceX Starship target cost:** Aiming for under $100 per kilogram to orbit
data center growth chart
Why It Matters This represents a fundamental redefinition of what constitutes valuable infrastructure in the AI era. For decades, data center location was defined by access to cheap power, water for cooling, and proximity to end-users. The space-based data center proposal completely flips this logic: the optimal location is no longer on the ground but in specific orbits offering constant solar power and natural thermal dissipation. This shift isn't merely technical but economic and strategic, moving competition from limited terrestrial resources to theoretically unlimited space.
Potential winners include space launch companies like SpaceX, which see a massive new source of demand for their rockets that could sustain growth for decades. Satellite manufacturers like Thales Alenia Space, which has already developed thermal management systems for large telecommunications satellites, could pivot to this new market that promises to be more lucrative than traditional communications. And chip companies like Nvidia, whose processors are already flying in space, could develop orbital-specific versions that maximize efficiency under extreme radiation and temperature conditions.
Losers could be communities currently hosting large data centers that rely on their tax revenue and jobs. If the industry moves significantly to space, these localities might face decreased demand for their power and water resources, along with the loss of specialized jobs that have been pillars of local economies. Furthermore, countries that have heavily invested in terrestrial infrastructure could see their competitive advantage erode if space becomes the new standard for high-power computing.
What This Means For You For investors, this creates a new infrastructure asset class worth immediate attention and deep analysis. Traditional data center REITs could face significant pressure if orbital costs fall enough to make terrestrial location less competitive, especially considering rising regulatory and environmental costs. Meanwhile, pure-play space technology companies might see a new growth source that transforms their business models and market valuations.
What This Means For You For investors, this creates a new infrastructure asset class worth immediate attention and deep analysis. Traditional data center REITs could face significant pressure if orbital costs fall enough to make terrestrial location less competitive, especially considering rising regulatory and environmental costs. Meanwhile, pure-play space technology companies might see a new growth source that transforms their business models and market valuations.
1Diversify into space infrastructure: Consider strategic exposure to launch companies, satellite manufacturers, and specialized thermal management component suppliers with proven space technology.
2Monitor data center REITs: Watch for statements about how they plan to compete with the orbital value proposition and evaluate their adaptation plans for this new paradigm.
3Track launch costs: The tipping point will come when putting hardware in orbit becomes cheaper than environmental and resource costs on the ground—a moment that could arrive sooner than expected.
4Evaluate critical component companies: Identify companies providing thermal management systems, radiation protection, and space solar power, as these will be essential to this industry's success.
5Assess regulatory developments: Follow international space policy discussions that will determine orbital allocation and frequency rights for space data centers.
investor analyzing market charts
What To Watch Next Two key catalysts will determine whether this moves from theory to reality in the coming years. First, the success of the thermal management system Thales Alenia Space developed for large telecommunications satellites. If this technology can scale to handle heat generated by AI data centers—which is significantly greater than traditional communications—one major technical hurdle disappears. Tests planned for late 2026 will be crucial for validating this capability.
Second, the evolution of launch costs, which currently represent the largest economic barrier. SpaceX promises its Starship rocket will dramatically reduce cost per kilogram to orbit, possibly by up to 90% compared to current systems. If they deliver on this promise within the next 2-3 years, space data center economics could become viable much sooner than many anticipate, accelerating the transition from terrestrial infrastructure.
A third critical factor will be international regulation. Coordination between space agencies and telecommunications bodies will be essential for allocating orbits and frequencies, avoiding conflicts that could delay this nascent industry's development. Decisions made in forums like the International Telecommunication Union in 2027 could define the adoption timeline.
The Bottom Line Space-based data centers are no longer distant fantasy from science fiction novels. They're a logical and necessary response to the physical limitations facing AI expansion on Earth, where resources are finite and environmental costs grow exponentially. While technical challenges are significant—particularly heat management in an environment where temperature never drops below 80 °C—solutions are already in development, and early prototypes are demonstrating feasibility.
The Bottom Line Space-based data centers are no longer distant fantasy from science fiction novels. They're a logical and necessary response to the physical limitations facing AI expansion on Earth, where resources are finite and environmental costs grow exponentially. While technical challenges are significant—particularly heat management in an environment where temperature never drops below 80 °C—solutions are already in development, and early prototypes are demonstrating feasibility.
By 2030, we could see the first commercial-scale orbital data centers operating, marking the beginning of a new era in technological infrastructure. This would redefine not just AI infrastructure but also the economics of communities currently hosting these facilities and the geopolitics of access to high-power computing. The race for space as computing's next frontier has just begun, and those who position themselves strategically today could define the technological landscape for decades to come. The move to space isn't an option but a necessity to sustain the exponential growth of artificial intelligence that's transforming every aspect of our society.
