The race to develop artificial intelligence is creating an unprecedented demand for computing power and electricity. As terrestrial infrastructure encounters growing energy, land and environmental constraints, some of the world’s most ambitious technology companies are exploring an extraordinary alternative: moving AI computing infrastructure into orbit.
Rock Edge Research | Space Exploration | Investment Intelligence
Research and Analysis | 02 November 2026 | 15–20 -minute read
Written by Moalosi Moyane
The Energy Problem Behind Artificial Intelligence
Artificial intelligence is rapidly transforming the global economy. From generative AI and autonomous systems to advanced robotics, scientific research and financial modelling, the technology is becoming increasingly embedded in modern economic activity.
However, behind every sophisticated AI application lies an enormous computing infrastructure.
Training and operating advanced AI models requires specialised processors, enormous computing capacity and sophisticated data centres capable of processing vast quantities of information. These facilities consume substantial amounts of electricity, creating a new challenge for the technology industry: securing sufficient energy to sustain the next generation of artificial intelligence.
The International Energy Agency has identified data centres as an increasingly important source of global electricity demand. In the United States, the US Government Accountability Office has highlighted projections that data centres could account for as much as 12% of national electricity consumption by 2028.
Technology companies are consequently looking beyond traditional data centre locations, exploring alternative energy sources, advanced cooling technologies and new infrastructure models.
The challenge is no longer exclusively about developing more powerful computer chips. It is increasingly about securing the energy, infrastructure and resources required to operate them.
One particularly ambitious proposition is emerging: What if the next generation of AI data centres were built in space?
Harnessing the Sun Beyond Earth
The fundamental attraction of space-based computing is energy.
Unlike terrestrial solar installations, which are affected by atmospheric conditions, geographical limitations, weather patterns and the day-night cycle, carefully selected orbital environments can provide satellites with near-continuous exposure to sunlight.
Certain sun-synchronous orbits offer particularly attractive conditions for solar energy generation. In these environments, satellite-mounted solar panels can receive sunlight for substantially longer periods, reducing dependence on large onboard battery systems.
Google’s research suggests that solar panels operating in suitable orbital environments could generate up to eight times more energy than comparable panels on Earth under its stated comparison assumptions.
For an industry increasingly constrained by electricity availability, this possibility is attracting considerable attention.
The proposition is relatively straightforward: deploy satellites equipped with solar arrays and powerful computing processors, connect them through high-speed optical communication systems, and establish an interconnected network capable of processing artificial intelligence workloads.
Rather than relying exclusively on electricity generated by terrestrial power grids, these orbital computing systems could potentially harness solar energy directly from space.
The implications extend beyond electricity generation. Moving certain computing workloads into orbit could also reduce pressure on terrestrial land, water resources and electricity infrastructure.
However, abundant sunlight is only one part of the equation. The challenge is converting that energy into reliable, economically competitive computing capacity.
Google: Project Suncatcher Takes AI Into Orbit
One of the most significant developments in this emerging industry is Google’s Project Suncatcher.
Announced in November 2025, the initiative explores the possibility of creating interconnected constellations of solar-powered satellites equipped with Google’s Tensor Processing Units (TPUs), specialised processors designed for machine learning applications.
The company’s long-term vision involves placing computing infrastructure aboard satellites and connecting these systems through high-bandwidth optical communication links.
In October 2026, Google confirmed that its first Project Suncatcher prototype had successfully entered orbit following a launch aboard SpaceX’s Transporter-18 mission.
Developed in partnership with Planet Labs, the experimental satellite is designed to test how Google’s computing hardware performs under the demanding conditions of space, particularly radiation exposure, temperature extremes and the physical stresses associated with orbital operations.
This is an important milestone, although it is essential to distinguish between an experimental satellite and a fully operational orbital data centre.
Google is still investigating whether the technology can eventually support computing infrastructure at commercial scale.
Its next development priorities include testing more sophisticated satellite-to-satellite communication systems and determining how computing hardware can operate reliably over extended periods.
Project Suncatcher nevertheless demonstrates that the concept is progressing beyond theoretical research and into practical engineering experimentation.
SpaceX: Elon Musk’s Ambition to Build an Orbital AI Infrastructure Network
Google is not alone in pursuing this opportunity.
SpaceX, led by Elon Musk, has outlined an ambitious strategy to develop orbital AI computing infrastructure.
The company’s June 2026 prospectus provides a particularly revealing account of its plans.
SpaceX intends to develop specialised satellites equipped with AI processors, larger solar arrays and sophisticated thermal management systems. These satellites would operate as interconnected computing platforms, potentially forming an extensive orbital AI infrastructure network.
A critical component of this strategy is Starlink.
SpaceX has already established a substantial satellite communications network incorporating inter-satellite laser links. These systems allow information to travel between satellites rather than relying exclusively on conventional ground-based communications infrastructure.
The company intends to build upon this technological foundation to establish the communication architecture required for orbital computing.
Another important component is Starship, SpaceX’s heavy-lift launch vehicle.
Deploying large quantities of computing equipment, solar arrays and supporting infrastructure into orbit would require an enormous transportation capability.
SpaceX believes that reusable launch vehicles, high-volume satellite manufacturing and its existing space infrastructure could provide it with a competitive advantage.
Its prospectus anticipates beginning deployment of orbital AI computing satellites as early as 2028.
The long-term ambitions are considerable. SpaceX has discussed the possibility of deploying up to 100 gigawatts of AI computing capacity into orbit annually.
This is an ambitious corporate target, not an existing operational capability or guaranteed outcome.
Importantly, SpaceX itself acknowledges that orbital AI computing remains technologically complex, unproven at commercial scale and subject to substantial execution risks.
Nevertheless, its involvement signals how seriously some of the world’s largest technology and aerospace companies are considering the future of computing beyond Earth.
The Emerging Competition: Starcloud and the Orbital Computing Industry
The development of orbital data centres is not restricted to established technology giants.
Starcloud, an emerging space infrastructure company, is also pursuing the opportunity.
In November 2025, the company launched a spacecraft carrying an NVIDIA H100 processor into orbit, demonstrating the potential for high-performance computing hardware to operate in space.
The significance of these developments lies in the possibility of creating an entirely new infrastructure industry.
Companies involved in satellite manufacturing, semiconductor technology, optical communications, aerospace engineering, advanced materials and energy systems could potentially become important participants in the emerging orbital computing ecosystem.
This could create opportunities extending well beyond the companies directly building the data centres.
The Engineering Challenges: Space Is Not an Easy Operating Environment
Despite the enormous potential, constructing data centres in space introduces engineering challenges that do not necessarily exist in terrestrial facilities.
1. Cooling and Thermal Management
Perhaps the most significant misconception surrounding orbital data centres is that the cold temperatures of space make cooling computer processors easier.
In reality, space presents a complicated thermal management problem.
On Earth, data centres use air conditioning, liquid cooling and other technologies to remove excess heat generated by processors.
Space, however, is effectively a vacuum. There is no surrounding air through which heat can be dissipated through conventional convection.
Orbital computing infrastructure would therefore require sophisticated heat pipes, thermal management systems and large radiators capable of transferring excess heat away from sensitive electronics.
The larger the computing installation, the more demanding this challenge becomes.
2. Radiation and Hardware Reliability
Electronic components operating in space are exposed to radiation that can damage semiconductor components and disrupt computing operations.
Specialised shielding, radiation-tolerant electronics and sophisticated error-correction systems will be necessary to ensure that processors operate reliably.
The ability to manufacture computing hardware that can withstand prolonged exposure to these conditions will be critical.
3. Communication and Data Transmission
An orbital data centre cannot operate effectively in isolation.
Large AI workloads frequently require multiple processors to communicate and exchange enormous quantities of information.
This means satellite constellations must establish extremely high-bandwidth communication links, potentially using laser-based optical systems.
Maintaining reliable connections between rapidly moving satellites introduces additional engineering complexity.
4. Launch Costs and Commercial Viability
Perhaps the most important commercial challenge is the cost of transporting infrastructure into orbit.
Solar panels, computing processors, radiators, satellite platforms and supporting equipment all have weight.
Even with reusable rockets, the financial viability of orbital computing will depend on substantially reducing launch costs while increasing payload capacity and satellite manufacturing efficiency.
Hardware failures and replacement costs must also be considered.
A data centre on Earth can be serviced, upgraded and repaired. Replacing sophisticated computing equipment in orbit is considerably more complicated.
These challenges will ultimately determine whether the technology becomes a commercially viable infrastructure alternative or remains a specialised application.
The Investment Opportunity: An Entirely New Technology Supply Chain
For investors, the most interesting aspect of orbital data centres may not necessarily be the companies operating the satellites.
It could be the industrial ecosystem required to make this infrastructure possible.
The development of orbital computing could create additional demand across several interconnected industries.
Semiconductors and Advanced Computing
AI accelerators, specialised processors, memory systems and advanced semiconductor manufacturing will remain fundamental to the computing infrastructure.
Companies developing high-performance processors and radiation-resistant electronics could benefit from the emergence of this market.
Solar Energy and Advanced Materials
Orbital computing systems will require highly efficient solar arrays capable of generating substantial electricity while minimising weight.
This could increase demand for advanced photovoltaic technologies, lightweight materials and sophisticated energy conversion systems.
Aerospace and Satellite Manufacturing
The construction and deployment of orbital computing infrastructure will require satellite platforms, launch vehicles, propulsion systems and advanced manufacturing capabilities.
Companies involved in reusable launch technology and high-volume satellite production could occupy strategically important positions in the emerging industry.
Optical Communications and Photonics
Interconnected orbital computing systems will depend on sophisticated communication technologies.
Laser-based communication equipment, photonic components and high-speed optical transceivers could become increasingly important as satellite constellations evolve into distributed computing networks.
Thermal Management and Advanced Engineering
The removal of excess heat from high-performance processors in a vacuum presents an entirely different engineering challenge.
Companies developing advanced radiators, thermal materials, heat pipes and specialised cooling systems could become important suppliers.
The potential investment opportunity therefore extends across multiple industries rather than being confined to a handful of aerospace companies.
The Commercial Reality: Will Space Replace Earth-Based Data Centres?
The possibility of moving AI computing infrastructure into space is exciting. However, investors should distinguish between technological feasibility and commercial competitiveness.
Boston Consulting Group’s August 2026 analysis provides an important perspective.
Its modelling estimates that orbital data centres currently carry a cost premium of approximately 2.5 to three times compared with terrestrial facilities.
Even under its realistic improvement scenarios, a cost premium could remain over the next decade.
This suggests that orbital data centres may not immediately replace conventional infrastructure.
Instead, they could initially serve specialised applications where their unique advantages justify the additional expense.
These could include processing information generated by Earth-observation satellites, space-based scientific instruments and other systems already operating beyond Earth’s atmosphere.
Over time, improvements in launch economics, satellite manufacturing, solar technology and thermal management could expand the range of commercially viable applications.
The eventual opportunity may therefore be complementary rather than competitive: terrestrial data centres and orbital computing networks operating as interconnected components of a much larger global AI infrastructure system.
The Final Frontier of the AI Infrastructure Race
The development of artificial intelligence is creating an infrastructure race that extends far beyond software applications and semiconductor performance.
The next competitive advantage may increasingly depend on who can secure sufficient electricity, deploy computing infrastructure efficiently and deliver processing capacity at commercially sustainable costs.
Google’s Project Suncatcher and SpaceX’s orbital AI computing ambitions demonstrate that the search for computing capacity is beginning to extend beyond the physical limitations of terrestrial infrastructure.
There are substantial engineering, financial and regulatory challenges ahead. Commercial deployment at meaningful scale remains unproven, and the economics have yet to be established.
Nevertheless, the underlying investment proposition deserves attention.
The convergence of artificial intelligence, renewable energy, semiconductor technology, satellite communications and reusable space transportation could eventually establish an entirely new segment of the global technology infrastructure market.
For investors, the question is no longer simply how much computing power the world will need.
It is increasingly about where that computing power will be generated, how it will be supplied with energy and which companies will provide the infrastructure necessary to make it possible.
The AI revolution began on Earth. Its next infrastructure frontier may be in space.
ROCK EDGE RESEARCH — Identifying tomorrow’s investment opportunities, today.
Investment Disclaimer: This publication is for informational and educational purposes only and does not constitute financial, investment, or other professional advice, nor a recommendation to buy or sell any security. The views expressed are those of the author and may change without notice. Forward-looking statements regarding space-based AI infrastructure and related technologies are speculative and subject to significant technological, commercial, regulatory, and execution risks. Investors should conduct their own research and consult qualified advisers before making investment decisions. Rock Edge Research accepts no liability for any losses arising from the use of this material.