For decades, space-based solar power has existed mostly in academic papers and aerospace research. Recently, however, it has begun to appear in a different context: discussions around the long-term energy requirements of artificial intelligence.
A growing number of founders and technology leaders working at the intersection of AI, including Lado Okhotnikov, founder of Holiverse, noted that AI’s expanding demand may eventually test the limits of Earth-based power systems. Among the speculative ideas under discussion is space-based solar energy — the concept of collecting sunlight in orbit and transmitting it to Earth — which some teams are exploring as a long-term research direction rather than a near-term solution.
The discussion is not about imminent deployment. Instead, it reflects a broader shift: as AI scales, conversations about infrastructure and energy are moving beyond incremental improvements.
The Growing Appetite for Energy
AI infrastructure operates continuously, with little tolerance for power interruptions. Data centres, the backbone of AI, run 24/7 and require stable electricity to support training and inference workloads.
Estimates from energy research agencies suggest that global data centres consume roughly 1–2% of the world’s electricity today, with AI workloads accounting for a meaningful and growing portion. Some forecasts indicate that data-centre energy demand could double over the next decade as AI adoption and cloud computing expand.
In the United States, data centres already represent a meaningful share of total electricity consumption, with estimates around 4 % and projections pointing to continued growth as AI deployments scale. Renewable sources supply an increasing share of energy, but intermittency, grid congestion, and geographic mismatches between supply and demand present ongoing challenges.
Why Space-Based Energy Is Being Considered
The idea of orbital solar power is simple in principle. Solar panels in space receive continuous sunlight, unaffected by night cycles or weather. Energy could be transmitted wirelessly — via microwave or laser — to Earth-based receiving stations, then converted into electricity for the grid.
The physics behind the concept is well established. Small-scale wireless power transmission has been demonstrated, and space agencies have studied orbital solar power for decades. In theory, it works. For instance, the European Space Agency and Caltech have demonstrated small-scale wireless power transmission from space prototypes back to Earth, showing that the physics is sound.
The challenge is one of scale. Supporting even a fraction of global AI infrastructure would require orbital systems far larger and more complex than anything currently deployed in space.
Building Infrastructure Above Earth
Deploying solar arrays in orbit involves unprecedented engineering and investment. Arrays must be manufactured, launched, and assembled, potentially with autonomous or semi-autonomous systems. Maintenance is another concern: radiation, micrometeoroids, and material degradation all affect lifespan, and repairs in orbit are slow and costly.
On Earth, receiving stations would require upgrades to grid infrastructure to handle high, continuous energy flows safely. Taken together, these challenges highlight why orbital solar power remains a long-term research avenue rather than a practical near-term solution.
Reliability and New Dependencies
AI workloads are sensitive to power disruptions. Introducing orbital energy adds new layers of dependency: power would rely not only on terrestrial grids but also on infrastructure far beyond Earth’s surface, exposed to space weather, debris, and technical failures.
Whether such systems could match or exceed the reliability of ground-based generation — without extensive redundancy — remains unknown. Even with improved grid integration and energy storage, orbital power introduces risks that terrestrial systems alone do not face.
Space-based energy raises complex governance questions, as discussed by the International Academy of Astronautics and the UN Committee on the Peaceful Uses of Outer Space. Who would own orbital infrastructure? How would transmission be regulated? Which countries or organizations would have priority if supply is constrained?
Unlike national electricity grids, which operate under established frameworks, orbital systems would likely require new international agreements covering safety, access, and dispute resolution. Currently, such frameworks do not exist.
A Marker, Not a Forecast
As experts likeLado Okhotnikov, founder of Holiverse, emphasize, there is no deployment timeline for AI powered by space-based solar energy. Economics, technology, and governance all remain uncertain. Yet its presence in public discussions signals a shift: as AI’s electricity demand grows — potentially doubling in the next decade — industry leaders are forced to think beyond incremental improvements.
Efficiency gains in hardware, advanced cooling techniques, smarter grids, and better energy management are likely to handle most of the load. Space-based energy, for now, functions more as a thought experiment than a prediction. It reflects a recognition that Earth-based systems may eventually face limits, and that technologists are beginning to consider the possibilities that lie beyond them.
Whether AI data centres will ever be powered from orbit remains uncertain. What is clear is that the discussion itself illustrates the scale of the challenge and how far the industry’s thinking has already extended.