Meta Platforms has signed a deal to power its artificial intelligence data centers with solar energy harvested in orbit and beamed to Earth, Bloomberg reported in late April 2026. The agreement marks the first known contract between a Big Tech hyperscaler and a space-based solar provider, and it underscores just how aggressively the company is hunting for electricity sources that can keep pace with the explosive growth of AI workloads.
The concept is deceptively simple. Solar panels stationed in geostationary orbit collect sunlight around the clock, unblocked by clouds, nightfall, or seasons. That energy is converted into a focused microwave or laser beam aimed at a ground receiver, which turns it back into grid-ready electricity. For a company whose AI training clusters can burn through megawatts continuously for weeks at a stretch, the promise of truly uninterrupted clean power is hard to ignore.
Why Meta is looking beyond the grid
AI models have become voracious consumers of electricity. Training a single large language model can require tens of megawatts sustained over months, and inference demand grows every time a new feature rolls out to billions of users. The International Energy Agency projected in early 2026 that global data center electricity consumption could more than double by 2030, with AI accounting for the largest share of that increase.
Meta has already moved to lock in conventional clean energy. The company holds long-term power purchase agreements for terrestrial wind and solar farms across multiple continents, and reporting throughout 2024 and 2025 indicated it was exploring nuclear options alongside rivals Microsoft and Amazon. But traditional renewables come with well-known limits: solar panels stop producing after sunset, wind turbines depend on weather, and grid congestion in key data center corridors from northern Virginia to the Netherlands has made new interconnections slow and expensive.
Space-based solar, at least in theory, sidesteps all of those problems. An orbital array in geostationary orbit receives sunlight more than 99 percent of the time, and a dedicated beam link could deliver power directly to a receiver near a data center campus, bypassing congested transmission networks entirely.
The science is real, but the engineering gap is wide
The physics behind orbital power beaming have been studied for decades. A technical report produced by Lawrence Livermore National Laboratory and published by the U.S. Department of Energy around 2004 laid out the feasibility of laser and optical power transmission from space, examining system design, performance modeling, and atmospheric interference. While that study is now more than two decades old and predates the current AI energy boom, it established that the concept rests on proven electromagnetic principles, not speculation.
More recently, the field has moved from paper studies to hardware. In 2023, Caltech’s Space Solar Power Demonstrator (SSPD-1) successfully transmitted energy wirelessly in orbit for the first time, a milestone that shifted the conversation from “could this work?” to “how fast can it scale?” The European Space Agency’s SOLARIS initiative and the UK’s Space Energy Initiative have both funded feasibility work aimed at commercial deployment within the next decade.
Still, the distance between a laboratory proof of concept and a commercial power plant in the sky is enormous. Every conversion step, from sunlight to electricity, electricity to beam, beam through the atmosphere, and beam back to electricity, loses energy. No publicly available peer-reviewed study has yet demonstrated end-to-end efficiency figures that would make orbital solar cost-competitive with terrestrial alternatives at scale. Launch costs have fallen dramatically thanks to reusable rockets, but lofting the massive solar arrays and beam-forming hardware needed for gigawatt-class output would still require hundreds of heavy-lift flights.
What the deal does and does not tell us
Bloomberg’s reporting confirms that Meta has entered into an agreement, but key details remain undisclosed. The company has not named the orbital provider, specified the expected power output, or indicated when it anticipates receiving its first kilowatt-hours from space. Whether the contract is a firm power purchase agreement, a research partnership, or an option on future capacity is unclear. Meta has not filed any related disclosure with the Securities and Exchange Commission, and no company executive has spoken on the record about the arrangement.
That ambiguity matters. A binding, large-scale PPA would signal genuine confidence that the technology is approaching commercial readiness. A smaller research agreement or option contract would suggest Meta is placing an early bet to secure priority access if the technology matures, a hedging strategy rather than a power procurement plan.
Regulatory and community hurdles ahead
Even if the engineering works, beaming concentrated energy from orbit to Earth raises questions no government has fully answered. Aviation authorities would need to verify that power beams pose no risk to aircraft flying through or near the transmission corridor. Spectrum regulators would need to allocate and protect the frequencies used for energy transfer, ensuring they do not interfere with communications satellites or ground-based wireless networks. As of May 2026, no country has enacted a comprehensive regulatory framework for receiving beamed power from space.
On the ground, siting a receiver station presents its own challenges. A rectifying antenna (rectenna) for microwave beams or a tuned photovoltaic array for laser transmission would require a significant land footprint near the data center or a robust grid connection to move the power. Communities that have already clashed with data center developers over noise, water use, and transmission line construction may view an unfamiliar energy receiver with skepticism, particularly if the technology is poorly explained or perceived as hazardous.
What this signals for the industry
Meta’s willingness to put a contract behind space-based solar, even with so many unknowns, sends a clear message to the energy and aerospace sectors: the largest electricity buyers on the planet are running out of conventional options and are ready to fund unconventional ones. That signal alone could accelerate venture investment in orbital power startups, push government agencies to fast-track feasibility studies, and prompt regulators to begin drafting rules.
The company is not acting in isolation. Microsoft has signed nuclear power agreements, including a deal to restart a unit at Three Mile Island. Amazon has invested in small modular reactor developers. Google has pursued geothermal and advanced nuclear contracts. Each of these moves reflects the same underlying pressure: AI workloads are growing faster than clean electricity supply, and the companies driving that growth are competing to lock in every viable source they can find.
Space-based solar sits at the far end of that spectrum, higher risk but potentially higher reward. If the technology scales, it could offer something no earthbound source can match: continuous, weather-independent, land-light clean power delivered on demand. If it stalls, Meta will have spent relatively little compared to its overall capital budget and will still hold options on a technology that may mature later.
Why the gap between ambition and delivery still matters
For now, no orbital solar system has transmitted commercially meaningful amounts of power to the ground. No independent audit has validated the cost projections that space-solar companies use to attract investors. And no regulator has approved a ground receiver for operation. Meta’s deal is best understood not as proof that space solar works, but as proof that the AI industry’s appetite for clean electricity has grown large enough to make even the most ambitious energy bets look rational.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
