Somewhere in Meta’s long-term energy planning, someone decided that the sun on Earth’s surface wasn’t enough. The company has agreed to purchase up to 1 gigawatt of electricity generated by solar panels in orbit and beamed down to the ground, a deal with startup Overview Energy that stands as the most unconventional power bet any major technology company has made in the race to fuel artificial intelligence.
The agreement, first reported by Bloomberg in late April 2026, would channel space-transmitted electricity directly to Meta’s AI data centers. At full capacity, 1 GW rivals the output of a large nuclear reactor. Because orbital solar arrays receive sunlight nearly around the clock, the system would operate at or near a capacity factor of 1.0, meaning it could theoretically deliver close to its full rated output continuously. On that basis, 1 GW could supply roughly 750,000 American homes, using the U.S. Energy Information Administration’s average household consumption of about 10,500 kilowatt-hours per year.
Meta has no interest in powering neighborhoods. It wants to power GPU clusters.
Why Meta is looking to orbit
The deal reflects a problem every company building large-scale AI systems now faces: electricity demand is growing faster than the grid can expand. Training a single frontier AI model can consume energy equivalent to tens of thousands of households over the span of weeks, according to estimates from researchers at Epoch AI. Meta, which disclosed plans in early 2025 to spend tens of billions of dollars on AI infrastructure, needs power sources that are massive, reliable, and available around the clock.
Terrestrial solar is cheap but only generates during daylight and depends on weather. Nuclear plants offer steady baseload power but take a decade or more to build. Space-based solar sits in a different category: arrays orbiting above the atmosphere receive sunlight nearly 24 hours a day, uninterrupted by clouds, nightfall, or seasonal shifts. The energy is converted onboard and transmitted to ground-based receiving stations, called rectennas, via microwave or laser beams.
The concept has a long pedigree. NASA studied it in the 1970s, and the U.S. Air Force Research Laboratory has funded space-based solar experiments for years. In June 2023, Caltech’s Space Solar Power Demonstrator (SSPD-1) successfully transmitted power wirelessly from a satellite in low Earth orbit, the first time any institution had done so from space. The amount of power was tiny, on the order of milliwatts, and the transmission distance was short. But it proved the physics work outside a lab. What no one had done, until this deal, is attach a Fortune 10 company’s name to a contract contemplating gigawatt-scale delivery.
What the deal actually says
The “up to” language matters. It signals a ceiling, not a guaranteed purchase volume. This structure is common in long-term power procurement tied to emerging technology: Meta is reserving capacity it can draw on as Overview Energy launches hardware and builds ground infrastructure, rather than committing to take delivery of 1 GW on day one.
Neither Meta nor Overview Energy has released financial terms. Whether the arrangement is structured as a traditional power purchase agreement, a milestone-based contract, or something else entirely will shape how seriously energy analysts and investors treat it. A phased approach, where Meta pays more as the system proves itself, would be a rational hedge. A large upfront capital commitment would signal deeper conviction.
Meta’s future capital expenditure filings may offer the clearest signal. The company’s quarterly earnings reports, which detail infrastructure spending, will be watched closely for any line items or commentary pointing to space-based energy investments.
The company behind the panels
Overview Energy is not a household name, and its public profile remains thin. The startup has discussed space-based solar power concepts publicly, but detailed information about its engineering leadership, manufacturing capacity, funding history, and prior hardware demonstrations has not been widely reported. For a project that would require launching massive solar arrays into orbit, maintaining them for years, and operating a ground-based power conversion system, the supplier’s track record matters as much as the buyer’s ambition.
The broader space-based solar sector is small but gaining momentum. The European Space Agency’s Solaris program is actively studying the feasibility of orbital solar for Europe’s energy mix. Caltech’s research continues. A handful of startups have entered the field, but none has operated a commercial-scale system. If Overview Energy’s deal with Meta progresses to actual hardware in orbit, it would represent a leap from laboratory proof-of-concept to industrial-scale energy delivery, a gap no one has bridged.
How it compares to other tech energy plays
Meta is far from the only tech giant scrambling for power. Google signed an agreement with Kairos Power for small modular nuclear reactors in late 2024. Microsoft struck a deal with Constellation Energy to restart Unit 1 at Three Mile Island in Pennsylvania. Amazon has become one of the world’s largest corporate buyers of clean energy. Each move reflects the same underlying pressure: AI workloads are pushing data center power consumption into territory that existing grid capacity and standard renewable contracts cannot easily cover.
What sets Meta’s deal apart is the sheer novelty of the supply source. Nuclear, geothermal, and next-generation battery storage are proven or near-proven technologies with established regulatory pathways. Space-based solar power has never delivered commercial electricity to anyone. The gap between concept and kilowatt-hour is wider here than in any comparable corporate energy agreement.
That gap also means the potential upside is larger. If space-based solar works at scale, it offers something no terrestrial source can match: near-continuous generation with a minimal land footprint. A rectenna takes up ground space, but the solar collection happens roughly 22,000 miles overhead in geostationary orbit. For a company that may need dozens of gigawatts within the next decade, that kind of scalability is worth exploring, even at high risk.
What has to go right
Turning this agreement into delivered power requires clearing several hurdles that no one has cleared at commercial scale.
Launch economics. Placing large solar arrays in orbit remains expensive. Even with falling launch costs driven by SpaceX and other providers, the capital required to deploy enough panels to generate 1 GW is substantial. The arrays must also be assembled or unfolded in space, adding complexity and risk to every mission.
Transmission efficiency. Converting sunlight to electricity, then to a microwave or laser beam, transmitting it through the atmosphere, and reconverting it to grid-compatible electricity on the ground introduces losses at every stage. Published estimates for end-to-end efficiency vary widely, and if too much energy is lost in transit, the cost per delivered megawatt-hour could dwarf terrestrial alternatives, even with the advantage of near-constant sunlight.
Regulation. Beaming concentrated energy from orbit to the ground raises questions that span multiple federal agencies. The Federal Communications Commission would likely need to approve spectrum use for microwave transmission. The Federal Aviation Administration would be involved in launch coordination. Safety regulators would want assurances about beam intensity, drift, and failsafe mechanisms. No clear permitting pathway for commercial space-to-ground power transmission exists in the United States today.
Cost competitiveness. Utility-scale terrestrial solar in the U.S. now comes in below $30 per megawatt-hour in favorable markets, according to recent Lazard LCOE analyses. Space-based solar would need to approach that range, or offer reliability and capacity advantages significant enough to justify a premium, for the economics to hold up against alternatives Meta could pursue on the ground.
Timeline. Even optimistic projections from space-based solar advocates place commercial-scale delivery years away. Overview Energy has not publicly committed to a date for first power delivery. Meta’s AI infrastructure needs, meanwhile, are growing now. The mismatch between urgency and readiness is the deal’s most obvious tension.
Skepticism from the ground
Not everyone views the announcement as a serious near-term energy play. Space-based solar has long attracted skepticism from energy analysts and aerospace engineers who question whether the economics can ever compete with rapidly cheapening terrestrial renewables paired with battery storage. The core concern is straightforward: even if the physics of orbital solar collection and wireless power transmission are sound, the cost of launching, assembling, and maintaining hardware in space may keep the delivered price of electricity far above grid alternatives for decades.
Grid experts have also raised practical questions about integrating beamed power into existing electricity infrastructure. A rectenna receiving a continuous microwave or laser beam would behave differently from a conventional power plant, and interconnection standards for such a source do not yet exist. Without a regulatory framework or a single commercial reference project, some observers see the Meta deal as closer to a research sponsorship than a binding energy procurement.
Supporters counter that similar skepticism greeted early commitments to offshore wind and utility-scale batteries, both of which eventually reached commercial viability after anchor customers helped drive down costs. Whether space-based solar follows the same trajectory or remains a stranded concept is the open question the Meta-Overview Energy deal now forces the industry to confront.
A contract written in ambition
The Meta-Overview Energy agreement is best understood as a statement of direction, not a delivery schedule. It tells the market that one of the world’s largest technology companies considers space-based solar power worth contracting for, even before the technology has been demonstrated at anything close to commercial scale.
Whether those orbital panels ever send a single watt to a Meta server rack depends on breakthroughs in launch logistics, transmission engineering, and regulatory cooperation that have not yet materialized. But Meta’s willingness to put its name, and presumably its capital, behind this bet reveals something about the AI industry’s power crisis that quarterly earnings calls and press releases tend to understate: the companies building the future of artificial intelligence are running out of conventional ways to keep the lights on.
Warren Cohen is a finance writer based in Phoenix, Arizona, covering personal finance topics including credit, banking, and beginner investing. He earned his degree in business administration from Arizona State University and began his career working in consumer finance, where he gained direct experience with lending and credit systems. He now writes for personal finance websites and fintech platforms, focusing on clear, practical content that helps readers make informed financial decisions.
