Space-Based Solar Just Signed Its First Paying Customer

John Bucknell spent his early career pushing propellant through rocket engines, a discipline that rewards moving energy through a hostile environment and losing as little of it as possible. In 2019 he aimed that instinct at a stranger problem: sending electricity down from orbit.

On September 30, the Michigan company he founded, Virtus Solis, signed the first commercial power-purchase agreement for space-based solar. Brae Systems, a Chicago developer building GPU data centers on the sea floor, agreed to buy 100 megawatts of round-the-clock beamed power — 876,000 megawatt-hours a year — with an option to reach 250 MW within three years. It is the first time anyone has contracted firm baseload electricity generated in space.

A power plant that never sees night

The premise is a plain asymmetry. A solar farm on the ground fights clouds, winter and the daily loss of the sun. A satellite in the right orbit does not. Virtus Solis parks its arrays in Molniya orbits, long elliptical loops that hold a spacecraft in sunlight for most of a twelve-hour lap while it lingers high over the northern hemisphere.

Each array is a bed of 1.65-meter solar tiles assembled in orbit by robots, converted to radio waves, and beamed to a receiver on the ground. The company's engineering target is 70% end-to-end efficiency. A first public demonstration of a complete wireless transfer is not due until March 2027.

What Brae is actually buying

Brae is not a utility. It sinks sealed GPU data centers to the seabed, where cold water handles cooling and land is not a constraint. Water solves heat. It does not solve the power bill, or the years a large new load can spend waiting in an interconnection queue. Brae is betting orbital generation skips both.

AI workloads require an unprecedented leap in both energy availability and thermal efficiency. Brae engineered underwater data centers to unlock natural free liquid cooling and bypass terrestrial bottlenecks.— Vishnu Indukuri, chief executive, Brae Systems

The sale is narrower than the slogan. Bucknell is marketing the grid's constraints, not its carbon.

“High-performance AI data centers demand uninterrupted, gigawatt-scale power that strained terrestrial grids simply cannot deliver cleanly,” he said. The beam lands without a transmission line, a battery farm or a permitting fight.

The money arrived first

Capital has moved ahead of the hardware. Tracked investment in space-based solar power rose from about $1.07 billion in 2024 to $1.9 billion in 2025 and roughly $5.3 billion in 2026, by the count of analysts who follow the sector. Meta has reserved up to a gigawatt from a rival, Overview Energy. Star Catcher raised $65 million in May and is preparing an orbital power-beaming demonstration this fall.

As we wrote in September, Google's Project Suncatcher pushed the first AI chips into orbit to test whether compute can live where the sun never sets. The two bets are cousins. One moves compute to the energy; the other moves the energy to the compute.

The direction of travel is a response to a real squeeze. Data centers are forecast to add 80 to 160 terawatt-hours of demand by 2030, and the grids that would serve them are already congested. A seller who can promise firm power without a queue has something worth paying for — if the promise holds.

How the beam reaches the ground

Beaming power is older than the satellites that would carry it. Researchers moved energy by microwave in the 1960s, and NASA published a full space-solar reference design in 1980. What blocked the idea was never the transmitter. It was the cost of lifting mass into orbit and the difficulty of assembling large structures once there.

Both constraints have moved. Reusable launch has pulled the price per kilogram down by an order of magnitude, and rideshare missions now fly small payloads on a schedule rather than a hope. Robotic assembly is the harder half, and it is the subject of Virtus Solis's 2024 partnership with Orbital Composites, meant to build the arrays tile by tile in medium-Earth orbit.

The transmission itself is deliberately undramatic. Virtus Solis uses radio-frequency beams rather than a laser, on the argument that radio passes through weather with almost no loss and stays diffuse enough to be safe across a receiving field. The company claims 99.9% uptime and a beam no more intense, in aggregate, than ordinary sunlight.

Competitors have picked different trades. Star Catcher points optical beams at satellites already in orbit, selling power to spacecraft rather than to the ground. Overview Energy aims a near-infrared beam at existing terrestrial solar farms so they keep producing after dark without new land. Cowboy Space, formerly Aetherflux, is chasing a mix of remote sites and orbital compute. The approaches disagree about the customer, not about the physics.

Government interest has followed the capital. Space-solar power beaming has drawn feasibility contracts from agencies that want power at remote or austere sites, a market where the alternative is diesel hauled in at high cost.

What the last first looked like

Space-based solar has a precedent, and it is cautionary. In 2009 Pacific Gas and Electric signed a 200 MW power-purchase agreement with a startup called Solaren, to be delivered from orbit starting in 2016. The contract was cancelled in 2015 before a single watt arrived, after Solaren failed to finance the hardware.

Bucknell argues the inputs that killed it have since collapsed. Solar cells, power electronics and, above all, launch costs sit at a fraction of their 2009 levels. He is pitching an eventual price near $30 per megawatt-hour, roughly the cost of firm fossil generation. That figure is a target, and targets have a poor record in this sector.

The scale is also easy to misread. A 100 MW plant is not one machine. Virtus Solis's own architecture implies on the order of 100,000 small satellites for that capacity, assembled in orbit, each carrying part of the array. The March 2027 demonstration will beam power across a lab, not from a constellation.

Does orbital baseload survive its first dispatch?

The agreement is real, and so is the money behind it. What remains unproven is the claim underneath the paperwork: that electricity beamed from orbit can cost less than the grid it avoids. Until a receiver in Illinois actually lights up, this is a contract with the future.