A Troy, Michigan startup has signed a 20-year contract to beam solar power down from orbit into an underwater data center, betting that microwave energy from space can solve the AI industry's ballooning appetite for electricity. Virtus Solis Technologies will supply 100 megawatts of orbital solar baseload power to Brae Systems, a Chicago-based developer building modular data centers in marine environments, delivering an estimated 876,000 megawatt-hours of electricity annually once the arrangement is fully operational.
The deal, reported by pv magazine USA, includes an option letting Brae Systems expand its draw to as much as 250 megawatts within three years of initial commercial operations. If that option is exercised, Virtus Solis would deliver 2.19 million megawatt-hours annually, turning what is already billed as a landmark contract into what the companies describe as a multi-billion-dollar lifetime commitment. Brae Systems intends to use the power for underwater servers designed to simplify cooling and sidestep land-use restrictions that have slowed conventional data center construction.
Virtus Solis plans to build its orbital arrays by launching thousands to potentially millions of small modular satellites, each about 1.65 meters across, that autonomously link together in orbit to form structures spanning half a kilometer to three kilometers, according to Space Frontier Foundation. Each unit pairs solar cells with a 10 GHz radio-frequency transmitter, and the company will assemble the hexagonal, snap-fit modules in orbit with help from a dedicated space tug satellite. The approach is meant to avoid the cost and complexity of launching one massive monolithic structure.
Why Molniya Orbits and Why Now
Virtus Solis plans to place its stations in highly elliptical Molniya orbits, which have a 12-hour orbital period and keep a station in view of ground power receivers for eight hours at a stretch, meaning three units are needed for full-time coverage. Company co-founder and CEO John Bucknell said Virtus Solis chose the orbit specifically to take advantage of reusable rockets and cut launch costs. Bucknell, who spent 15 months at SpaceX starting in 2011 helping develop the Raptor engine that powers Starship, founded Virtus Solis in 2019, according to background reported by Latitude Media.
The timing lines up with a milestone in the rocket program Virtus Solis is counting on: SpaceX's Starship Super Heavy reached low-Earth orbit for the first time on September 28, demonstrating the kind of high payload capacity and reusability the company says makes orbital solar financially feasible. Virtus Solis is also working with roughly six vendors to manufacture a photovoltaic cell it expects will be perovskite-based, and the company says its techno-economic model runs on 240 separate parameters.
Federal money is already backing part of the technology. In January 2025, Virtus Solis received a $1.92 million award from the Department of Energy's Advanced Research Projects Agency-Energy under its Vision OPEN 2024 program, funding meant to develop long-distance, high-efficiency wireless power transfer electronics, per the agency's own project documentation. That hardware is expected to feature in the company's planned March 2027 public demonstration, where Virtus Solis is aiming for end-to-end wireless power transfer efficiency exceeding 50 percent — a steep jump from the roughly 4 percent efficiency recorded in prior industry demonstrations, Latitude Media reports.
Ground Receivers Designed to Share Land With Farming
On the ground, ground receivers will have roughly the footprint of utility-scale solar arrays and the form factor of a First Solar Series 7 panel, but with a twist: the panels are designed to be about 80 percent transparent to sunlight, allowing them to support agrivoltaics underneath. Virtus Solis says existing solar arrays could even be repowered to receive space-based solar transmissions instead of being built from scratch. The microwave energy beamed down is described as about equal in intensity to mobile-phone energy.
The first receiving station will be built in Illinois, co-located with a pilot data center, giving regulators and the public an early look at how the system performs before wider rollout. Separately, Virtus Solis plans to fly a 100-kilowatt pilot solar plant in low-Earth orbit within about 24 months, intended to let federal regulators and the public evaluate an operating space-based power plant ahead of commercial scale-up. Company officials also note the Molniya orbit approach can serve southern latitudes that other orbital configurations struggle to reach.
Vishnu Indukuri, speaking on behalf of Brae Systems, said the company's direct 20-year supply of firm clean power will provide predictable power costs and zero carbon emissions — a pitch aimed squarely at data center operators spooked by volatile electricity markets. That pitch arrives as global data center electricity consumption is projected to nearly double, from 485 terawatt-hours in 2025 to 950 terawatt-hours by 2030, driven largely by AI computing workloads, according to the International Energy Agency. That crunch is pushing developers like Brae Systems toward unconventional sites, including underwater ones, a trend Hoodline has covered before in the context of a Dallas startup chasing floating data centers to escape Texas water fights.
Underwater Cooling Has Precedent, but Not at This Scale
The subsea concept Brae Systems is pursuing has real-world grounding. Microsoft's Project Natick experiment showed in 2020 that underwater server pods had a failure rate one-eighth that of land-based control groups, relying on surrounding ocean water for passive thermal cooling, and Microsoft concluded the project in 2024 after validating those reliability benefits. China has already pushed the concept to commercial scale, launching full commercial operations for an underwater data center off Shanghai in May 2026 following a year-long trial, as previously reported by The Maritime Executive.
Still, commercial subsea GPU deployment at megawatt scale remains untested in U.S. waters, even as the underlying cooling science has been proven out. Virtus Solis and Brae Systems are targeting first commercial power delivery around 2030, a timeline that requires Virtus Solis's satellite constellation rollout to align with Brae's planned deployment of its first megawatt-scale subsea data center, per Latitude Media's reporting on the agreement.
The Economics Remain Unproven
A January 2024 NASA report offers a sobering baseline for the economics at stake. It estimated that space-based solar power operating in 2050 would face levelized electricity costs of $0.61 to $1.59 per kilowatt-hour under current technology, but found those costs could fall to as little as $0.03 to $0.08 per kWh with major advances in launch pricing, cell efficiency, and automated in-orbit assembly. Satellite manufacturing and launch currently account for more than 90 percent of space-based solar's lifecycle costs, underscoring how much the entire business case depends on cheaper, reusable rockets like Starship actually delivering on their promise.
That leaves the Virtus Solis-Brae Systems agreement resting on two assumptions that remain unproven: ultra-low launch costs driven by Starship, and a dramatic jump in wireless power transmission efficiency from roughly 4 percent to more than 50 percent at the March 2027 demonstration. Both companies are early-stage startups attempting unproven technological models on a 2030 commercial timeline, and while the multi-billion-dollar contract signals real corporate appetite for novel baseload clean power, it remains contingent on both sides hitting a string of technical milestones in the years ahead.
