A team of engineers who built SpaceX's Starlink from whiteboard to mass production stepped out of stealth on June 26 with a startup that inverts the economics of the satellite industry: Eclipse Space argues that governments and enterprises should not have to rent space-based connectivity from someone else's constellation — they should be able to own one outright.
Starlink pedigree, a fabless model
Eclipse Space launched publicly from Redmond, Washington — the same city where SpaceX operates its primary Starlink production facility — after roughly a year of quiet development. Led by former SpaceX satellite payload engineering manager Derek Huerta, the company has about 30 employees, nearly half of whom came directly from the Starlink program, including 13 engineers from the constellation's earliest internal phase. The company has taken external investment from early-stage firms including Space Capital, Tectonic and Ubiquity, though it has not disclosed the amount.
Borrowing from the chip industry
Eclipse's core idea is lifted from semiconductors. In a fabless model, a company designs the product and owns the intellectual property while contracting physical production to specialized partners — the approach that let Qualcomm, Nvidia and Apple's chip division design processors without owning a factory. Eclipse will design spacecraft, retain the IP and define the manufacturing process, then hand assembly to regional partners. Customers — governments or enterprises — end up owning the constellation without first building a SpaceX-scale vertically integrated operation.
What ships first
The company plans to deliver prototype hardware before the end of 2026, including a phased-array antenna and a telemetry, tracking and command radio. A full demonstration spacecraft — a bus weighing under 100 kilograms — is targeted for 2027, carrying in-house radiation and plasma sensors to validate the architecture. Operational satellites are specified with dual S-band phased arrays, E-band backhaul, V-band inter-satellite links and eight kilowatts of power, with more than 20 spacecraft stacking on a single Falcon 9. Eclipse's first commercial target is direct-to-device connectivity, letting standard smartphones connect to low-Earth-orbit satellites without special hardware.
A sovereignty argument
The pitch is explicitly geopolitical. After Ukraine's wartime dependence on Starlink exposed the risks of renting critical infrastructure, sovereign satellite capability has become a national priority — the EU is building IRIS2 while China advances the Guowang and Qianfan megaconstellations. Eclipse aims at governments caught in between, too small for a state-scale program yet wary of depending on a foreign operator. The company has even embedded an AI engineering team, acquired from the group behind Rendered.ai's Agent Studio platform, directly into its spacecraft design workflow.
Eclipse enters a market where private capital is pouring into orbital infrastructure, from ICEYE's $450M Series F for SAR satellites to Star Catcher's orbital power grid and AST SpaceMobile's BlueBird direct-to-device push. Whether a fabless model can hit Starlink-class performance under regional assembly is the question its 2026 prototypes will begin to answer.
Reporting based on coverage from Tech Times and SpaceNews.