Google’s latest frontier for artificial intelligence may soon extend beyond Earth itself, as the company explores putting data centers into orbit to relieve pressure on land, power, and local communities.
Early ambitions and rising demand
The surge in AI compute has pushed tech and infrastructure companies to consider space as a new home for servers, motivated by abundant solar energy and escape from local opposition to massive ground facilities.1 But launching large amounts of hardware remains technically and economically difficult, with today’s terrestrial data centers still far cheaper once satellite construction and launch costs are included.1
Cowboy Space spots a bottleneck (2024–2028)
In response to that bottleneck, Cowboy Space Corporation—originally founded in 2024 as Aetherflux to beam solar power from space—pivoted to orbital data centers after realizing it could instead use that energy in orbit.2 The company concluded there “aren’t enough rockets to put data centers in orbit around the Earth, and they’re too expensive,” and that existing launch providers could not offer enough capacity or competitive unit economics compared with Earth-based centers.2
On May 11, 2026, Cowboy Space announced a $275 million Series B round at a $2 billion valuation to fund its own rocket program, aiming for a first launch by the end of 2028.2 Its plan is to integrate data centers directly into the second stage of custom-built rockets, bypassing today’s limited launch capacity.
Google, SpaceX and the orbital data center race (2026–2030s)
On May 12, 2026, reports emerged that Google and SpaceX are in talks to launch orbital data centers, with SpaceX pitching investors on space as the “cheapest place to put AI compute within the next few years” as it prepares a massive IPO.1 Elon Musk has publicly touted orbital data centers as cheaper to operate and less exposed to local backlash, while advocates say space sidesteps the land-use fights facing U.S. data center buildouts.1
Google is also speaking with other rocket-launch companies and pursuing its own initiative, Project Suncatcher, which targets prototype satellite launches by 2027 and full concepts not until the mid‑2030s.12 That longer timeline contrasts with SpaceX’s aggressive framing and Cowboy Space’s 2028 launch goal, underscoring a key tension: visionaries see orbital compute as inevitable, but rockets, costs, and timelines remain stubborn constraints.