Google and SpaceX Explore Orbital Data Centres Amid AI Infrastructure Pressures

Google is reportedly in discussions with SpaceX regarding a potential rocket launch agreement to deploy data centres into orbit, according to reports from the Wall Street Journal. The initiative, known as Project Suncatcher, seeks to address the escalating energy and land requirements of artificial intelligence infrastructure by utilising space-based solar power. Google Research has outlined plans to deploy prototype satellites by 2027 in a dawn-dusk sun-synchronous low Earth orbit, a trajectory designed to maximise exposure to continuous sunlight.

The push for orbital infrastructure is driven by the urgent need for expanded computing capacity as the AI boom intensifies. Current terrestrial data centres are vast, energy-intensive facilities that consume significant amounts of electricity and water. SpaceX chief executive Elon Musk has publicly argued that global electricity demand for AI cannot be met through terrestrial solutions, stating that space-based AI is necessary for long-term scaling. Musk has previously noted that harnessing even a fraction of the Sun’s energy would require vastly more power than current civilisation uses.

Despite the strategic appeal, Google Research has identified several technical obstacles that must be overcome. In a vacuum, heat cannot be dissipated via air as it is on Earth; instead, systems must radiate heat carefully. Furthermore, the lack of immediate access for repair crews necessitates new engineering standards to ensure autonomous operation. The project also faces the challenge of increasing congestion in low Earth orbit, with independent researchers warning that collision risks are rising in already crowded regions.

Economic viability remains a critical barrier to implementation. While SpaceX currently charges tens of millions of US dollars per launch mission, analysts indicate that orbital data centres would need launch costs to drop significantly to be financially feasible. Specifically, costs would likely need to fall to below $1,000 per kilogram to make the venture economically sensible.

The status of the talks between Google and SpaceX is based on reports from the Wall Street Journal and has not been independently confirmed by both parties in this source. The timeline for Project Suncatcher prototypes is subject to clearing the aforementioned technical and economic hurdles, leaving the feasibility of space-based AI infrastructure as a long-term prospect rather than an immediate reality.