TLDRs
- Google is in talks with SpaceX to explore orbital data center infrastructure development.
- Project Suncatcher aims to test early space-based computing systems by 2027.
- Rising AI energy demand is pushing tech firms toward space power solutions.
- High launch costs remain the biggest barrier to viable orbital data centers.
The talks reflect growing concerns about rising energy demands from artificial intelligence workloads and the long-term limits of terrestrial data center expansion.
The initiative is tied to Google’s broader space computing effort, known as Project Suncatcher, which aims to test early-stage orbital infrastructure in the coming years. While still in its experimental phase, the project signals that Google is seriously evaluating space as a viable extension of its global cloud and AI computing network.
Space Data Ambitions Grow
Google’s discussions with SpaceX center on leveraging rocket launch capabilities to deploy experimental data center components into orbit. The idea is to eventually build systems that could harness uninterrupted solar energy in space while reducing reliance on Earth-based power grids.
The company previously revealed that it plans to launch two prototype satellites by early 2027 as part of Project Suncatcher. These early missions are expected to test the feasibility of operating computing systems in space, where conditions differ significantly from those on Earth, including radiation exposure and thermal management challenges.
AI Energy Pressure Mounts
The push toward orbital computing is being driven largely by the explosive energy requirements of artificial intelligence. Google’s data center electricity consumption has surged dramatically in recent years, rising to tens of millions of megawatt-hours annually as AI model training and inference workloads scale up.
Industry-wide, the pressure is even more pronounced. Analysts estimate that U.S. data centers could account for as much as 12% of total electricity consumption by 2028 if current trends continue. This looming demand has forced major tech firms to explore alternative infrastructure models that can support future growth without overwhelming existing energy systems.
Space-based infrastructure is being viewed as a long-term solution, particularly due to the potential for constant solar energy availability in orbit.
High Costs Remain Barrier
Despite its promise, orbital computing faces major economic hurdles. Estimates suggest that building a 1-gigawatt space-based data center could cost nearly three times more than constructing an equivalent facility on Earth. The financial gap is primarily driven by launch expenses, which currently sit at around $3,600 per kilogram of payload.
For space data centers to become economically viable, launch costs would need to fall dramatically, potentially below $200 per kilogram. Google’s own long-term projections suggest such reductions may only be achievable in the 2030s, assuming rapid advances in reusable rocket technology and space logistics.
Even industry leaders remain cautious. Executives across cloud and infrastructure firms have pointed out that key components such as server racks remain too heavy and that permanent, large-scale orbital installations have not yet been proven feasible.
Industry Race Intensifies
Google is not alone in exploring space-based computing. The broader industry is beginning to experiment with hybrid models that combine Earth-based infrastructure with early-stage orbital capabilities.
SpaceX itself is working with AI companies and partners on expanding compute capacity, while startups backed by major chipmakers are also investigating orbital data concepts. At the same time, firms like Anthropic have signed major agreements to secure large-scale terrestrial compute capacity, including access to high-density GPU clusters in the United States.
This combination of Earth-bound expansion and future-facing orbital planning highlights a transitional phase in the AI infrastructure race. Companies are scaling up traditional data centers while simultaneously preparing for a potential long-term shift into space-based systems.


