Google plans to begin constructing space-based data centers by 2027, according to CEO Sundar Pichai. The goal of Project Suncatcher is to leverage solar energy more efficiently to power data centers, which consume a massive amount of energy. “One of our ambitious goals is to create space-based data centers to use solar energy more efficiently, which is 100 trillion times more abundant than what we currently generate on Earth,” Pichai stated. Initial steps towards realizing this vision will be undertaken in early 2027. Google, in collaboration with Planet, a company specializing in satellite Earth imaging, will launch two pilot satellites to test equipment in Earth’s orbit. Pichai believes that soon space-based data centers will become the new standard. “I am confident that in about ten years, we will consider this a commonplace method for building data centers,” he remarked.
Google is not the only company exploring space as a solution to the energy efficiency challenges of data centers. Previously, the startup Starcloud, backed by Y Combinator and Nvidia, launched its first satellite equipped with artificial intelligence. According to forecasts by Starcloud CEO Philip Johnston, carbon emissions from space data centers will be ten times lower than those from terrestrial ones, even considering launch emissions.
The power consumption of data centers is causing increasing concern due to the expanding computing power needed for AI development. In December 2024, the U.S. Department of Energy released a report stating that the load on data centers has tripled over the past ten years and could triple again by 2028. In 2023, data centers consumed more than 4% of the country’s electricity, and by 2028, they are expected to consume up to 12%. This project might be a crucial step towards more sustainable AI technology development. Utilizing solar energy in space could significantly reduce the carbon footprint of data centers and decrease their reliance on terrestrial energy sources. However, the cost of constructing such centers remains unknown, making the project’s implementation a challenging task.
