Google Takes Aim at Space-Based AI Data Centers with Project Suncatcher

Google's Project Suncatcher is taking its first step towards establishing AI data centers in space. The project aims to send AI processors into orbit aboard a SpaceX rocket, marking a significant move towards creating a full-fledged data center in space.

The test satellite will carry four processors that will run Google's Gemini AI models for 15 minutes at a stretch before needing to shut down and cool off. The ultimate goal is to create a network of satellites that can share the work of running AI models and beam their responses back to Earth.

Google says that using space-based data centers can help reduce the energy consumption of AI processing, which is projected to use around 3% of the world's electricity by 2030. In orbit, sunlight is abundant and free, with solar panels generating up to eight times as much power as they would on the ground.

However, scaling up this technology poses significant challenges. Each chip and solar panel must be hauled into orbit, and every answer those chips produce must come back down to Earth. Getting information between Earth and space is also a major issue, with some data compression being necessary due to limited bandwidth and energy savings.

Google's engineers have acknowledged these challenges but remain optimistic about the project's potential. 'As a first step, we tried to find reasons that it was impossible,' said Travis Beals, senior director at Google who worked on Project Suncatcher. 'But we gradually became convinced that it might actually work.'