Google will launch a satellite equipped with Artificial Intelligence chips next week to test Project Suncatcher. This initiative aims to investigate the possibility of solar-powered computing structures being installed in space in the future.
As reported by The New York Times, this test will also serve to analyze how components behave under adverse conditions, such as radiation, extreme temperatures, and vacuum. A crucial aspect to verify is the system developed to dissipate heat generated by the chips without relying on air circulation.
The launch will be conducted in collaboration with Planet Labs, utilizing SpaceX's shared Transporter-18 flight. In low Earth orbit, the satellite will carry four of Google's Tensor Processing Units (TPUs) to confirm if they can execute AI tasks accurately.
The concept for this idea arose in response to the growing energy consumption required to operate artificial intelligence systems. In orbit, future data centers would have the advantage of receiving sunlight for significantly longer periods, overcoming certain restrictions imposed by current terrestrial infrastructure.
Currently, the mission is in its initial phase, being only a prototype. The primary goal is to collect data in orbit and identify potential flaws in the design.
Key technical challenges
One point of attention is the heat generation by the chips during information processing. While fans assist in dispersing this heat in conventional computers, there is no air available in the space environment for this function. To overcome this, Google implemented a solution based on heat pipes and radiators, which transfer heat from the components to structures capable of releasing it into space.
Radiation represents another significant challenge, as it can cause modifications to the data processed by the components, a phenomenon known as 'bit flips'. Although Google has already subjected its TPUs to tests at the University of California, Davis, the experience in orbit will be indispensable for understanding the hardware's reaction under real conditions.
Looking ahead, Google plans to launch two satellites in 2027 to test high-bandwidth laser connections. This type of communication will be essential for connecting future clusters of space computers.
Other companies, such as SpaceX and Starcloud, are also researching computing structures in space. However, experts point out that the technology is still far from commercially viable operation due to factors such as high launch costs, engineering complexities, and bottlenecks in satellite manufacturing.
Project expectations
James Manyika, Senior Vice President of Research at Google, expressed moderate expectations for the project. He stated that, being completely frank, they do not expect anything operationally useful in the coming years, comparing the effort to a long period of research that preceded the development of the company's autonomous vehicles.
Therefore, the first Suncatcher flight is not intended to deliver a functional orbital data center. Google's focus is primarily on determining whether the fundamental elements can survive and operate in the space environment before scaling the idea to something larger.
