Google plans to determine solar-powered synthetic intelligence knowledge facilities in area to unravel Earth’s power insufficiency.
Google is planning a radical step to beat the restrictions of energy-hungry synthetic intelligence knowledge facilities on Earth by sending AI chips to area on solar-powered satellites. This groundbreaking challenge, which the corporate calls “Challenge Suncatcher,” goals to determine space-based knowledge facilities that run on steady photo voltaic power. This is not going to solely present uninterrupted processing energy with clear power but additionally scale back the environmental influence brought on by excessive electrical energy consumption on Earth.
“House could also be the very best place to scale AI sooner or later,” says Travis Beals, a senior director at Google. As a part of the challenge, which continues to be within the experimental stage, Google plans to put its personal Tensor Processing Unit (TPU) chips on satellites outfitted with photo voltaic panels. These panels will be capable to generate power eight instances extra effectively than these on Earth.
First Trial Satellites to Launch in 2027

Nevertheless, Google faces important technical hurdles. The corporate should set up connections between satellites able to transferring tens of terabits of information per second. Moreover, the chips should stand up to the extraordinary radiation in area. Google claims its Trillium TPUs can endure the five-year mission period.
The challenge can be formidable by way of price. Though sending a knowledge middle to area is presently very costly, Google’s evaluation means that by the mid-2030s, a middle in area may function at roughly the identical price as one on Earth, on an power foundation. Because of this, the corporate plans to launch its first trial satellites in 2027 in collaboration with a agency referred to as “Planet.”
You Would possibly Additionally Like;
Observe us on TWITTER (X) and be immediately knowledgeable in regards to the newest developments…
Copy URL
Observe Us

