Launched on May 14 aboard a Long March 2D rocket from the Jiuquan Satellite Launch Center, these satellites are designed to diminish dependence on ground-based computers. The cold vacuum of space will be utilized for natural cooling while processing data with a total capacity of 1,000 peta operations per second, as asserted by the Chinese government.
Wang Jian from Zhejiang Lab underscored the notion of space as a venue for AI applications during the Beyond Expo Technology Conference. He remarked on the resurgence of opportunities for exploring future applications in the decades ahead.
These satellites will communicate using lasers, with one featuring an X-ray polarization detector for cosmic research. Each satellite manages an AI model containing 8 billion parameters, executing 744 tera operations per second, with the potential to scale to five peta operations in total. This pioneering array represents the first operational implementation of its kind, contrasting with existing land-based technologies like Microsoft’s AI Copilot+, which operates at approximately 40 tera operations per second.
The ainewsarticles.com article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
