Artificial intelligence increasingly requires so much space and power that we may run out of both on Earth. As a solution, tech companies are looking to create data centres in space that harness solar power. While this would demand less cooling, it could create other costs and problems.
The enormous amount of data required by AI has led to data centres being built across the world, and “global power requirements could double by the end of this decade as companies train larger AI models”, said Scientific American. A move to space could “resolve long-standing challenges around powering data centre computation in a carbon-efficient manner”.
The sun’s rays are “direct and constant for solar panels to collect”, said The Wall Street Journal. There are “no clouds, no rainstorms, no nighttime”, and “demands for cooling could also be cut because of the vacuum of space”.
Tech giants, including Amazon, Google, OpenAI and Elon Musk’s xAI, are “running into physical limits to their AI ambitions on Earth”, said The Verge, but launch costs are “still prohibitively expensive”.
Space-based computing will “not become cost-effective unless rocket launch costs decline substantially”, said Scientific American. Experts also warn that these systems could have “even bigger environmental and climate effects than their earthly counterparts”.
Data centres “visible in the night sky at dawn or dusk” could disrupt astronomers who “rely on twilight to hunt for near-Earth asteroids” and it could worsen the space junk problem, as “more hardware is launched and more debris and fragments fall back through the atmosphere”. |