Welcome to Eye on AI with AI reporter Sharon Goldman. In this edition: Data centers in space are possible, but not ready for launch…accenture AI links promotions to logins… AI pioneer Fei-Fei Li's startup World Labs raises $1 billion. nvidia deal meta Signals a new era in computing power.
The AI industry is literally on a power trip and it's getting desperate. Data centers already account for about 4% of US electricity use, a share expected to more than double by 2030 as gigawatts of electricity are required to run and train AI models. Analysts estimate global data-center electricity demand could grow as much as 165% by the end of the decade, even though new generation and transmission infrastructure is years behind what is needed. In response, hyperscalers are making deals to build their own gas plants, exploring small nuclear reactors, and looking for electricity wherever they can find it.
Against that backdrop, it's not surprising that some of the industry's biggest players are starting to look to outer space for solutions.
in a feature story published This morning, I'm reflecting on how — even as tech companies are on track to spend more than $5 trillion globally on Earth-based AI data centers by the end of the decade — Elon Musk is arguing that the future of AI computing power is in space, powered by solar power. Musk has suggested that economics and engineering could align within a matter of years, even predicting that within five years there could be more AI computing capacity in orbit than on Earth.
The idea of orbital space stations themselves is not new. As far as 2015 is concerned, Luck Was already asking the question: What if we put servers in space??
What has changed is the urgency. Today's power outages have pushed the concept back into serious conversation, with startups like StarCloud paying attention and big tech leaders like Ex. Google CEO Eric Schmidt, Alphabet CEO Sundar Pichai and Amazon's Jeff Bezos are all turning their attention to the possibilities of launching data centers into orbit.
However, while Musk and other bulls argue that space-based AI computing could become cost-effective relatively quickly, many experts say anything coming at a meaningful scale is decades away. Barriers related to power generation, heat dissipation, launch logistics, and cost still make it impractical – and for now, the overwhelming majority of AI investments are flowing into terrestrial infrastructure. He argues that small-scale pilots of orbital computing may be possible in the next few years, but space remains a poor substitute for Earth-based data centers for the foreseeable future.