The Guardian has published a behind-the-scenes piece about how its global technology reporting team has increasingly shifted from covering screens and software to investigating the physical infrastructure underpinning the AI boom, chiefly the vast datacentres being built worldwide. The change matters because these structures come with tangible constraints — power, water, land, chips and grid capacity — that determine whether the industry's grand promises are realistic, and reporters argue that these physical realities are what will ultimately make or break the AI expansion.

The article highlights a recent investigation by AI reporter Aisha Down, who travelled to rural Scotland and found that an £8.2bn AI complex had misrepresented its plans to be powered entirely by on-site renewables; a related earlier story revealed that a four-acre London site earmarked for an AI supercomputer was still being used as a scaffolding yard. Global technology editor Dan Milmo separately reported on the growing number of large datacentre projects being challenged or cancelled around the world, citing a well-funded Welsh site as an example of how hard such builds are to deliver. US tech editor Blake Montgomery describes the datacentres as among the most massive and complex structures humanity has created, underlining how the reporting now overlaps with energy, environment and local community concerns.