Who’s afraid of the big, bad GPU?
A feature from The Verge examines the environmental and ethical costs of GPUs, the chips powering the current AI boom, arguing that scrutiny should extend beyond AI data centres to the broader history and use of graphics processors. The piece notes that hundreds of thousands of GPUs are being installed in data centres worldwide, but that the same chips also power smartphones, cars and gaming PCs, raising the question of why AI alone is singled out for its environmental footprint when manufacturing, energy and water use, and e-waste concerns apply across the whole industry.
The article traces GPU origins back to 1970s arcade graphics hardware, predating Nvidia's 1999 "world's first GPU" marketing claim, and cites Staffordshire University ethics professor Catherine Flick on how gaming has long served as a testbed for technologies with significant ethical implications, including AI. It also highlights Nvidia's revenue growth, with data-centre revenue rising from around $0.6 billion in fiscal 2018 to over $62 billion in fiscal 2026, and discusses how the US, home to the largest concentration of data centres, is seeing communities increasingly confront the local impact of new hyperscale facilities being built to support AI expansion.
- GPUs' environmental costs extend well beyond AI data centres alone
- Nvidia's data-centre revenue soared from $0.6bn to $62bn since 2018
- Communities increasingly grapple with new AI data centres locally
New here? Start with this
Nvidia's GPU chips, originally designed to render graphics for video games, have become the essential hardware for training and running artificial intelligence systems, and demand for them has driven a huge expansion of data centres around the world. This has brought growing attention to the environmental costs involved, including the energy and water needed to run these facilities and the manufacturing and disposal of the chips themselves.
The debate explored in this piece is whether it is fair to focus criticism specifically on AI's use of GPUs, given that the same chips are found in everyday devices such as smartphones, cars and gaming computers, all of which carry similar environmental costs. It also looks at the long history of GPU technology, which stretches back decades before Nvidia popularised the term, and how gaming has often been the first place where new computing technologies were tested before being put to wider, more contentious uses.
This matters because AI's rapid growth has made Nvidia one of the most valuable companies in the world and turned data centres into a major issue for communities living near them, particularly in the United States. Understanding where GPUs came from and how widely they are used helps explain why arguments about AI's environmental impact are often bound up with broader questions about the technology industry as a whole.
AI Americas Art Celebrity Culture Elections Entertainment Environment Politics Science Technology World