A.I.'S Environmental Impact Will Threaten Its Own Supply Chain (www.nytimes.com)

Researcher Kate Crawford visited Spruce Pine, N.C. — a small mining town whose extracted materials underpin modern electronics — and found the local landscape and community hammered by Hurricane Helene, underscoring how climate shocks intersect with the environmental costs of A.I. Crawford’s reporting and accompanying Times Opinion video argue that the raw materials, land, water and power that make machine learning hardware and data centers possible are being depleted and damaged by extractive practices, leaving residents exposed and the supply chain fragile. As she puts it, generative A.I. increasingly “competes with you for your power, water and land.” The piece signals a structural risk for the A.I./ML field: model development and deployment aren’t just computational problems but materially intensive ones. Concentrated mining and heavy resource use for chip manufacture, cooling and energy-intensive training mean that extreme weather, local environmental degradation and social impacts can bottleneck hardware supply and operational resilience. For practitioners and policymakers this implies urgent priorities — transparent supply chains, energy- and water-efficient model design, recycling and circular sourcing, and regulation to mitigate local harms — if the industry hopes to scale safely without undermining the very ecosystems and communities it depends on.