There has been growing mainstream media coverage on the heavy energy and water demands of artificial intelligence (AI) systems, especially by popular generative models like ChatGPT. Headlines depict AI as an industry greedily guzzling resources, framing the issue as the technology itself having potentially disastrous environmental impacts. However, the history of AI reveals different interpretations. AI emerged conceptually in the 1950s and developed slowly over decades, enabled by expanding computational power predominantly within academic contexts. Private tech companies have since monopolized AI research and development, concentrating the data and compute resources in a few corporations like Meta, Google, and OpenAI. Thus, AI today is largely owned by big tech companies and “data brokers” and is more a marketing term than a coherent field.

Data companies have powered AI’s recent advances through massive data centers straining energy grids worldwide. While data centers are increasingly “greening,” AI’s scope grows exponentially, demanding more infrastructure, land, water, minerals, human labour, and electricity. When discussing solutions, discourse typically weighs AI’s potential benefits against sustainability costs. Lost in this debate are AI’s links to extractive logics serving tech-capitalism’s endless growth mandate. AI is thereby tethered to the same environmentally damaging industries and practices fueling the climate crisis. Beyond framing AI as a resource-hungry industry itself, analyses should trace how AI necessitates and perpetuates extractivism throughout its material lifecycle.