GenAI’s silent cost: the looming environmental crisis

The rapid evolution of generative artificial intelligence (GenAI), particularly large language models such as ChatGPT and Gemini, has triggered an unprecedented expansion of computing infrastructure. Yet, behind this rapid technological advancement lies a significant and largely overlooked consequence: the looming crisis of electronic waste (e-waste). According to The Washington Post, tech companies such as OpenAI and Google rely on resource-intensive advanced chips to power GenAI models, and frequent equipment upgrades are accelerating e-waste generation (De Vynck, 2024). By 2030, GenAI alone may generate 1.2 to 5.0 million tons of e-waste globally, roughly equivalent to discarding 13 billion iPhones (Wang et al., 2024). Unlike conventional information technology equipment, GenAI hardware—particularly graphics processing units and high-performance servers—undergoes rapid obsolescence, often requiring replacement every 18 to 24 months. MIT Technology Review reported that the accelerated hardware cycle stems from the relentless pursuit of greater performance and efficiency, leading to mass disposal of functioning but outdated equipment (Crownhart, 2024). The Global E-Waste Monitor 2024 reports that the world produced 62 million tons of e-waste in 2022 alone and is on track to reach 82 million tons by 2030. Yet,