AI’s rapid growth is causing an increase in electronic waste and resource consumption. Data centers are upgrading hardware frequently, leading to millions of tons of discarded components. Large language models (LLMs) like ChatGPT and Claude could generate 2.75 million tons of e-waste annually. The AI industry is also using vast amounts of resources; for example, ChatGPT’s water consumption is significant.
By 2030, e-waste could contain almost a million tons of lead, 6,000 tons of barium, and concerning amounts of other toxic elements. Proper recycling of metals in e-waste could generate $70 billion for the economy and incentivize infrastructure development. Solutions include extending server lifespan, reusing components, and repurposing outdated AI servers for other tasks.
Environmental groups are prioritizing addressing the issue before it becomes unmanageable.