The rapid growth of artificial intelligence (AI) is generating a significant increase in electronic waste (e-waste), with estimates indicating that large language models (LLMs) could produce 2.75 million tons of e-waste annually by 2030. The AI sector's increased complexity necessitates frequent hardware upgrades, leading to discarded functional equipment. A study projects that by 2030, the e-waste generated could rise to about 16 million tons if LLM usage becomes widespread. This growing waste stream is expanding at a rate of 110% annually, far surpassing traditional e-waste growth rates. North America accounts for 58% of AI-related e-waste. The environmental impact includes toxic materials such as lead and cadmium entering ecosystems. Recycling could recover valuable metals worth $70 billion, with suggestions for solutions like extending server lifespans and repurposing hardware to mitigate e-waste. Experts emphasize the urgency of addressing this issue before it escalates further, highlighting the need for both corporate responsibility and improved recycling infrastructure.

Source 🔗