The rapid expansion of artificial intelligence infrastructure could generate between 395 million and 617 million metric tons of retired electronic equipment between 2025 and 2050, according to a report published by the Basel Action Network (BAN). This projected volume represents triple the current volume of equipment, driven by the massive hardware demands of data centers housing servers, accelerators, racks, cooling systems, and power distribution units.
The Scope of AI-Driven Electronic Waste
According to the Basel Action Network white paper titled “The Coming AI Waste Wave,” the hardware supporting artificial intelligence requires vast electromechanical infrastructure. Servers, accelerators, and racks account for 13% of a data center’s electro-mechanical infrastructure by mass, while power supply and distribution systems comprise 34%. Cooling systems represent 35% of the total infrastructure mass, backup power systems make up 15%, and networking equipment accounts for 3%, according to the report.
For comparison, the UN Global E-Waste Monitor estimated that the world generated 62 million metric tons of material in 2022, with approximately 22.3% of that total properly recycled. Scaled up based on that estimate, the amount generated in 2050 would be about 165 million metric tons — but that’s not including the kind of increased demand AI could create. The new BAN projections suggest that AI-specific retirement cycles could push total annual generation far beyond baseline expectations.
Understanding AI Waste Contagion
The report introduces the concept of “AI waste contagion,” referring to electronics outside data centers that may be retired early as personal devices and telecommunications networks need to be updated to work with AI. According to the Basel Action Network, IT equipment in homes, businesses, and gaming rooms will be affected.
This contagion effect could add between 16.2 million and 30.7 million metric tons of retired equipment annually to the global e-scrap total by 2050. When combined with data center retirements, total annual e-waste generation could reach 211 million metric tons, representing a 28% increase over BAN’s previous estimates and triple today’s annual total.
Mitigation Strategies and Future Policy
Jim Puckett, founder of the Basel Action Network, stated via email that the goal of the white paper series is to slow down the AI boom rather than halt it entirely. BAN’s model’s inputs include how fast capacity is growing, how much equipment each added gigawatt requires and how long that equipment lasts.
Subsequent releases in the four-part white paper series will examine whether it will be possible to reuse, refurbish and repurpose material to reduce the impact. The upcoming publications will also address the toxicity of the stream and propose policy responses BAN suggests to manage the outcome.