AI Boom Sparks E-Waste Crisis: Millions of Tonnes Predicted

A new report highlights that the accelerating AI boom is set to generate a significant new stream of electronic waste. Between 2025 and 2050, retired AI equipment alone is projected to contribute between 395 and 617 million tonnes of waste. This staggering amount could fill 15 to 23 million shipping containers, enough to encircle the Earth six times if placed end-to-end.

The increased waste is attributed to rapid technological advancements driving the need for specialized hardware like AI accelerators alongside GPUs. Global data center capital expenditure is expected to reach $7 trillion by 2030, with $4.3 trillion dedicated to electronic hardware. A critical point is that current estimates often overlook 87% of data center hardware, including networking, power, storage, and cooling systems.

Furthermore, AI accelerators and servers are predicted to be replaced every 2.5 years, significantly faster than the typical five to seven-year cycle for general-purpose servers. With 100GW of additional data center capacity anticipated globally by the end of this decade, the scale of potential e-waste is substantial. To mitigate this, recommendations include designing more durable AI data center equipment and exploring recycling and second-use options.