AI Boom Could Fuel 617 Million-Tonne E-Waste Wave
Artificial intelligence may appear weightless, but its expansion relies on servers, GPUs, power-distribution systems, cooling equipment, backup power and networking gear. The nonprofit Basel Action Network said earlier forecasts focused largely on computing hardware and omitted about 87% of a data center’s electromechanical infrastructure. Counting the full equipment stack sharply increases the estimated environmental burden from rapid replacement cycles.
BAN said in a report released on Sept. 15, 2026, that AI-related equipment retired between 2025 and 2050 could total 395 million to 617 million metric tonnes, enough at the upper end to fill 23 million 40-foot shipping containers. Global e-waste could reach 211 million tonnes annually by 2050, with AI accounting for 15% to 20% of the total.
All Coverage
2 original reportsThe Backstory
The history behind this eventBig Tech Data Centre Boom Threatens Carbon Surge
Surging demand for AI training and inference is pushing technology groups to build vast data centres packed with power-hungry chips and cooling systems. The expansion is becoming a major test of corporate climate pledges and global emissions targets, particularly where electricity grids remain dependent on fossil fuels and additions of renewable generation and transmission capacity are failing to keep pace.
A Financial Times analysis found that 60 large data centres planned by Big Tech could generate annual carbon emissions equivalent to those of 27 coal-fired power plants or about 24 million cars. The estimates highlight the potential environmental cost of the AI infrastructure race, as companies secure unprecedented amounts of electricity and computing capacity while clean-energy supplies and grid upgrades lag behind projected demand.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →