AI data centers could trigger a far bigger e-waste wave than expected, report warns

A new report from the nonprofit Basel Action Network (BAN) warns that electronic waste generated by the AI industry could grow enough by 2050 to fill 23 million shipping containers — roughly equivalent to circling the globe six times if 40-foot containers were lined up end to end.

The estimate is significantly higher than previous studies because BAN's calculation goes beyond servers and GPUs to include the broader infrastructure that supports data centers: power supply and distribution systems, cooling equipment, backup power, and networking gear. The report also introduces a category called "AI Waste Contagion," which accounts for telecommunications infrastructure and personal devices that may become obsolete and get replaced sooner because of AI-driven advances.

BAN founder Jim Puckett said that while AI may feel weightless, every model depends on a massive amount of specialized hardware, and if companies and governments fail to plan for this coming waste, the current AI buildout could turn into an even more severe toxic waste crisis than the one already underway.

Globally, 68.3 million tons of e-waste are produced each year, but less than a quarter of it is formally collected and recycled. Most ends up in informal waste streams, where burning or burying equipment exposes workers and surrounding communities to toxic substances such as lead and chromium. The US, home to more data centers than any other country, still has not ratified the Basel Convention, the international treaty meant to limit trade in hazardous waste, and American recyclers continue to ship e-waste abroad, where it often ends up in unregulated backyard recycling operations.

BAN projects that total global e-waste will nearly triple by 2050, reaching up to 211 million metric tons annually, with AI responsible for 15 to 20 percent of that total. Factoring in all categories, the group estimates 70,000 metric tons of e-waste per gigawatt of data center capacity. Citing a McKinsey projection of data center capacity reaching up to 219 gigawatts by 2030, BAN calculates that AI-related equipment retired between 2025 and 2050 could add up to between 395 and 617 million metric tons of e-waste, or roughly 8.6 to 13.1 million metric tons per year. Earlier studies, which focused narrowly on servers and accelerators, missed about 87 percent of a data center's electro-mechanical infrastructure, according to BAN.