AI datacenter buildout to trigger 'tsunami' of toxic e-waste, report warns

The report projects that the electronic trash AI generates in that period could fill enough shipping containers to encircle the globe six times, and that AI-driven equipment retirement alone will reach up to 13 million metric tonnes a year by 2030 — far above the estimates of earlier studies. That waste typically contains lead, mercury, cadmium and forever chemicals, along with valuable metals that are difficult to reclaim.

"Already electronic waste is the fastest-growing waste stream on the planet and only one-fifth is properly managed," said Jim Puckett, co-founder of BAN, which campaigns against the dumping of toxic waste on poorer communities. The report says no plan exists to deal with the material the AI buildout will generate.

To reach its figures, BAN counted 4,871 new datacenters in the construction pipeline in the United States, adding that while not all are clearly designated for AI use, many will be. It estimated the computing power analysts expect to be built — roughly 100 gigawatts of new capacity worldwide by 2030 alone, and growing beyond that — then weighed the equipment required, including server racks of 1,360kg each, switches of up to 30kg and hundreds of kilograms of copper cable connecting each rack. Cooling, power distribution, networking and energy storage equipment add thousands of tonnes of copper, steel and concrete.

Such equipment is not expected to be designed for repair or reuse and would need to be completely replaced every two to five years, the report projects. BAN also assumed that AI's computing requirements will force the rapid replacement of consumer computers, phones and telecommunications gear.

Vince Beiser, author of Power Metals, a book on the global difficulties of reclaiming valuable metals from electronic goods, said the problem is already poorly handled. "We're doing a terrible job of handling e-waste in the US," he said. "Some of it gets dumped overseas. Most of it just gets dumped in landfills, which is a problem because it can leach all kinds of nasty things into the water, into the soil."

Datacenter construction has drawn bipartisan opposition across the US over the space and energy it requires. Last week a group of environmental and health experts called on the federal government to pledge to protect citizens' health from the repercussions of the AI buildout, citing estimates that air pollution from the electricity generation the datacenters need could result in hundreds of thousands of new asthma cases and as many as 1,300 premature deaths a year in the US alone. The Trump administration has sought to weaken environmental safeguards that would normally govern a buildout on that scale.

A 2025 report by McKinsey & Company estimated that achieving the tech industry's planned rollout of artificial intelligence will require nearly $7tn in capital outlays worldwide over five years to build datacenters running millions of computer servers around the clock. BAN's white paper notes that this spending "is roughly the cost of ending world hunger for the next 75 years — being spent in just five". "We all find ourselves unwitting participants in the AI horserace," the report says, "as IT users everywhere find themselves using AI whether they would like to or not, and yet almost always, without being aware of the greater consequences of doing so."

Editor's Summary

BAN's white paper projects that the AI datacenter buildout will generate up to 13 million metric tonnes of e-waste annually by 2030 and could triple global e-waste generation within 25 years, with lead, mercury, cadmium and forever chemicals among the materials lacking a disposal plan. The findings arrive as datacenter construction faces bipartisan local opposition in the US and as the Trump administration moves to weaken environmental safeguards. The report also contrasts the nearly $7tn in planned global datacenter spending over five years against what it says is the cost of ending world hunger for 75 years.