AI e-waste is emerging as one of the most serious environmental challenges of the next decade, with a new report revealing dimensions that exceed any previous estimate. The rapid development of artificial intelligence and the data centers that support it is creating a cycle of e-waste production that threatens to flood the planet with toxic materials. According to the findings, existing estimates of e-waste have not taken into account the full scope of the infrastructure required to support AI.
See also: Meta Hyperion: Investment of over $50 billion for the largest data center
The new report highlights the urgent need for companies and governments to address the looming waste crisis linked to the rapid development of artificial intelligence. It is an issue that affects not only the environment but also the public health of millions of people worldwide, particularly in developing countries where most of the world's e-waste ends up.
The global e-waste problem was already huge before the AI boom. 68.3 million tons of e-waste are generated worldwide each year, and less than 25% of it is properly collected and recycled. The remaining 75% ends up in informal waste collection systems, where hazardous materials — lead, mercury, cadmium, brominated flame retardants — pose serious risks to workers’ health and the environment.
AI e-waste and the role of the Basel Action Network
The Basel Action Network (BAN), a non-profit organization that monitors the illegal trade in toxic waste, is the one who wrote the extensive report on AI e-waste. The organization points out that the United States, the country with the largest number of data centers in the world, has not ratified the Basel Convention — the international treaty that regulates the cross-border movement of hazardous waste. This allows the continued export of e-waste to countries where it often enters informal recycling processes.
This situation endangers the health of millions of people, especially children, who are exposed to toxic substances. In countries in Africa, Asia and Latin America, informal workers — often minors — dismantle electronic waste with their bare hands, inhaling toxic fumes from burning cables to recover copper. The health consequences are devastating: neurological damage, cancer, respiratory problems and developmental disorders in children.
BAN also expands the definition of AI e-waste beyond the obvious servers and GPUs . The report includes power supplies, cooling systems, networking devices, and the entire supporting infrastructure required to run a modern AI data center . This broader approach dramatically changes the calculations for the overall waste footprint of the AI industry.
AI e-waste: The worrying numbers
BAN ’s predictions for the future of AI e-waste are impressive. The organization estimates that global e-waste will grow to 211 million metric tons per year by 2050 , with AI contributing 15 to 20 percent of that total. To put that into perspective: that equates to tens of millions of tons of hazardous materials the planet will have to manage every year.
The report also introduces a new concept: “AI Waste Contagion .” This concept describes how advances in artificial intelligence are impacting and accelerating equipment retirement in telecommunications and personal devices. As new AI require more powerful hardware, older devices — from smartphones to routers — are being replaced more quickly, creating a cascade of waste across the entire technology ecosystem.
See also: Intel and AMD "lock" multi-year agreements with Chinese data centers
BAN estimates that data centers will generate approximately 70,000 metric tons of e-waste for every gigawatt of capacity. With projections showing that data center could reach 219 gigawatts by 2030, the expected e-waste from equipment retired between 2025 and 2050 could total 395 to 617 million metric tons. This translates to an annual retirement of 8.6 to 13.1 million metric tons of equipment due to the advancement of artificial intelligence.
The impact of AI e-waste in Europe and Greece
Europe is better positioned than the US when it comes to e-waste legislation, as the European Union has ratified the Basel Convention and is implementing strict regulations through the WEEE (Waste Electrical and Electronic Equipment) Directive. However, the rapid growth of data centers in Europe — with countries such as Ireland, the Netherlands and Germany hosting huge facilities — means that Europe will also face significant AI e-waste in the coming years.
For Greece, the challenge is twofold. On the one hand, the country must manage its own e-waste in accordance with European directives. On the other hand, the growing demand for data centers in the Mediterranean — with Greece emerging as an attractive destination due to its climate and geographical location — creates new responsibilities. Proper management of AI e-waste is not only an environmental obligation, but also an economic opportunity for job creation in the circular economy.
Solutions to the AI e-waste problem
Addressing the AI e-waste requires action on multiple levels. Tech companies — from Google and Microsoft to Meta and Amazon — must adopt circular economy policies that include repairing, reusing, and responsibly recycling their equipment. Some companies have already taken steps in this direction, but efforts remain insufficient given the scale of the problem.
Governments, for their part, need to enact stricter regulations for the management of e-waste from data centers, invest in recycling infrastructure, and ensure that companies are held accountable for the lifecycle of their equipment. Ratification of the Basel Convention by the U.S. would be a critical first step in reducing the illegal export of e-waste to developing countries.
On a technical level, the AI industry needs to invest in more efficient chips and systems that require less power and have a longer lifespan. Developing GPUs and processors that can be upgraded rather than completely replaced, as well as using recycled materials in the construction of new equipment, are critical strategies for reducing the AI e-waste footprint .
See also: AWS data centers: Drone attacks caused extensive damage
The problem of AI e-waste is not just an environmental issue — it is a question of social justice, public health, and sustainable development. Artificial intelligence holds the promise of solving many of humanity’s problems, but if its development creates new waste crises that affect the planet’s most vulnerable populations, then the cost of progress is too high. The time to act is now — before the numbers get out of control.
