AI competition to generate 617 MT of e-waste by 2050: Basel Action Network (BAN)

The rapid expansion of artificial intelligence (AI) could generate between 395 million and 617 million metric tonnes of electronic waste worldwide between 2025 and 2050, according to a new assessment by environmental organisation Basel Action Network (BAN).

The estimate covers electronic equipment used across AI data centres, including servers, graphics processing units (GPUs), networking equipment, power systems, cooling infrastructure and backup power facilities. BAN study says that it is significantly higher than previous estimates, which largely focused on servers and computing chips.

By 2030, annual waste from AI-related electronic equipment could reach 8.6 million to 13.1 million tonnes, highlighting the growing environmental challenge posed by the global AI boom.

Mounting e-waste beyond imagination

BAN held that the remaining share comprises networking systems, power supply and distribution equipment, cooling infrastructure and backup power facilities.

As AI computing capacity expands, the replacement of this supporting infrastructure could significantly increase the volume of discarded electronics.

The rapid pace of technological advancement could further aggravate the problem. BAN mentions that older equipment may be replaced before the end of its functional life as companies adopt newer, more powerful chips and systems to meet growing computing demands.

Extending equipment lifespans through repair, reuse and refurbishment, alongside the recovery of valuable materials, could help reduce resource consumption and the environmental burden of disposal. Improperly managed e-waste can release hazardous substances, including lead and mercury, posing risks to soil, water and human health, said BAN survey.

Furthering data security challenge

Disposal of AI infrastructure also raises concerns about data security. Retired servers, storage devices and networking equipment may contain confidential business information, system details and login credentials.

Companies need to ensure secure handling, proper data erasure and documented tracking throughout the disposal process.

The BAN findings underline the need for responsible management of AI infrastructure, combining secure equipment disposal with greater reuse, repair and recycling as the technology industry continues to expand.