Generative AI Boom Could Generate Millions of Tons of E-Waste by 2030, Study Warns

Generative AI could generate millions of tons of e-waste by decade's end, study finds

A team of urban environmentalists at the Chinese Academy of Sciences' Institute of Urban Environment, working with a colleague from Reichman University in Israel, has attempted to estimate the amount of e-waste that will be generated over the next several years due to the implementation of generative AI applications. In their study, published in Nature Computational Science, the group attempted to add up all the circuit boards, batteries and other pieces of electronic hardware used to drive generative AI applications as they outlive their usefulness. As generative AI applications like ChatGPT have taken the world by storm, one overlooked aspect of their rise is the hardware used to run them. Such applications are typically run on specialized GPUs plugged into specialized computers. They are typically housed together in data centers and server farms and there are a lot of them. Generative AI apps are resource and energy intensive, and because they have become critical for some users, large stores of batteries ensure operation in the event of outages. Unfortunately, all such equipment has a shelf life. As it ages or becomes obsolete, it is replaced. The old hardware then becomes e-waste. In this new effort, the research team attempted to estimate the total amount of such e-waste that will be generated between now and the end of this decade. To make their estimates, the research team estimated the amount of hardware typically used to run a given application at a standard data center/server farm and the average shelf life for each of its components. They then identified the number of such data centers. They made educated guesses about the expected demand for such applications and their services in the years ahead. Finally, they integrated all their data into a computer model programmed to make such types of estimates. The model showed that if things remain on their current trajectory, the AI industry could produce somewhere between 1.2 to 5.0 million metric tons of e-waste by the end of the decade. It also showed annual production of e-waste increasing from 2.6 thousand metric tons in 2023 and potentially reaching up to 2.5 million metric tons per year by the end of the decade. The researchers note that such huge amounts of waste production could be avoided if the industry adopts a circular economy approach in which hardware is recycled.

Generative AI Has a Massive E-Waste Problem

Katherine Bourzac is a freelance journalist based in San Francisco, Calif. Private investment in generative AI has grown from about US $3 billion in 2022 to $25 billion in 2023, and about 80 percent of private companies expect AI to drive their business in the next 3 years, according to Deloitte. Keeping up with the latest advancements means upgrading GPUs, CPUs, and other electronic equipment in data centers as newer, more advanced chips become available. And that, researchers project, will lead to an explosion in the production of electronic waste. A study published last week in the journal Nature Computational Science estimates that aggressive adoption of large language models (LLMs) alone will generate 2.5 million tonnes of e-waste per year by 2030. "AI doesn't exist in a vacuum; it relies on substantial hardware resources that have tangible environmental footprints," says study coauthor Asaf Tzachor, a sustainability and climate researcher at Reichman University, in Israel. "Awareness of the e-waste issue is crucial for developing strategies that mitigate negative environmental impacts while allowing us to reap the benefits of AI advancements," he says. Most research on AI sustainability has focused on these models' energy and water use and their concomitant carbon emissions. Tzachor worked with Peng Wang and Wei-Qiang Chen, both professors at the Chinese Academy of Sciences, to calculate the potential increase in e-waste associated with generative AI. The study is intended to provide an estimate of the potential scale of the problem, and the researchers hope it will spur companies to adopt more sustainable practices. Electronic waste contains toxic metals and other chemicals that can leach out into the environment and cause health problems. In 2022, the world produced 62 million tonnes of e-waste in total, according to the United Nations Global E-waste Monitor. This waste stream is growing five times faster than recycling programs, the UN found. In the coming years, AI could make a significant contribution to the problem. Tzachor says e-waste associated with generative AI includes discarded GPUs, CPUs, batteries used for backup power in data centers, memory modules, and printed circuit boards. The study details four potential scenarios for generative AI adoption -- ranging from limited to aggressive expansion -- and projects potential e-waste expansion from a 2023 baseline of 2,600 tons per year. Limited expansion of AI use would generate a total of 1.2 million tonnes of e-waste from 2023 to 2030; aggressive use would result in a total of 5 million tonnes over that period. Tzachor says given current trends, the aggressive scenario is most likely. The study isn't comprehensive -- it only considers large language models, not other forms of generative AI. Tzachor says the team focused on LLMs because they're among the most computationally intensive. "Including other forms of AI would increase the projected e-waste figures," Tzachor says. In theory, adopting more advanced chips should help server farms do more with less, and produce less waste. But each upgrade results in a net increase in the waste stream. And given current trade restrictions on semiconductors, upgrading is not always an option. Countries that don't have access to the most advanced chips may generate more waste as a result. A one-year delay in upgrading to the latest chips will result in a 14 percent increase in e-waste, according to the study. One of the best ways to mitigate this AI waste stream is to find ways to reuse electronic equipment -- what Tzachor calls downcycling. Servers that are no longer cutting edge can be repurposed for hosting websites or doing more basic data processing tasks, or they can be donated to educational institutions. Most tech companies -- including

Researchers warn AI boom could generate up to 5 million tons of e-waste

A new peer-reviewed study published in the journal Nature Computational Science warns that the rapid growth of generative artificial intelligence (AI) could generate up to 5 million tons of electronic waste (e-waste) annually by 2030. The international collaboration was led by Peng Wang from the Chinese Academy of Sciences and included contributions from scientists in Israel. Researchers estimate that if immediate measures are not implemented, AI could generate between 1.2 and 5 million metric tons of e-waste by the end of this decade. This potential increase is equivalent to discarding between 2.1 and 13 billion units of the iPhone 15 Pro or more than 11,000 loaded Boeing 747 airplanes. The study emphasizes that this surge in e-waste could worsen the global toxic trash crisis, including 1.5 million tons of printed circuit boards and 0.5 million tons of hazardous batteries. The rise in electronic waste is attributed to the rapid expansion of AI applications and data centers, which demand frequent upgrades of high-performance computing hardware. Generative AI models, such as large language models used in applications like ChatGPT, are highly resource-intensive, requiring powerful servers, processors, and storage solutions to operate effectively. This dependence on rapid improvements in hardware infrastructure and chip technology results in short life cycles for advanced processors and storage equipment, leading to a surge of discarded electronics. Leading tech companies are spending heavily to build and upgrade data centers to power generative AI projects and to stock them with powerful computer chips. As the AI boom continues, the older chips and equipment could amount to extra electronic waste equivalent to throwing out 13 billion iPhones annually by 2030, according to a study from academics in China and Israel. Most of the e-waste would be clustered in regions like North America, Europe, and East Asia, where most data centers are concentrated. E-waste includes discarded electronic devices such as computers, smartphones, chargers, wires, and larger server systems, defined as any product with a battery or plug. It is the planet's fastest-growing waste stream and is rapidly outstripping the capacity of recycling facilities. The vast majority of electronic waste is not recycled, with much of it ending up in landfills or exported to lower-income countries. In these countries, people manually break apart old devices to access copper and other metals, exposing workers to harmful substances such as mercury and lead. This improper disposal leads to the release of hazardous materials, harming ecosystems and human health. Researchers have offered solutions to reduce the electronic waste caused by AI. Implementing certain practices, including circular economy strategies, could reduce AI-related e-waste by up to 86%. By extending the lifespan of existing computer infrastructure, reusing parts, and recycling valuable materials like copper and gold, e-waste generation could be significantly reduced. The authors suggest that applying a circular economy strategy could prevent the generation of more than three million tons of waste. Asaf Tzachor, an associate professor at Reichman University and one of the study's authors, said, "We hope this work brings attention to the often-overlooked environmental impact of AI hardware," according to The Washington Post. He added, "AI comes with tangible environmental costs beyond energy consumption and carbon emissions." However, challenges exist in implementing these strategies. Tzachor stated, "Challenges like data security concerns and the need for high-performance hardware can make reuse and recycling more complex." Some experts, like Ana Valdivia from the University of Oxford, point out that options like reusing GPUs are not always viable. Valdivia told El Comercio Perú, "GPUs cannot be inserted into a circular economy because it is very expensive to recycle their components. 100% of a GPU ends up incinerated or in a landfill." Geopolitical restrictions on semiconductor imports may intensify the increase in e-waste generation from AI. U.S. restrictions on the sale of advanced GPUs to countries such as China force data centers to use outdated server models, resulting in more e-waste. Stay updated with the latest news! Subscribe to The Jerusalem Post Newsletter Subscribe Now The study underscores the need for responsible use of generative AI and proactive strategies for managing electronic waste to reduce the harmful effects of pollution. Peng Wang expressed deep concern regarding the competition between the expansion rate of generative AI and the adoption of the circular economy. He said, "Given the unprecedented increase in demand for this technology, to win this battle, shock measures should be implemented imminently," as reported by El Comercio Perú. Electronic waste is already a significant and growing global problem. According to an annual United Nations report, a record 62 million tonnes of e-waste was produced in 2022. Projections from the United Nations Institute for Training and Research foresee that e-waste could soar to 82 million tons by 2030. In light of this alarming projection, the authors of the study suggest several strategies to limit electronic waste generation. The researchers write, "Relatively simple strategies could have a major effect on e-waste generation." Dismantling, renovating, and reassembling obsolete modules, such as GPUs, for less intensive forms of computing could reduce e-waste by 42%. Extending the lifespan of these devices by just one year could prevent the generation of more than three million tons of waste. Shaolei Ren, Associate Professor of Electrical Engineering and Computer Science at the University of California, Riverside, said, "Electronic waste is a critical issue, although often overlooked, when considering the future social impact of generative AI. This article draws attention to the problem of electronic waste generated by generative AI, and I believe it will invite a deeper debate," according to Página/12. The study highlights that if things remain on their current trajectory, the AI industry could produce somewhere between 1.2 to 5.0 million metric tons of e-waste by the end of the decade. This increase would be as much as a thousandfold over 2023 levels, which is equivalent to "throwing away between 2.1 and 13 billion units of the iPhone 15 Pro." As generative AI continues to expand, the environmental footprint of this technological advancement becomes a pressing concern. The study emphasizes the need for immediate measures to reduce the environmental impact of electronic waste generation, warning that without action, the consequences could be devastating. Sources: Página/12, ABC Digital, Home, DIE WELT, El Comercio Perú, Yahoo News, Australian Broadcasting Corporation, The Washington Post This article was written in collaboration with generative AI company Alchemiq

AI boom may spur 5 million tons of e-waste by 2030: Study warns

A recent study conducted by environmentalists at the Chinese Academy of Science in collaboration with Reichman University in Israel raises serious concerns about the environmental impact of generative artificial intelligence (AI). The demand for specialized hardware required to power apps like ChatGPT is skyrocketing, making older electronic devices outdated, according to the study. "Our findings indicate that this e-waste stream could increase, potentially reaching a total accumulation of 1.2 to 5 million tonnes during 2020 to 2030," said the researchers, South China Morning Post reported.

Generative AI could cause 10 billion iPhones worth of e-waste per year by 2030

The immense and quickly advancing computing requirements of AI models could lead to the industry discarding the e-waste equivalent of over 10 billion iPhones per year by 2030, researchers project. In a paper published in the journal Nature, researchers from Cambridge University and the Chinese Academy of Sciences take a shot at predicting just how much e-waste this growing industry could produce. Their aim is not to limit adoption of the technology, which they emphasize at the outset is promising and likely inevitable, but to better prepare the world for the tangible results of its rapid expansion. Energy costs, they explain, have been looked at closely, as they are already in play. However, the physical materials involved in their life cycle, and the waste stream of obsolete electronic equipment ... have received less attention. Our study aims not to precisely forecast the quantity of AI servers and their associated e-waste, but rather to provide initial gross estimates that highlight the potential scales of the forthcoming challenge, and to explore potential circular economy solutions. It's necessarily a hand-wavy business, projecting the secondary consequences of a notoriously fast-moving and unpredictable industry. But someone has to at least try, right? The point is not to get it right within a percentage, but within an order of magnitude. Are we talking about tens of thousands of tons of e-waste, hundreds of thousands, or millions? According to the researchers, it's probably towards the high end of that range. The researchers modeled a few scenarios of low, medium, and high growth, along with what kinds of computing resources would be needed to support those, and how long they would last. Their basic finding is that waste would increase by as much as a thousandfold over 2023: "Our results indicate potential for rapid growth of e-waste from 2.6 thousand tons (kt) [per year] in 2023 to around 0.4-2.5 million tons (Mt) [per year] in 2030," they write. Now admittedly, using 2023 as a starting metric is maybe a little misleading: Because so much of the computing infrastructure was deployed over the last two years, the 2.6 kiloton figure doesn't include them as waste. That lowers the starting figure considerably. But in another sense, the metric is quite real and accurate: These are, after all, the approximate e-waste amounts before and after the generative AI boom. We will see a sharp uptick in the waste figures when this first large infrastructure reaches end of life over the next couple years. There are various ways this could be mitigated, which the researchers outline (again, only in broad strokes). For instance, servers at the end of their lifespan could be downcycled rather than thrown away, and components like communications and power could be repurposed as well. Software and efficiency could also be improved, extending the effective life of a given chip generation or GPU type. Interestingly, they favor updating to the latest chips as soon as possible, because otherwise a company may have to, say, buy two slower GPUs to do the job of one high-end one -- doubling (and perhaps accelerating) the resultant waste. These mitigations could reduce the waste load anywhere from 16 to 86% -- obviously quite a range. But it's not so much a question of uncertainty on effectiveness as uncertainty on whether these measures will be adopted and how much. If every H100 gets a second life in a low-cost inference server at a university somewhere, that spreads out the reckoning a lot; if only one in 10 gets that treatment, not so much. That means that achieving the low end of the waste versus the high one is, in their estimation, a choice -- not an inevitability. You can read the full study here.

AI will add to the e-waste problem. Here's what we can do about it.

That's a relatively small fraction of the current global total of over 60 million metric tons of e-waste each year. However, it's still a significant part of a growing problem, experts warn. E-waste is the term to describe things like air conditioners, televisions, and personal electronic devices such as cell phones and laptops when they are thrown away. These devices often contain hazardous or toxic materials that can harm human health or the environment if they're not disposed of properly. Besides those potential harms, when appliances like washing machines and high-performance computers wind up in the trash, the valuable metals inside the devices are also wasted -- taken out of the supply chain instead of being recycled. Depending on the adoption rate of generative AI, the technology could add 1.2 million to 5 million metric tons of e-waste in total by 2030, according to the study, published today in Nature Computational Science. "This increase would exacerbate the existing e-waste problem," says Asaf Tzachor, a researcher at Reichman University in Israel and a co-author of the study, via email. The study is novel in its attempts to quantify the effects of AI on e-waste, says Kees Baldé, a senior scientific specialist at the United Nations Institute for Training and Research and an author of the latest Global E-Waste Monitor, an annual report. The primary contributor to e-waste from generative AI is high-performance computing hardware that's used in data centers and server farms, including servers, GPUs, CPUs, memory modules, and storage devices. That equipment, like other e-waste, contains valuable metals like copper, gold, silver, aluminum, and rare earth elements, as well as hazardous materials such as lead, mercury, and chromium, Tzachor says. One reason that AI companies generate so much waste is how quickly hardware technology is advancing. Computing devices typically have lifespans of two to five years, and they're replaced frequently with the most up-to-date versions. While the e-waste problem goes far beyond AI, the rapidly growing technology represents an opportunity to take stock of how we deal with e-waste and lay the groundwork to address it. The good news is that there are strategies that can help reduce expected waste. Expanding the lifespan of technologies by using equipment for longer is one of the most significant ways to cut down on e-waste, Tzachor says. Refurbishing and reusing components can also play a significant role, as can designing hardware in ways that makes it easier to recycle and upgrade. Implementing these strategies could reduce e-waste generation by up to 86% in a best-case scenario, the study projected.

Power-sucking GenAI also set to create mountain of e-waste

New modelling shows tech could massively increase current volumes of electronic landfill by 2030 Computational boffins' research claims GenAI is set to create nearly 1,000 times more e-waste than exists currently by 2030, unless the tech industry employs mitigating strategies. The study, which looks at the rate AI servers are being introduced to datacenters, claims that a realistic scenario indicates potential for rapid growth of e-waste from 2.6 kilotons each year in 2023 to between 400 kilotons and 2.5 million tons each year in 2030, when no waste reduction measures are considered. The team writes: The multi-national research team led by Peng Wang, professor in material circularity at the Chinese Academy of Sciences, considered four scenarios with varying degrees of generative AI production and application, ranging from an aggressive scenario with widespread applications to a conservative scenario with more specific applications. Under the scenario with the most AI growth the world could create as much as 2.5 million tons of e-waste each year. "Our study aims not to precisely forecast the quantity of AI servers and their associated e-waste, but rather to provide initial gross estimates that highlight the potential scales of the forthcoming challenge and to explore potential circular economy solutions," the researchers say in a paper published in Nature Computational Science today. The research analysis focuses on AI servers that include GPUs, CPUs, storage, memory units, internet communication modules and power systems. Ancillary machinery such as cooling and communication units was excluded from this study. The researchers point out that the weight of Nvidia's latest Blackwell platform in a rack system -- designed for intensive LLM inference, training and data processing tasks -- tips the scales at 1.36 tons, demonstrating how material-intensive GenAI can be. Other predictions suggest AI's installed computational capacity could increase approximately 500-fold from 2020 to 2030. Meanwhile, e-waste resulting from the introduction of GenAI could increase because of geopolitical restrictions on semiconductor imports. Not all is lost, though. The study shows that if the tech industry introduces circular economy strategies along the GenAI value chain, it could result in reducing e-waste generation by between 16 and 86 percent. "This underscores the importance of proactive e-waste management in the face of advancing GAI technologies," the researchers said. ®

The AI boom may unleash a global surge in electronic waste

Most e-waste is never recycled. Data center upgrades for AI projects could significantly add to the problem, researchers say. SAN FRANCISCO -- The Silicon Valley arms race to build more powerful artificial intelligence programs could lead to a massive increase in electronic waste, research published Monday warns. Leading tech companies are spending heavily to build and upgrade data centers to power generative AI projects and to stock them with powerful computer chips. If the AI boom continues, the older chips and equipment could amount to extra electronic waste equivalent to throwing out 13 billion iPhones annually by 2030, the study from academics in China and Israel said.