Was this email forwarded to you? Sign up here

🗞️ Issue 35 // ⏱️ Read Time: 7 min

Hello 👋

As we conclude our three-part series on AI's environmental impact, we turn our attention to a less visible but equally critical issue: the growing mountain of e-waste generated by the industry’s growing appetite for the latest AI hardware. Just as plastic revolutionized industries but left a lasting environmental footprint, AI hardware is reshaping our world while potentially creating a "digital landfill" for future generations.

The Digital Landfill

The Issue: Hardware waste

While AI continues to push the boundaries of what's possible in technology, it's leaving behind a less celebrated legacy: mountains of obsolete hardware. We’ve previously written about the role of GPU’s in the development of AI technologies. According to the 2024 UN Global E-Waste Monitor, 62 million tonnes of e-waste was produced in 2022, an 82% increase from 2010. Even more concerning, this figure is projected to reach 82 million tonnes (that’s the equivalent to about 546,667 blue whales! 🐋) by 2030.

👉🏼 Due to the presence of lead and other toxic materials, e-waste currently makes up around 70% of the world’s surface-level toxic pollution. Exposure to e-waste toxins is linked to various health issues in humans, including:

• Damage to the nervous system, blood, kidneys, brain, and other organs 🧠

• Respiratory problems 🫁

• Digestive issues 🍥

• Bone problems 🦴

• Developmental issues in children 🍼

👉🏼 A 2020 survey by Statista found that servers are replaced frequently, with 42% of data center managers and IT staff reporting that they refreshed their data center servers every two to three years, and 26% stated that they did so every year.

👉🏼 A study by the watchdog group Basel Action Network found that 40% of the e-waste supposedly recycled in the U.S. was actually exported. Most of it ended up in developing countries – usually in Asia and Africa – where recycling is typically unlicensed and unregulated.

So Where Does All The Waste Come From?

Older hardware quickly becomes obsolete as more powerful AI chips and high-performance GPUs are developed. This constant cycle of innovation and replacement, driven by the increasing demand for AI capabilities, results in a continuous stream of discarded hardware.

Data centres, the heart of AI operations, contribute significantly to the estimated 50 million tons of electronic and electrical waste produced worldwide each year. Only 20% of this e-waste is recycled, leaving a vast amount of potentially harmful materials in landfills or informal dumping grounds.

The reasons as to why 80% of the e-waste is not being recycled at the moment, depends on several different factors

Lack of infrastructure: Many regions lack proper e-waste collection and recycling facilities.

Economic incentives: Proper e-waste recycling can be expensive, and the economic incentives are often insufficient.

Product Design: Many electronic devices are not designed with easy recycling in mind, making the process more difficult and costly. Phones, for example, contain around 70 small pieces. Alternatives where this has been taken into consideration when designing the product can have as little as 8 components.

Valuable materials: While e-waste contains precious metals, the extraction process can be complex and not always economically viable.

Consumer Behaviour: Many consumers are unaware of proper e-waste disposal methods or the importance of recycling electronics. Do you know how to recycle your phone the next time you get a new one?

Stockpiling: People often keep old devices at home instead of recycling them, contributing to the low recycling rates. This is your sign to get rid of that laptop you were using in 2005.

Regulatory Challenges: Insufficient legislation and illegal export of e-waste to countries with weaker environmental regulations.

According to a paper in the Circular Economy and Sustainability journal, 23 of the world's 30 critical raw materials are found in server, storage and networking equipment. Meaning the resources we rely on for the world’s data centres are in low global supply and have no viable substitutes. This underscores the urgent need for data centres to minimize the mining of these materials by reusing existing resources - that is, to recycle the already existing hardware instead of throwing it to landfill.

How Major Contributors are Managing Their E-waste

Below we’re listing some of the initiatives of the biggest tech companies to deal with e-waste. While these initiatives appear to align with many industry best practices, it's important to note that without independent assessments or more detailed metrics, it's challenging to fully evaluate their effectiveness. Transparency in reporting recycling rates, volumes processed, and environmental impact reductions would provide a clearer picture of how these programs compare to broader industry standards.

Addressing the Challenge

As we grapple with the growing e-waste crisis, several strategies emerge as potential solutions:

1. Design for Longevity: Developing hardware with longer lifespans and easier upgradeability can significantly reduce waste and emissions. A 2019 study found that extending the lifetime of the EU’s stock of smartphones and other electronics by five years would save almost 10 million tonnes of emissions (CO2eq) annually by 2030. This is equivalent to taking 5 million cars off the roads for a year.

2. Robust Recycling Programs: Implementing comprehensive recycling initiatives for AI-related hardware is crucial. Apple's robot Daisy, capable of disassembling 200 iPhones per hour, serves as an example of innovative recycling technology.

3. Cloud-Based Solutions: Encouraging the use of cloud-based AI services can reduce the need for individual hardware, potentially decreasing overall e-waste.

4. Regulatory Frameworks: Implementing stringent regulations on AI's resource consumption and hardware disposal can drive industry-wide change.

5. Research Investment: Prioritizing research into more environmentally friendly AI technologies is essential for long-term sustainability.

6. Entrepreneurship: If you are someone that thinks problems are made to be solved - here’s your chance. With such a big amount of scarce, valuable resources being wasted, there are plenty of ways to solve the problem and make some money from it, too. One person’s trash, another person’s treasure, as they say.

It's important to note that addressing these challenges isn't about halting progress. Rather, it's about ensuring that our technological advancements don't come at the cost of irreversible environmental damage. By tackling the e-waste issue head-on, we can pave the way for an AI future that's both innovative, sustainable and relevant.

As we conclude this series on AI's environmental impact, it's clear that the path forward requires for innovation to integrate principles of sustainability. By understanding the impact of AI, we can build better and make sure that AI is actually useful, and solves more problems than it creates.

Until next time.
On behalf of Team Lumiera

Emma - Business Strategist
Allegra - Data Specialist

Lumiera has gathered the brightest people from the technology and policy sectors to give you top-quality advice so you can navigate the new AI Era.

Follow the carefully curated Lumiera podcast playlist to stay informed and challenged on all things AI.