Over the past year, my research has first highlighted the rapidly growing power demand of AI systems, followed by an assessment of the associated carbon and water footprints.
My latest research addresses another, often overlooked consequence of the expanding AI infrastructure: electronic waste.
• Substantial AI e-waste persists, underscoring the need for data center transparency.
The full article is available open access and can be accessed here: doi.org/10.1016/j.re...
• AI systems may contribute less to global e-waste than previously anticipated.
• The gap highlights the need for supply-chain data and realistic AI server lifespans.
• 2030 AI e-waste could still match Denmark, Norway, or Austria’s 2022 e-waste.
While this is not the first attempt to examine AI-related e-waste, it is the first estimate grounded in supply-side data, rather than inferred from demand-side assumptions.
The main findings are as follows:
• By 2030, AI servers could generate 131.0–224.8 kilotons of e-waste per year.
Over the past year, my research has first highlighted the rapidly growing power demand of AI systems, followed by an assessment of the associated carbon and water footprints.
My latest research addresses another, often overlooked consequence of the expanding AI infrastructure: electronic waste.
Let’s start the year by reflecting on the colossal growth in the resource consumption of AI and what it means for 2026.
At the start of 2025, the worldwide power demand of AI systems was approximately 9.4 GW. By the end of the year, this demand had likely increased to around 23 GW.
Links:
Power demand of AI systems: doi.org/10.1016/j.jo...
Carbon and water footprint of AI: doi.org/10.1016/j.pa...
The balance between private gains and public costs is profoundly skewed, and the industry has little incentive to address this imbalance openly. Achieving sustainable growth in data centers and AI will therefore require a fundamental rebalancing—one that must begin with transparency.
My research shows that the sector as a whole resists accountability and withholds critical information needed to assess the true impacts of (generative) AI.
What does seem certain, however, is that data center operators will continue to shift the full cost of their growing negative externalities onto society.
For 2026, the upper end of these ranges effectively marks the starting point for the current year. How this will evolve remains uncertain, with the most pressing question being whether technology companies will be able to secure sufficient power to meet their rapidly growing demand.
Operating AI systems worldwide may therefore result in a carbon footprint comparable to that of New York City in 2025, while the water footprint could fall within the range of global annual bottled water consumption.
This power demand translates into a carbon footprint of between 32.6 and 79.7 million tons of CO₂ emissions in 2025, while the associated water footprint could reach 312.5–764.6 billion liters.
This level of electricity consumption is comparable to that of Bitcoin mining and is approaching half of total global data center electricity consumption (ex crypto mining) in 2024. To put this into further perspective, it is nearing the amount of electricity required by a country such as the UK.
Let’s start the year by reflecting on the colossal growth in the resource consumption of AI and what it means for 2026.
At the start of 2025, the worldwide power demand of AI systems was approximately 9.4 GW. By the end of the year, this demand had likely increased to around 23 GW.
The UK’s largest planned data centre—a project by US firm QTS—is understating its true water consumption by a factor of 50, says new analysis by @watershed-i.bsky.social & @aiarmsracebook.bsky.social @digiconomist.bsky.social No mistake; AI is a heavy, polluting industry. shorturl.at/KXQxJ
HEY friends: recently seen heaps of "disclosures" from AI companies of the energy cost per question you type into chatbots (most recently from Google)?
Presenting per-query 'efficiency' is part of big tech's efforts to hide its extremely real and happening-right-now climate impacts!
NEW POST -->>>
It also ignores that Google’s total power demand has gone up by a massive 50% over the past two years (driven by AI) despite all the reported efficiency gains (a classic Jevons’ paradox).
Google released some new data on the environmental impact Gemini AI prompts, but omitted so many details you can hardly consider it useful. Instead, it does paint an overly rosy picture of the impact concerned.
It also ignores that Google’s total power demand has gone up by a massive 50% over the past two years (driven by AI) despite all the reported efficiency gains (a classic Jevons’ paradox).
Google released some new data on the environmental impact Gemini AI prompts, but omitted so many details you can hardly consider it useful. Instead, it does paint an overly rosy picture of the impact concerned.
Bitcoin mining has been nothing more than a huge energy waste in past years. Now AI is about to surpass Bitcoin’s power demand. Emerging tech is rapidly driving up the power demand of worldwide digital infrastructure with the inevitable consequence being strained power grids and more fossil fuel use
Microsoft’s sustainability report that covers 2024 is now available - and it contains some interesting highlights so let’s dive in!
blogs.microsoft.com/on-the-issue...
The Verge covering my latest research on the power demand of AI and how this is likely to soon surpass another energy hog in the digital space: Bitcoin mining.
www.theverge.com/climate-chan...
At the same time, there’s not even a breakdown of scope 3 emissions per region - so it’s not remotely possible to see what is going on here.
“We are working with governments, suppliers, peers, industry associations, and others [] to improve carbon-free electricity access in markets where we have a significant supply chain footprint, most notably in Korea, Japan, and Taiwan, where the majority of Microsoft's semiconductors are sourced.”
The key regions for AI chip manufacturing (Taiwan, Korea and Japan) are highlighted as a challenge (as these regions heavily rely on fossil fuels):
Another thing that is not clear from the report is how AI semiconductor manufacturing is impacting the report. Microsoft only writes:
“At Microsoft, 97% of our emissions are in the Scope 3 category, with the majority of these emissions within our supply chain.”
“As a sector, building materials such as steel and cement are currently some of the highest contributors globally to the carbon cost of new construction, together producing an estimated 13.5% of global carbon emissions.”
It’s also worth noting that building out all this new (AI) data center infratructure also has a significant impact: