Reads, takes and links. Posted here, before I forget them.
-
Are the big hyperscalers reporting power consumption in Germany like the laws say they’re supposed to?
I noticed something recently that I can’t find an clear explanation for, so I’m posting it here to help understand what’s going on. In the European Union, there is a sort of law called the Energy Efficiency Directive, which I’ve written about before, and this relevant when we think about data centres and decarbonising the…
-
Default to calling it ‘carbon pollution’, not ‘carbon emissions’, when talking about climate change
I’ve heard a few people use this specific framing of greenhouse gas emissions when talking about climate change, but it hasn’t become a real habit for me. This is a quick note for me to refer back to later. There is a now lots and lots of peer reviewed research, and advice from communications professionals…
-
Can you phase out oil and gas in the UK with CFDs at 14 GBP per MWh?
This is a quick note, because a report I was keeping an eye out for is now in the public domain. As I understand it, it was commissioned by Dale Vince of Ecotricity, and presented at the UK Labour party in September, and lays out a ‘mid-transition’ argument for effectively bailing out oil and gas…
-
Even with proprietary GenAI models, you can influence the environmental impact of use through supplier choice
I came across an interesting paper, How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference via a post on social media by a friend of mine, Asim Hussein, and one of the details surprised me enough to want to capture it here. What was worth writing down then? It’s on page…
-
Google’s expenditure on AI is eye watering
Before I saw this chart, I knew that the hyperscalers had been spending huge amounts on datacentre infrastructure, but I had some vague idea that Amazon was spending the most, followed by Microsoft, then and Google. Boy was I wrong. Google has been outspending the other huge hyperscalers since 2018 This chart is from an…
-
Finding nice images in the vector space, and AI the tool vs AI the project
It’s the weekend, I’m trying to close down a bunch of tabs, and one tab I hadn’t realised I had open, was this piece by Andy Masley about AI-generated images he’s been creating with Midjourney. Before it disappears into the either, I figured it was worth jotting a few thoughts down. Here’s the quote(s) that…
-
What’s the single thing you would do to make a web service more sustainable these days?
From between 2018 and April 2025, I was an organiser of an online community called climateAction.tech. I’ve written about it a few times before on this blog. In the community slack, someone who I respect asked a valid question: For context, they work in a business selling what I consider domain names at the higher…
-
A few notes on the quest for £1/kg H2, and relating it to fossil free datacentres
This post from Rivan, a hydrogen startup, makes a few key points that I think are really interesting in the context of the research into fossil-free molecules you might need for aviation, shipping, all year round power generation, and a fossil free internet. I’ve mainly written for my future self as much as anyone else,…
-
Things you forget to ask when doing a podcast about AI power spikes and hardware standards
I’m sharing this here to come back to later, as I host a podcast called Environment Variables and I recently did an interview with some folks who are working on getting a standards group off the ground focussed on how datacentre hardware and software work together. There was one claim that during the interview I didn’t…
-
Is the problem just GenAI?
I’m trying to make sense of some of the figures shared when we talk about the environmental impact of AI, and what the specific drivers of it are, in terms of the specific kinds of AI being deployed around the world. Here are some notes, and a request for help to point somekind of useful…
Got any book recommendations?