Sweating for digital waterfalls; PS5s out of China; Bad reception on fast trains; and exploding coal turbines! It's GTG Links 65
Greetings, GTG readers – I trust those of you in the northern hemisphere are making the most of the late summer vibes; meanwhile, in Melbourne, spring almost sprung. I know because I was swooped by the world's most aggressive magpie today.
Lots more data centre details, some history, some cool AR device component diagrams, and a bunch of climate impacts to witness and take note of. Lets dive in.
Solidarity with Arkane workers 💪
Wonderful and encouraging to see game workers standing up for Palestinian lives.

And Stephen Totilo has an interview with one of them.

Coding waterfalls through summer heat
This story from the development of an influential Japanese shoot-em-up from 1990 goes into the material-heat nexus of game development. Brendan Keogh pointed out the similarities to the opening chapter of Digital Games After Climate Change.
The background on this waterfall, from the game's director Takatsuna Senba (the whole interview is pretty great stuff: web.archive.org/web/20191230...)
— Sam Barlow 🔥 (@mrsambarlow.bsky.social) 2025-08-23T20:11:24.438Z
Sony has shifted PS5 production outside of China
This seems like a pretty big change – the piece doesn't say where outside of China they're now being produced (India? Vietnam? Taiwan?) but clearly it's a big change. I can only speculate about how much of a hassle this must have been, and what sort of issues might crop up from this. Will be very interesting to see the effects.

Diminishing graphical returns via Indiana Jones' hair
Graphical sufficiency – you're hearing it more and more, folks!
There's never been a better example of the diminishing returns in graphical improvements. I thought this was a post from Hard Drive.
— Mat Piscatella (@matpiscatella.bsky.social) 2025-08-21T22:52:41.201Z
VR/AR device energy consumption
I came across this great paper on “Energy Demand in AR Applications” recently. It goes into some really detailed, almost forensic energy analysis of specific features and components of the HoloLens 2 device. Just cast your eyes over this gorgeously detailed system diagram, phwoooaaar! You don't see that every day.

Some of the findings of the energy testing might be cautiously generalizable to other AR/VR devices as well:
With efficiency in mind, we aim to deliver certain visual information to the user, using as little as possible power drawn (Watts). Aside from the trivial conclusion that a smaller/simpler scene implies smaller energy consumption, the optimizations might include:
• the selection of engine—our results favor Unity over Unreal.
• the scene design should remove or hide unused objects from the user’s view; this is because we have identified that visual coverage of the display by the contents contributes to the overall energy consumption.
• by employing occlusion of objects in UI/UX, if used smartly (e.g., layered layout), it could reduce the energy drawn by displaying mostly first-plane objects.
• also, the size of the displayed object may matter; however, it would be a trade-off with usability since larger objects are easier to notice and reach.
A plugin to adapt Wordpress to be “grid-aware”

Data Centre Energy Efficiency Analysis
The ever-excellent Green Web Foundation has done some analysis of the EU energy efficiency directive's new data centre disclosures so we don't have to. Most interesting are the graphs comparing PUE (power usage effectiveness: basically, how much electricity is "wasted" on cooling rather than useful computation) vs WUE (water usage effectiveness). Small DCs appear less cooling efficient, however larger ones are less water efficient – my guess is this is because the bigger ones are all using evaporative cooling rather than heat pumps/air conditioning. Maybe the economics work out that way for hyperscalers? But that's a problem for fresh water supplies if DC expansion continues...
Also the renewable energy factors were surprisingly high for all sizes of DC… though ~70% of renewable energy came via "Guarantees of Origin" (in other words, certificates of having purchased renewables from a marketplace) rather than power purchase agreements, which are a more direct arrangement, and which often result in actual new renewable generation being built. So, some good news, but not quite perfect.

AI query energy intensity in context
Ketan Joshi has a really good analysis of the new Google AI text query calculations that was recently published, putting it into a broader context that I think is quite valuable. I've seen too many people taking the numbers and comparisons that Google chooses to make at face value and using that to dismiss concerns over the energy use of AI, which seems like a mistake. Very very small things done in great numbers can still really add up to big impacts – this seems to be the lesson we have to keep re-learning over and over.

And some potential DC impacts very close to home - data centres are applying to use too much water in Melbourne. But I need that water to live!
New niche DC energy measurement method
I came across this new(ish) method for measuring bare metal server energy consumption, which proposes to do it via reading data from *water cooling* temperatures from servers that are liquid cooled. Note that this isn't referring to evaporative water cooling for entire DCs, but specific servers that are water cooled on chips – though I'm not sure prevalent that is? Maybe it's mainly in high performance computing contexts? Maybe not.
Anyway, it's interesting to me because, a) I hadn't considered the privacy dimension to renting bare metal servers, and b) it just seems like a cool thing to be able to do, if somewhat niche in application. Here's the abstract:
Numerous cloud providers offer physical servers for rental in bare metal paradigm. This mode gives customers total control over hardware resources, but limits cloud providers’ visibility of their usage. Accurately measuring server energy consumption in this context represents a major challenge, as installing physical energy meters is both costly and complex. Existing energy models are generally based on system usage data, which is incompatible with the general privacy policies of bare-metal server contracts. To deal with these problems, it is imperative to develop new approaches for estimating the energy consumption of these servers. This paper presents an original non-intrusive method for estimating the energy consumption of a server cooled by direct-chip liquid-cooling, based on the coolant temperature and the processor temperature obtained via IPMI. Our approach is evaluated on an experiment carried out on 19 bare metal servers of a production infrastructure equipped with physical wattmeters.

Climate insurance is the pointy edge of the climate wedge
If you want to know where a climate financial crisis is most likely to start, it's in insurance – which makes this warning all the more worrying.

Meanwhile, the climate continues to bring major impacts that scientists have been warning of for years...
Flood devastation on island of Cape Verde

Arizona heat deaths

Heat wave in Iran

And it got so bad it disrupted the energy system.
Here's a curveball for sustainable travel – bad phone service!
Turns out there’s are some real physical challenges to getting a good internet connection on high speed rail, particularly in Europe. A minor annoyance, perhaps, but also a potential barrier to more sustainable transport options.

Now for some good news!
Batteries are meeting a quarter of the Californian evening peak energy demand! Yeeeeow! Big batteries rule, and the Australian and Californian experience is really showing just how effective a little bit of storage (relatively speaking) can be.
Batteries and solar are fast becoming the backbone of a clean, flexible power system🔋☀️ In our #GER2025, we saw California batteries supplied ~20% of evening peak demand in June 2024, displacing gas. Fast forward to today: they’re now averaging 28% ⚡ ember-energy.org/lat...
— Ember (@ember-energy.org) 2025-08-21T13:00:59.000Z
And your random bit of energy trivia
Do you want to know how and why a coal turbine in Queensland blew up in 2021? It involves the turbine going from being an energy generator to a massive power sink – basically turning into a giant electric motor until it blew itself apart. Absolutely fascinating this, if you've ever wondered about how big generators like this work.
Super recommend this one!
Thanks for reading Greening the Games Industry. If you made it this far, hell yeah. GTG readers – especially GTG Supporters who unequivocally rock and help GTG keep existing with a monthly or annual contribution – make this all worthwhile.
I'll be back soon with more from what we've been up to at the Sustainable Games Alliance – keep your eyes peeled for that, because we're getting awfully close...!
Till next time.