Many of us are used to calculating how much CO2 we are saving as a result of the items we repair, using either Fixometer or Farnham’s Carbon Calculator, but has anyone any experience of converting those figures in water saving. We know that both energy and manufacturing are water intensive and that water is one of our most treasured resources available.
I know there are calculation formulas out there, but how relaiable are they?
Great question! Totally agreed that it would be great to know roughly how much water we could save by not manufacturing each new product.
When we were building the current version of the Fixometer’s CO2e calculator back in 2021, we did consider adding water consumption as an additional metric. At the time we found that there simply wasn’t enough product-specific information out there about water consumption at each of the various stages of a product’s lifcycle (extraction, manufacturing, transport, use, end of life etc.)
Press here for more detail about that
From the 1,600+ reports and papers we sourced, we found that most companies only report the minimum types of impacts they can. While that often (though not always) included CO2e, it very rarely included water consumption. To pick one example, Apple (one of the companies that actually provides more environmental impact data about its products) still doesn’t report on water use for its latest devices (see their recently announced M5 MacBook Air report here). They do have a ‘[u]Water Strategy[/u]’, but again, there are no real figures here.
Most of the water consumption figures we did find came from papers in academic journals. But while useful, due to differing methodologies and the relatively small number of products covered, we couldn’t determine a predictable relationship between CO2e production and water consumption, especially when comparing products across different categories (e.g. textiles vs. electrical devices vs electronics).
As a result, we weren’t able to add a water consumption calculator.
I have to confess that I haven’t really kept up with this closely enough over the last couple of years. But I’d be a little sceptical of formulae that claim to calculate water consumption for a few reasons:
CO2e figures themselves are generally estimates at best. Calculating water consumption based only on this is essentially estimating based on an estimate, making it potentially very unreliable.
This is especially true if the calculator doesn’t account for the type of product in question. E.g. it takes something like 10 litres of water to produce 1 kg of plastic, but somewhere around 7,000-10,000 litres to produce 1 kg of cotton. But producing plastic is far more cabon-intensive; 1 kg of cotton releases around 8 kg of CO2e, roughly equivilent to a single plastic bag weighing mere grams. And electronics/electricals are much more complex.
Like CO2e, water consumption will vary considerably depending on where and how the product was manufactured.
All that said, I do need to update myself in this area. Do you have examples of these calculation formulae?
I agree with James: CO2e is dubious enough, so water will just be even more dubious.
Also, in the same way that manufacturers have a vested interest in minimising their environmental impact to boost their credibility, I think that obsessing over numbers like grammes of CO2e & litres of water is bad for repairers because it’s an “easy win” but misses so many other issues and may encourage us to focus on high-CO2e or high water consumption devices, whereas in fact some devices may be far worse in other respects.
For example, I personally would prefer to see measures like:
“How many people (especially children) died while mining the minerals used to make this device?”
“How many guns did the mining of those minerals buy for a murderous dictator’s army?”
“How many days were lives shortened due to the pollution emitted during the manufacture of this device?”
In a perverse way, any broken device that queries an LLM (what too many confusingly call AI) is likely to use less water than a working one.
Kat Austen who is doing an artistic and research project about the impact of water of a data centre in Korea wrote today:
How much water is consumed annually by embodied AI is hidden behind a lack of reporting, but the most recent scientific article on the subject estimates that the yearly footprint of these models could equal the that of the global consumption of bottled water (de Vries-Gao, 2026). And indeed the water used up by AI is usually the purest of the pure — because if it were dirty it would damage the processors, clog up the pipes: it would not be good for the body. What happens to this water? Primarily, water is used to remove heat from the processors.
And of course the impact on the life living in this water is negative.
On a different note related to Dave’s point, the war in Iran is making visible some usually invisibile flows of material required to create some tech, for instance that of helium. Here’s the Hankoreh about it:
The semiconductor industry, which has benefited from the AI boom, is also tense about supply chain hiccups originating in the Middle East. Helium is a critical element for wafer cooling, an important process in semiconductor manufacturing. The country’s chip industry relies on imports of helium, 64.7% of which come from Qatar. A whopping 97.5% of imports of bromine, used in the wafer etching process, come from Israel.
The Ministry of Trade, Industry and Resources is evaluating the supply situation of 14 product categories, including bromine, helium, and various metrology tools (high-precision instruments necessary for inspecting and verifying the properties of chips and wafers). The ministry is exploring measures to diversify relevant supply chains if the Middle East situation stretches into the long term.
It’s complex and one can only approximate the demands of some of the resources needed and the impact on humans, non-human animals and our planet.