It was during COVID, I was talking to someone who works for one of the world leaders in the cloud and, when the conversation turned to the environment, he said to me, “We‘re doing some sh… but some cool sh… and as long as public opinion stigmatizes air travel without taking an interest in us, everything is fine”.
That sentence has been going around and around in my head. Not because I’m a climate activist, far from it, but because I don’t like it when people are uninformed or misinformed, no matter the subject.
But for some years now, the subject of the environmental impact of digital technology has been increasingly present, at least in the business world, but without the general public realizing to what extent they are concerned. It took the explosion of AI for this to become a real topic in the media.
The recent release of the study [FR] Environmental Impacts of Digital Technology in the World 2025 by the Green IT association is an opportunity to take stock of the subject.
Digital technology is not immaterial at all
I have always seen a problem with the rise of digital technology, particularly in business: of course we see the equipment but we do not see the flow of information, the dysfunctions in the way organizations operate that it generates and the impact they have on people (Hyperconnectivity in the workplace: digital becomes a burden).
Digital technology will remain a problem as long as we fail to quantify and visualize its uses, to objectify the intangible.
The same is true of what digital technology produces. Of course we see the devices and we know that they have been produced in one way or another. But we don’t see what is needed to circulate and store information… It’s easy to say to yourself “it’s in the cloud” and pass the buck, but when we talk about the cloud, we’re actually talking about the Chernobyl cloud.
“If you can’t see it, you can’t deal with it”. Fortunately, if you can’t see it, some people measure it, and you’ll see that digital technology is not as intangible as it seems. This was the case, in the scope of collaborative uses, in the latest Lecko report on the state of the art of internal transformation (Sustainable digital: no more hypocrisy ) and it is therefore the case in this report on the overall scope of digital technology.
A few figures to give an idea of the scale of the subject.
In 2023, there will be 6 active digital devices per internet user, and I find it hard to believe that this will decrease. This is a global average that therefore takes into account less advanced countries with low purchasing power, but do the math at home. You are certainly well above that average, while some populations are just starting to get equipped.
With 6 devices per internet user, the annual carbon footprint of digital technology is 1.8 billion tons of CO? equivalent, or 5.5 times that of France.
If we set ourselves the goal of limiting global warming to 1.5°, digital technology alone consumes 40% of the sustainable carbon budget per Internet user! That leaves you with only 60% to get around, get dressed, keep warm, manage… live what…
Finally, the impact of digital technology on the depletion of mineral resources now exceeds that of global warming.
A good videoconference is better than a long trip, but you might think twice before turning on your camera or embarking on a long binge-watching session on Netflix to kill a long winter Sunday.
Digital technology contributes to planetary limits
Three figures to understand how much digital technology has a major impact on our environment and that it takes a lack of good faith to look the other way.
It therefore already represents 27% of the global sustainable budget for CO2 and it is clear that it is growing exponentially compared to other sectors that are more in the spotlight and are forced to make efforts.
It also consumes 18.7% of the world budget in mineral resources.
Finally, it represents 15.3% of the budget in fine particle emissions.
Small devices have a much greater impact than data centers
At this stage, we might think that on our small scale, with our smartphones and computers, we don’t weigh much compared to what data centers and the network in the broad sense produce. Well, we are wrong.
Smartphones, connected objects, and computing in the broad sense represent 50% of the impacts. This is logical, given the number of users and the proliferation of devices.
The network and data centers, meanwhile, account for 23% and 21% of the impact respectively.
But with the rise of AI, its huge requirements and data centers continuing to spring up like mushrooms, nothing says this won’t change in the future.
“Plug, baby, plug”: the growing impact of AI
With the announcements made in recent weeks, we cannot fail to focus on AI.
Although it has only been around for 2 years, generative AI and the servers dedicated to it alone already account for 1 to 5% of global digital impacts and 4% of greenhouse gases.
To better understand these orders of magnitude, consider that data centers consume 506 TWh per year, the equivalent of the electricity consumption of Germany.
Will this decrease? I don’t see how. On the contrary, I expect an explosion.
Unpromising approaches to reducing the impact of digital technology
The report recommends certain approaches that are in fact quite in line with the alarm recently sounded by researchers, and I doubt that they will please anyone ([FR]“AI and ecological disruption: moving away from the techno-solutionist illusion”). The idea is that, contrary to popular belief, technology will not solve the problems it creates on its own and that it will be necessary to reduce equipment and usage.
The first is that less equipment is needed. Therefore, repair should be encouraged and replacement limited.
If things are done in some countries with regard to repair, I don’t see how we can limit replacement. This is already the case for individuals, who will find it difficult to say that they can no longer buy the latest smartphone, but also for businesses, for simple performance reasons.
Then there is the extension of product life and the fight against hardware and software obsolescence.
As far as material obsolescence is concerned, there is already legislation to this effect in some countries, but fighting against software obsolescence is more or less like trying to reverse the course of rivers and the direction of the wind.
Then there is the issue of optimizing usage. It’s about avoiding digital waste, arbitrating between real need and the superfluous.
So good luck with that. First of all because we’re talking about imperceptible things, or at least less so than saying that you don’t need to take your 4×4 to go shopping 10 minutes away on foot, and secondly because we’re getting into the heart of people’s lives. “We’re stopping Netflix and going to read books”? I don’t believe it, and besides, you need paper for a book.
And of course the report calls on the public authorities to encourage digital sobriety and regulate in this direction.
Let’s just remember that the ecological backlash has already contributed to the results of the last US elections and that it is a tune that we are beginning to hear in many countries, even if everything is being done to make it barely audible.
Bottom line
The environmental impact of digital technology is a reality and a problem.
Is there a way to fight it? Of course.
Is it realistic? In my opinion, absolutely not.
Already because we are touching the heart of people’s lives, but that is not the most important thing. For those who did not push the analysis of the last AI summit too far (AI Summit: Real Success or Covering Up Europe’s Flaws?), it is not just about tech and struggles between businesses for global leadership.
Tech and especially AI have become the playground for a struggle for political and ideological domination of the world between different countries or blocs of countries, and the economic development of tech is now the growth driver on which all countries are betting.
Curbing usage and limiting the consumption of hardware and data centers does not mean boring businesses, and encouraging people to be frugal is going against the strategy not of one or two states but of all the major world powers.
We often hear that the last drop of available oil will go into an airplane, and I totally agree. Now I would add that the last drop of water will go into a data center and the last gram of rare metals into a battery.
Image: IT and environment by wk1003mike via Shutterstock
