green IT

Is GreenXAI an illusion?

GreenXAI, an illusion?

GreenXAI, an illusion? 308 163 Cordny

Green IT, that’s one of my favorite subjects these days here and on LinkedIn.

But did you know TestingSaaS also works in the field of Explainable AI aka XAI?
And that he combines it in GreenXAI.
Let me explain.


Is GreenXAI an illusion?

source image: https://en.imna.ir/news/807163/AI-Emerges-as-a-Vital-Tool-for-Environmental-Protection-Sustainability

What is XAI?

AI is for people a black box. You put some data in the AI app (like ChatGPT) and you receive the output. But can you explain how the output is created?

Here XAI joins the stage. This is a collection of methods and processes, enabling AI users to understand and trust the AI output.
The so called post-hoc methods are applied after the model is trained.
SHAP and LIME are examples of these, which I use daily in my work.

Because I am also interested in green IT, I was wondering how green these post-hoc methods are.


How green is XAI?

Well, they are not, because they run additional computations on top of the AI model. This means they increase runtime resulting in extra CPU/GPU usage, consuming more energy compared to just running the model alone.

But is that the whole story?

No, they expose bias and errors faster resulting in a reduction of wasted compute on poorly performing models.
Also they enable more efficient model design. By identifying which features truly matter, you can retrain a smaller, leaner, greener model.
And last, but not least:  they can be paired with tools like CodeCarbon to connect explanations with sustainability metrics. This increases the transparency in energy measurement.


Is greenXAI an illusion?

So, is green XAI an illusion by using these post-hoc methods?

SHAP and LIME are not inherently “green IT” methods, but they can play a green role within the AI development lifecycle by preventing waste and helping optimize models for efficiency.

I will still use these methods because I want to find out more about how the AI output is created. In the meantime I will make my coding greener when possible.

Are you interested in my GreenIT and XAI work? Just contact me and let’s see how I can help you.

Why you can't measure the AI consumption of different cloud computing vendors without a proper standard for GreenIT?

Comparing apples with oranges in Green IT

Comparing apples with oranges in Green IT 1200 768 Cordny

Comparing apples with oranges?
Yes, that’s what I think is going on when GreenIT professionals are comparing cloud computing vendors on their energy costs per LLM query.

Comparing apples with oranges?
Yes, that’s what I think is going on when GreenIT professionals are comparing cloud computing vendors on their energy costs per LLM query.
Last week Google published an article about how environmental impact on AI inference is measured by them. And whole LinkedIn went wild. And it was polarizing, supporters and critics falling over each other trying to shout the hardest.

But what did Google measure? That’s what TestingSaaS will find out.


Measuring environmental impact on AI inference by the cloud/AI providers

Google

First of all, Google measured the energy costs of a single Gemini text prompt (text, not another medium, which costs a lot more energy). The study focuses on a broad look, including not only the power used by the AI chips that run models but also by all the other infrastructure needed to support that hardware like water consumption, cooling etc.


The estimation results: the median Gemini Apps text prompt uses 0.24 watt-hours (Wh) of energy, emits 0.03 grams of carbon dioxide equivalent (gCO2e), and consumes 0.26 milliliters (or about five drops) of water.


How Google did this is explained in their technical paper. It goes too far to explain all here.
To have a better undertanding of these numbers Google stated:


The Gemini Apps text prompt uses less energy than watching nine seconds of television (0.24 Wh) and consumes the equivalent of five drops of water (0.26 mL) and 0.03 grams of carbon dioxide (market estimate)

And although the Google scientists also mentioned some critical remarks in their paper and article (median, market estimate etc.) LinkedIn went in critical mode. Just query LinkedIn on ” google gemini AI energy’ and you will find enough positive and negative posts on this subject.

Mistral

Last July, Mistral AI published a full life cycle assessment (LCA):

the environmental footprint of training Mistral Large 2: as of January 2025, and after 18 months of usage, Large 2 generated the following impacts:

20,4 ktCO₂e, 

281 000 m3 of water consumed, 

and 660 kg Sb eq (standard unit for resource depletion). 

the marginal impacts of inference, more precisely the use of their AI assistant Le Chat for a 400-token response – excluding users’ terminals:

1.14 gCO₂e, 

45 mL of water, 

and 0.16 mg of Sb eq. 

Awesome, now we can compare the results with Google, or not?

Comparing Google and Mistral AI energy costs

That’s like comparing apples with oranges.


Why?

Just look at what is measured:

– the “marginal by prompt” (Google),

– the “total cost of the cycle” (Mistral).

What Google measured is completely different compared to Mistral.

Why is this wrong?
Well, if you compare different things it’s like comparing apples with oranges. They are completely different, so no comparison can be made.
It’s not a standard you can compare.

What to do now?

So instead of criticizing the reports of how AI energy consumption is measured by the cloud computing vendors, why not figure out together what a suitable measurement standard could be for AI energy consumption in GreenIT?
Or not Green Software Foundation?

That would make the world less polarizing as it already is.
We’re engineers, let the politics out of it!










Green IT vs. Sustainable IT

Green or sustainable IT?

Green or sustainable IT? 1024 1024 Cordny

Last week I had a great meetup with like-minded people about IT and getting it more greener.
Members from the Green Software Foundation and Sustainable IT Netherlands communities together.
Passion for technology converged with a shared commitment to sustainability.
That evening something was lingering in my mind, but I could not grasp it.
A few days later it struck me:

Green IT is not equal to Sustainable IT

Let me explain.

Why is Green IT not equal to Sustainable IT?

Green IT and sustainable IT, both terms are used frequently on social media, especially on LinkedIn, when promoting the use of IT measures against climate change. We use them interchangeably without even knowing it. But they are not the same.
We first have to see what these terms mean seperately.

What is Green IT?

In my honest opinion, Green IT, as also stated by the Green Software Foundation (GSF) is software that is responsible for emitting fewer greenhouse gases like CO2. So, less CO2, less energy, less waste etc.
It’s a tech thing, trying to solve a prblem quick, result driven.
But what’s then the difference with Sustainable IT?

What is Sustainable IT?

Sustainable IT refers to the design, manufacture, use, and disposal of IT systems and infrastructure with minimal negative impact on the environment, while also being socially responsible and economically viable.
The last part of this definition shows the difference with Green IT: while also being socially responsible and economically viable. Let me explain this further.

The difference between Green IT and Sustainable IT?

Sustainable IT is not only about IT, it involves also the social interaction. It’s a process, get IT sustainable.
How? By implementing processes, not only involving IT pro’s, but other people too.
And processes take time, it’s not a quick fix, it needs commitment. It is slower.
And it also affects the economy. How can we build solutions for sustainable IT that lasts a long time?

In other words, Green IT is a part of Sustainable IT.

What is the role of TestingSaaS here?

TestingSaaS has a green mission: help to reduce the emission of green house gases by IT.
With his knowledge of software testing, documentation, and green IT Cordny Nederkoorn
helps Small and Medium-sized Enterprises get their software testing and document creation greener, so they can create sutainable products.
Not only green IT, but also sustainable IT, by creating awareness at these companies and clients.

Would you like to discover what TestingSaaS can do for your organization?

📅 Schedule a free exploratory call via
https://lnkd.in/eAXUVjBS

or send me a direct message.

Let’s build sustainable IT together, for a lasting world!!









TestingSaaS goes green IT

TestingSaaS on a new mission: a green mission!

TestingSaaS on a new mission: a green mission! 502 356 Cordny

Remember I said it was time for some experiments?

Well, they started.

TestingSaaS goes green IT, just like its logo.

It started with a spark, erupting into a flame and last Wednesday a flow started to spread.

How did this happen?

The spark which erupted the flame: a book called Green IT

A few months ago I was walking through my favorite bookshop in Oss like Ernest Hemingway did a hundred years ago in the famous bookshop Shakespeare and Company in Paris. I always relax here, just browsing through books. Then I saw a book called Green IT by Jan Hoogstra and Eric Consten.

While reading the cover I got intrigued and I bought it. Yeah, the current IT (especially AI) is having a problematic impact on the climate. And this book gives , next to some theory, also practical examples how to decrease this impact. Software, hardware, networks, data centers and utilities, they all can help this do it. Companies and organisations like TNO, AFAS and the government are doing it.

This got me thinking, how could I help with my company?
Not only by getting more sustainable with my company, but also more as a real player in this new field.

A spark became a flame.

Then I met Wilco Burggraaf on LinkedIN.

The flow: Meeting the people from Sustainable IT Netherlands and the Green Software Foundation

Following Wilco, the Dutch Green Software Champion, expanded my network in Green IT, with people like Robert Keus, a social entrepreneur revolutionizing the way technology intersects with society. And developing a  first green AI chat, reducing the impact AI has on our environment by running on sustainable infrastructure and by repurposing heat. Chatbots eeh, where did I do that before ?


Man, I had to meet these people, but how? As an entrepreneur I’m also quite busy.
Let’s see if there are meetups where they are involved, and yes, there are.

One of them was last Wednesday, the 28th of May 2025.  A meetup from Green Software Foundation and Sustainable IT Netherlands communities, where passion for technology converges with a shared commitment to sustainability.
Hosted by Thorsten Picard at Capgemini HQ Utrecht.
This was the time to get that flow going!

The Green IT flow

The evening was wonderful. I finally met Wilco and Robert and a lot of other people, a real organic gathering.

I heard about the ‘Green Software Foundation’, and I was very happy to also meet people from ‘Sustainable IT Netherlands’. Corina Milosoiu and Chris Stapper, very delighted to have met you.

But a meetup is not a meetup when there are no talks.

Robert kicked it of, together with Cas Burggraaf, an energetic and eye-opening session, which proved that a talk about GenAI can be about so much more than just numbers and figures.

Then the stage was ready for a lighthearted talk by Mirko van der Maat from Capgemini about Sustainability in Architecture and Barbapapa. Oh man, Barbapapa, forgotten memories.

At the end it was time for some pitches, which were received very well.

But hey, what is a meetup without some drinks and snacks, very well facilitaded by our host?

Time for some networking. Great talks with very passionate people with one thing in common: Green IT!

TestingSaaS going green: the future?

Ok, we had a spark that erupted in a flame, becoming a flow.

Well, I want a good Aussie bush fire, I want to create a flood.

Yes, I remember your books Rijn Vogelaar.

And I can’t do it alone.

With my new friends from the Green Software Foundation and Sustainable IT Netherlands communities I can.

How? By Creating Content through Testing!

To be continued!!