Blog
25/09/2025
AI’s Hidden Costs: Why GreenOps and FinOps Must Converge
Artificial Intelligence is changing the way organisations work, innovate, and compete at a blistering pace. But with that opportunity comes cost and sustainability challenges that many leaders are not yet equipped to address.
Two emerging disciplines, GreenOps and FinOps, are at the heart of tackling this problem alongside the more traditional ITAM practice. AI can be, and is, purchased and managed in multiple different ways, commonly as SaaS – which incorporates ITAM and FinOps practices, on public cloud, and on premises - which is often the greenest option.
This post is not about giving you all the answers. Instead, it is about helping you have that first conversation, whether inside your business, at an event, or in a discovery workshop. Awareness is the first step towards better decisions.
What Do We Mean by GreenOps?
At its simplest, GreenOps stands for Green Operations. In practice, we are usually talking about Cloud GreenOps, measuring and managing the environmental impact of digital infrastructure and AI workloads.
Currently, the field is still relatively new. Groups such as the TBM Council and the Green Software Foundation are trying to define standards, but much of the work comes down to two things:
- Carbon impact (CO2e): The greenhouse gas emissions from running workloads. CO2e, “carbon dioxide equivalent”, is a standard way of comparing all greenhouse gases on a common scale. For example, methane is 14 times more impactful than carbon dioxide.
- Water usage: Massive amounts of water are used to cool data centres. In some regions, this creates real trade-offs with local communities that also need that water.
To put it in perspective:
- Training GPT 3 used an estimated 5.4 million litres of water.
- GPT 4’s training consumed energy equivalent to 6,000 US homes for a year.
- Even writing a 100-word email with AI consumes water, roughly a pint.
It sounds extreme, but that is the reality: AI is resource intensive. (Source: AI Needs Your Help)
The good news is that you don't need to have the perfect data or solutions on day one. What matters is getting started. Tools like Cloud Carbon Footprint or commercial solutions like Green Pixie can provide visibility into your digital carbon footprint. Once you know, you can begin to improve.
What About FinOps for AI?
If GreenOps is about sustainability, FinOps is about cost and financial value.
FinOps is the discipline of bringing financial accountability to cloud and now AI spend. And when it comes to AI, this is more important than ever:
- AI SaaS costs can balloon quickly. Tools such as Copilot or ChatGPT often start at a low price, sometimes even as a loss leader, but as reliance grows, vendors may increase their pricing.
- Building your own AI is a complex trade-off. Running models in the cloud may be more flexible, but on-premises deployments can be both more cost-effective and environmentally friendly.
- Token usage drives cost. If one team member prompts efficiently while another takes 20 times more tokens to reach the same answer, that inefficiency becomes real money.
FinOps for AI is not just about saving money; it is about avoiding surprises and making intentional choices between SaaS, cloud, and on-premises options.
Why Convergence Matters
Here is where it gets interesting: AI forces convergence.
- If you only think about cost, i.e. FinOps, you may choose an unsustainable option.
- If you don’t understand the user based and consumption based SaaS licensing models and usage in the organisation your AI efficiency could quickly get out of control.
- If you only think about sustainability, i.e. GreenOps, you may pick an option that is too expensive or impractical.
- If you do not think about either, you risk lock-in, price hikes, and reputational damage.
By bringing GreenOps and FinOps together (along with ITAM disciplines), organisations can weigh trade-offs across cost, speed, performance, and impact, which is a progression from the well known “iron triangle” into a more complex decision making paradigm
For example:
- Do you run AI inference in the cloud, or host a smaller model on-premises?
- Do you prioritise cheapest storage, or optimise for carbon efficiency?
- Do you accept higher costs today for a more sustainable footprint tomorrow?
The point is not that there is a single “right” answer; it is about making informed decisions instead of flying blind.
Where to Start
Our approach is straightforward: we help teams initiate the conversation. That is what our Kickstarter workshops are designed to do. They do not require a major commitment, and they do not demand perfection. They are about:
- Making sustainability and cost visible.
- Explaining key terms such as CO2e, PUE (power usage effectiveness), and token pricing.
- Showing practical examples and trade-offs.
- Identifying quick wins that build momentum.
Whether it is FinOps, GreenOps, or AI, the most significant risk is not making the wrong decision. It is not realising the impact of your choices at all.
Final Thought
AI is accelerating faster than any technology in history. The choices organisations make now about where and how they run AI will shape both their financial future and their environmental footprint.
Starting the conversation about GreenOps and FinOps today ensures those choices are intentional, not accidental.
Come and see us at SAMS Europe
We’re delighted to be sponsoring SAMS Europe in Amsterdam this October, and will be heading with our multi-skilled team. We’ll also be exhibiting in the main exhibition area so watch out for the stand and our team as you grab a coffee.
If you’ve got a specific challenge you’d like to discuss we highly recommend prebooking some time with the team as their diaries are already filling up. Book a meeting with this us without delay.
