Sustainable IT is entering a new phase of challenges. While technology continues to evolve rapidly, the rise of Generative Artificial Intelligence marks a major shift. Given the environmental and social impacts of AI, the field of Sustainable IT must reassess its approach. At the same time, the current geopolitical climate, where productivity is increasingly prioritized, sometimes at the expense of sustainability commitments, especially within EU regulations, also plays a key role in this needed reorientation.
Artificial Intelligence is rapidly becoming an expected pillar of productivity and competitiveness. From customer service bots to powerful generative tools, the promise is clear: streamlined processes, enhanced innovation, and economic advantage. Companies see AI as a critical lever for future growth. However, this transformation comes with hidden costs and risks that cannot be overlooked in a sustainable IT approach.
The rise of Generative AI has driven up energy consumption dramatically. While some actors, like Ecologits or Hugging Face, provide tools to evaluate environmental impact, measuring the carbon footprint of a single prompt or training cycle remains complex with LLM. The evaluation of AI’s footprint is still in its infancy. Meanwhile, AI-enhanced predictive maintenance or digital twins for example are being used to reduce environmental impacts, illustrating the dual nature of AI: both part of the problem and a potential help for certain problems.
As AI deployments scale, costs become a growing concern. Gartner warned in 2024 that expenses could reach five to ten times initial projections. Gartner explained this at its symposium in Australia, highlighting poor use cases, such as replacing simple searches with expensive LLM queries, and the fact that more complex queries mean more tokens.
Beyond cost, cybersecurity and sovereignty are rising concerns. AI empowers defensive tools, detecting anomalies and threats in real time. Yet, the same algorithms can be weaponized by malicious actors. Sovereignty, too, is at stake: the ability of organisations and nations to control their AI stack (data, models, infrastructure) is critical for long-term resilience. And now, with the advent of Agentic AI, autonomous agents that operate on behalf of users, we may be entering a new chapter in the AI revolution, raising even deeper ethical and systemic questions.
The AI boom doesn’t mean innovation can remain unchecked. We now face a fundamental challenge: how to innovate without exceeding planetary boundaries. Sustainability must be designed into the core of our digital systems. This doesn’t mean halting progress, but rather reframing it through new lenses: frugality, circularity, and ethics. The question is no longer just what technology can do, but what it should do.
Frugal AI exemplifies this mindset. It calls for efficiency not only in computing, but also in purpose. Do we need an LLM to answer this question? Can simpler models deliver sufficient value? Reducing unnecessary model use helps mitigate environmental and financial costs while promoting more deliberate design.
IoT, another frontier of digital expansion, poses similar questions. Connected devices are multiplying, increasing emissions and complexity. Sustainable IT requires serious strategies for device lifecycle management, from design and production to repair and reuse. Ecodesign and second-hand markets are essential pillars. Reparation, modularity, and reuse must become standard practices, not niche experiments.
Ethics is not a constraint but a catalyst for sustainable innovation. Governance frameworks must evolve from compliance tools to strategic assets. The concept of Holistic Return on Ethics (HROE) reframes value creation to include risk mitigation, stakeholder trust, and social impact.
Labs like Les Augures in France push these principles further by embedding user research, collective intelligence, and low-tech design into their innovation processes. Sustainable tech becomes not just what we build, but how we build it.
Regulation is meant to guide transformation, but today’s legislative landscape reflects tension between urgency and hesitation. The Ecodesign for Sustainable Products Regulation (ESPR) is a landmark step, but its implementation so far reveals the fragility of ambition in the face of economic pressure. The new working plan reduced its scope, and key categories like ICT were postponed. While the first rules on smartphones and tablets will take effect in June 2025, the broader strategy remains in flux.
The AI Act faces a similar fate. As of May 2025, rumours suggest that enforcement of some provisions could be delayed. This potential backpedaling aligns with the EU’s wider deregulatory push. While simplification may help businesses adapt, it would risk diluting the transformative impact of Europe’s AI governance framework.
Another key milestone, the European Accessibility Act, enters into force on 28 June 2025. This law could reshape digital inclusion, yet accessibility still struggles to gain traction in corporate strategies. Will the Act trigger a culture shift, or will it become another well-meaning policy with limited impact? The answer may depend on enforcement and on whether consumers and civil society hold companies to account.
Finally, the Corporate Sustainability Reporting Directive (CSRD) could be either a catalyst or a checkbox. With the Omnibus directive threatening to reduce its scope, the momentum for ESG transformation is at risk. Originally envisioned as a mechanism to align business models with sustainability goals, CSRD may now be sidelined by short-term economic considerations. For IT departments, this could mean missed opportunities to drive sustainable practices from the inside out.
The future of Sustainable IT will be defined by its ability to maintain its vision and adapt in a changing context. As AI reshapes industry standards and regulations continue to shift, the challenge is no longer to prove that transformation is necessary, but to decide how, and on what terms, it will unfold. The path ahead demands clarity of purpose, ethical conviction, and an uncompromising commitment to long-term resilience.
True digital sustainability means reconciling efficiency with purpose, productivity with equity, and growth with planetary boundaries. Whether through smarter regulation, frugal innovation, or ethics-driven design, the IT sector holds the tools and the responsibility to reimagine itself as part of the solution.
All these subjects will be debated by the speakers of GreenTech Forum Brussels 2025 during the conferences and round tables on 17 and 18 June at the Maison de la Poste in Brussels.
Programme details: https://www.greentech-forum-brussels.com/2025/program
Article written by Rémy Marrone for GreenTech Forum Brussels