AI is intensifying the strategic rivalry between the European Union and the United States, reshaping models of industrial policy and regulatory sovereignty. Amid a flurry of investment announcements, the exposure of security vulnerabilities and the contest over global standards, one critical factor remains largely in the shadows — seldom acknowledged, scarcely quantified and rarely debated: its environmental footprint.
The environmental blind spot of a strategic technology
The silence surrounding the impact of AI is surprising. A study carried out by Sopra Steria and Opsci.ai analyzing over 3 million posts about AI on social media reveals that its environmental impact accounts for less than 1 percent of the global conversation.1 Worse still, among the 100 most influential AI personalities,2 ecological concerns are only eighth on the list of subjects they discuss most, far behind technological and economic issues.
A study carried out by Sopra Steria and Opsci.ai analyzing over 3 million posts about AI on social media reveals that its environmental impact accounts for less than 1 percent of the global conversation
AI relies on energy-intensive infrastructure that consumes resources and water, the footprint of which remains largely underestimated, poorly measured and therefore little considered in industrial and political trade-offs. This misalignment can also be explained by the trajectory of the sector itself: driven by the rise of AI, the digital sector is one of the few areas whose environmental impact is continuing to grow, contrary to the climate objectives set out in the Paris Agreement. While American players are already crushing the AI market, technological dependence must not be compounded by a setback on Europe’s carbon trajectory.
This omission undermines the credibility of any European industrial strategy built on AI. To serve as genuine drivers of transformation, the leading AI companies must bring full transparency to their environmental trajectory — one they are progressively shaping for Europe.
Measuring for action: The need for transparency and rigor
We must not rush to condemn AI, but we must insist on setting the conditions for its long-term sustainability. This means measuring its impact objectively and transparently, equipping stakeholders with the tools for informed debate, and guiding decision-makers in their technological choices. Recent research indicates that the environmental footprint of a given model can vary significantly depending on where it is assessed, the energy mix of the countries hosting the data centers,3 the duration of the training, the architecture employed and the extent to which low-carbon energy sources are used.
Breaking through the methodological vagueness means providing developers, purchasers and decision-makers with common frames of reference, impact simulators, libraries of low-carbon models and low-carbon computing infrastructures. Numerous levers for action and choice exist, provided we have the necessary data and tools.
This requirement is not a regulatory whim but a strategic steering tool. Sustainability must be given as much weight as performance or security in industrial and economic trade-offs, because it determines the very viability of Europe’s strategic autonomy. At a time when free international trade faces headwinds, and as the second phase of the AI Act — in force since August 2025 — continues to overlook environmental sustainability, transparency on environmental impact must become a prerequisite for access to European markets, financing and large-scale deployment.
Making sustainability a central pillar of European competitiveness
Europe has an opportunity to seize. It has a robust standards base that is a powerful lever for competitiveness and responsible innovation, provided that it is supported by targeted investment, shared standards and an industrial strategy aligned with our climate objectives. But Europe can rely on something even more decisive: its people. We have world-class researchers, visionary entrepreneurs, and thriving companies that embody the best of technological and industrial excellence. The recent strategic partnership between ASML, a key supplier to the world’s semiconductor industry, and Mistral, an AI start-up, illustrates Europe’s capacity to connect its industrial and digital strengths to shape a sovereign and sustainable future4.
It would be dangerous to suggest that Europe’s technological strength could be built on deferred ecology. What is tolerated as a gray area today will be a competitive handicap tomorrow. Customers, investors and citizens will increasingly demand transparency. The emergence of responsible AI does not mean making it perfect, but making it readable, controllable and adjustable.
In a technological landscape dominated by two superpowers that have hitherto favored efficiency and technological competitiveness to the detriment of ethical safeguards, Europe can chart a singular course. It has the means to assert itself by defending responsible AI, at the service of the common good and in line with its fundamental values: the rule of law, individual freedom, social justice and respect for the environment. This orientation is not a brake on innovation, but on the contrary a lever for differentiation, capable of inspiring confidence in a digital ecosystem that is often perceived as opaque or threatening. By betting on ethical, explainable and sustainable AI, Europe would not be giving up global competition, but it would be redefining the rules of the game. More than ever, it must give priority to clarity, stringency and rigor. Only then will AI cease to be a technological equation to be solved and become a genuine project at the service of our society, consistent with our democratic and ecological imperatives.
- AI & environment: breaking through the information fog – Sopra Steria
- “The 100 Most Influential People in AI 2024”, Time Magazine
- ADEME – Arcep study on the environmental footprint of digital technology in 2020, 2030 and 2025
- https://www.politico.eu/article/dutch-asml-invests-in-french-mistral-in-huge-european-ai-team-up/
Follow