With the rise of generative artificial intelligence, data centers—which are under ever-increasing strain—must, now more than ever, meet a dual imperative. On the one hand, they must reduce their electricity consumption to contain costs and lower their carbon footprint. On the other hand, they must strengthen their security in the face of an unprecedented surge in cyber threats. In other words, digital infrastructure can no longer simply be high-performing—it must also be resilient and sustainable.

“Cybersecurity, often perceived as at odds with sustainability, can in fact be compatible with it,” emphasizes Agnès Comte, a sustainable IT specialist at the Banque de France. “There are numerous synergies: digital sovereignty, control over information systems, life cycle management, continuous improvement, risk analysis, interoperability…”

Finding the Right Balance

However, tensions remain and necessitate certain trade-offs. For example, redundancy in the cloud—which is crucial for ensuring availability—can result in a significant carbon footprint. In addition, regulatory requirements regarding backups or data compartmentalization can increase data flows and the number of data copies.

“Balancing operational constraints, budgetary requirements, and environmental goals remains a delicate task,” acknowledges Agnès Comte. To find the right balance, “we need to ask ourselves what we want to protect, the scenarios we fear, the type of intra- or interregional redundancy, the resources already available within the company, and the potential added value for customers,” she says.

With this in mind, operators are seeking to accelerate their transition to a “green cloud.” This involves, in particular, innovative cooling solutions, such as liquid cooling, which “significantly reduces a data center’s PUE (Power Usage Effectiveness),” notes Didier Ronsen, Managing Director of Stulz France.

GreenOps and FinOps Brought Together

Artificial intelligence algorithms also help optimize thermal management. Using data on temperature, airflow, and humidity, they automatically adjust the various cooling systems, resulting in significant reductions in electricity consumption.

This GreenOps approach is also based on eco-design in software: more efficient code and algorithms, data deduplication, automatic resource standby mode… These are all strategies that help limit unnecessary processing. Finally, cloud platforms rely on increased use of renewable and carbon-free energy.

Finally, this approach is fully aligned with a FinOps strategy aimed at cost optimization: reducing energy consumption automatically lowers the bill, especially given the current significant pressure on electricity prices.