From Critical Insight to Concrete Action: Mitigating the Environmental Impact of Data Storage (3/4)

Over the past two blogs, we explored how essential data storage is to building trustworthy AI, and how this rapidly growing storage demand is driving a mounting ecological footprint. While the first blog highlighted data as the oxygen of AI, the second exposed that this oxygen comes at a high cost: electricity consumption, water use, and CO₂ emissions.

This brings us to a pressing question: what is actually being done to mitigate this environmental impact? In this blog, we dive into concrete mitigation initiatives, evaluate their effectiveness, and in a follow-up post, we will discuss some critical perspectives on these efforts.

1. Switching to Renewable Energy Sources
   • Major tech players like Google, AWS, Microsoft, and Meta are investing heavily in solar and wind energy to make their data centers more environmental friendly. This includes Power Purchase Agreements (PPAs) that secure long-term access to green electricity. Some data centers —like Google’s since 2017— even run entirely on renewable energy on a net basis.
   • Example: Microsoft aims to become fully carbon-negative by 2030, including emissions from its data center operations (Microsoft Sustainability).
2. Cooling Innovation
   • To curb energy use, data centers are adopting more efficient cooling methods. Liquid cooling is replacing traditional air cooling and delivers better performance with less energy. Even more efficient is immersion cooling, where servers are fully submerged in non-conductive fluids, cutting cooling energy use by over 90%.
   • Location matters too: data centers in colder climates (e.g., Iceland, Scandinavia) take advantage of natural air cooling.
     • Meta’s data center in Luleå, Sweden uses outside air and runs fully on hydroelectric power.
     • The EcoDataCenter in Falun (Sweden) supplies its residual heat to a nearby wood-processing facility, boosting mutual energy efficiency.
3. Smarter Storage = Lower Impact
   • Storage optimization is critical — it reduces not only costs, but also energy consumption and carbon emissions. Less data means less infrastructure, less cooling, and a smaller environmental footprint. Key examples include:
     • Advanced compression (e.g., AV1 codec for video) cuts file sizes without quality loss.
     • Data tiering automatically moves rarely used “cold” data to energy-efficient storage layers.
     • Computational storage brings processing closer to the data itself, lowering data traffic and energy use.
4. AI for Energy Optimization
   • AI plays an increasingly important role in optimizing data center operations:
     • Real-time management of cooling, server load, and power consumption.
     • Predictive maintenance that prevents unnecessary energy loss by detecting failures early.
     • Example: Google DeepMind reduced cooling energy consumption by up to 40% through AI-driven controls.
5. European Use Cases
   • Several advanced edge and hybrid data centers across Europe illustrate these mitigation strategies in practice:
     • Stack Infrastructure, Copenhagen (Denmark): designed for ultra-low latency and fast data access. Part of a broader European edge strategy focused on modular architecture, energy efficiency, and heat recovery.
     • EcoDataCenter, Falun (Sweden): although primarily hyperscale, it includes edge capabilities for local industry. Operates fully on renewable energy and reuses residual heat.
     • Maincubes, Frankfurt (Germany): offers colocated solutions with edge support, low-latency performance, and green power contracts.
     • GTT Communications, Paris (France): supports edge workloads from metropolitan locations near enterprise hubs, connected to fiber backbones.

Modular & Scalable Architecture

Modular infrastructures only activate what is needed, avoiding waste and boosting efficiency. Edge data centers reduce network latency and energy consumption by bringing compute power closer to users.

Certification & Regulation

Certification and regulation are key building blocks of sustainable data infrastructure. ISO 50001 was created in response to rising energy prices, climate concerns, and the need for harmonized energy management standards. It is now globally recognized and forms the backbone of many carbon reduction and sustainability programs.

Data centers are increasingly assessed on sustainability using standards such as:

• ISO 50001 (energy management)
• LEED (Leadership in Energy and Environmental Design)
• BREEAM (Building Research Establishment Environmental Assessment Method)

At the European level, initiatives like the EU Green Deal and the EU Code of Conduct for Data Centres drive greater transparency and energy efficiency across the sector.

In summary, we can conclude that the sustainability transition of data centers is underway, but not without legitimate concerns. Questions around greenwashing, rebound effects, and transparency are becoming increasingly pressing. In the next blog, we’ll examine how credible these initiatives truly are, and what structural choices are needed to make data storage genuinely sustainable.