Data centers are under pressure—Here’s what needs to change

Our industry is engaged in an important dialogue to improve the efficiency and resilience of real assets through transparency and industry collaboration. This article is a contribution to this larger conversation and does not necessarily reflect GRESB’s position.

Data centers keep everything running—emails, banking, streaming, AI models, and traffic lights. But the environmental cost is adding up fast. We’re talking about serious electricity use, heavy water consumption, and constant hardware turnover.

In the US, roughly 100 million square feet of new data centers were built between 2017 and 2023, according to MSCI, and PERE reports nearly USD 35 billion was raised for data center investments between 2019 and 2023—and that’s just in the US. The sector’s growing fast, and so are the emissions. That’s a problem for anyone with long-term interests in infrastructure, especially as regulation tightens and costs climb.

So, what can be done to mitigate this? First, let’s look at the problems data centers are facing today.

Energy demand’s out of control

Powering servers, cooling systems, and backup infrastructure all adds up. A single large data center can use as much electricity as a small town. And with digital services scaling by the day (AI, 5G, smart cities), demand’s not slowing down.

Water usage is becoming a red flag

Cooling systems can burn through huge volumes of water. This is becoming a growing liability—especially in water-stressed regions. Investors have started paying attention, and they’re right to.

Hardware churn means e-waste piles up

Short equipment cycles create a steady flow of e-waste—metal, plastic, and minerals often lost after just a few years in service. That’s a lot of material going to waste, or worse, into dodgy disposal pipelines.

Construction isn’t exactly clean either

The construction of data centers involves materials with substantial embodied carbon, such as reinforced concrete, glass, and steel, leaving a big carbon footprint before the first server’s even switched on.

Supply chains are shaky

Data center expansion relies on a robust supply chain. Shortages and price increases in semiconductors and other components have already delayed expansion plans.

And then there’s regulation

More governments are bringing in carbon caps, energy performance requirements, and reporting standards. The pressure is rising for data centers to adopt more sustainable practices.

To mitigate these risks, anyone investing in data centers needs to consider how to reduce operational demand, design for resilience, and how accountability and influence are handled by the data center operator.

Energy and water efficiency, day-to-day performance, and onsite optimization.

The quickest wins in data center sustainability come from lowering day-to-day demand—energy, water, cooling, and everything in between. This isn’t just about carbon. Lower consumption means lower bills, less exposure to regulatory pressure, and more control over asset performance.

Start with cooling. Traditional air-conditioning systems are high on energy use and low on efficiency. Replacing them with newer tech (like liquid or immersion cooling) cuts electricity use significantly. In the right climate, free cooling (pulling in outside air) can take the load off entirely. Add in design tweaks like hot and cold aisle containment, and you’ve got a much leaner system.

The hardware matters too. Using low-power servers, energy-efficient storage, and virtualizing workloads means you can get the same output with fewer machines running. It’s also worth looking at how those machines are managed. Power management systems that automatically shut down idle hardware or adjust consumption based on load aren’t new, but they’re still underused.

Then there’s visibility. Data Center Infrastructure Management (DCIM) tools give operators a real-time view of energy use across the site. When combined with Building Automation Systems (BAS), they can expose inefficiencies and point to quick fixes. Artificial intelligence can be used to predict cooling loads, schedule maintenance, and smooth out usage peaks before they hit the grid.

Water’s part of the equation, too. Older evaporative cooling systems draw giant volumes of water. Air-based and closed-loop liquid systems are far less demanding. Where water use can’t be avoided, recycling systems help—but only if they’re properly monitored. Tracking Water Usage Effectiveness (WUE) is quickly becoming as standard as tracking energy.

These are all measures that reduce how much the data center consumes, which directly ties to OPEX, regulation compliance, and climate risk—all of which matter to investors.

Broader infrastructure choices and how the asset interacts with its environment—physically and economically.

Reducing operational demand is essential, but it’s not the full story. Decisions made in design, procurement, and energy sourcing will shape an asset’s long-term performance and market viability. That’s where resilience comes in.

Clean energy use is one of the most direct levers operators can pull. Long-term power purchase agreements (PPAs) offer a stable way to access renewable electricity, while on-site solar and wind installations are becoming more common for direct supply. But without battery storage, it’s all too easy to fall back on fossil fuel backups when supply dips. Storage systems allow operators to smooth peaks and make use of renewable power when it’s needed most, not just when the sun’s shining.

Waste heat is another underused opportunity. Capturing heat from servers and using it to warm nearby buildings or generate additional electricity through combined heat and power (CHP) systems can significantly improve overall efficiency. Some operators are already plugging into local district heating schemes. It won’t work everywhere, but where it does, it’s a no-brainer.

Hardware needs to be considered from a long-term perspective too. With typical equipment lifespans sitting around 3-5 years, e-waste builds up fast, so circular thinking needs to be part of the strategy. It starts with smarter procurement: buying less, refurbishing more, and keeping gear in use longer.

Cloud computing can be part of a long-term sustainability play—but only if it’s approached carefully and with the right partners. Migrating workloads can reduce IT footprint, but not all cloud providers are equal—check how they source energy, manage water, and handle hardware lifecycle before moving anything off-premise.

But resilience isn’t just about how a data center performs; it’s also about whether it gets built at all. In many markets, big barriers to development include zoning and tenant demand and, most pressingly, access to power. Grid constraints are becoming a critical risk, with some projects facing multi-year delays just to secure electricity. Add in growing community resistance, particularly around water use and job creation, and the path to delivery is looking tougher. For investors, this means factoring in utility risk and reputational exposure right from the site selection stage.

These efforts focus on longer-term strategy and risk resilience, guiding decisions that position a data center as a better long-term investment. These aren’t daily optimizations—they’re structural choices that can define a site’s future value.

Transparency, market collaboration, and influence on policy.

Data centers are under increasing scrutiny from investors, regulators, and the markets they serve. Just as with any other asset class, they are expected to demonstrate performance, plan ahead, and work collaboratively with the wider system.

Operators leading the charge aren’t waiting for policy to catch up. They’re engaging with policymakers, shaping future regulations, and making sure their sites are ready before the rules land. They’re also working closely with grid operators and utilities to align on supply, manage demand, and avoid getting caught out by price spikes or capacity limits.

Beyond compliance, there’s a growing push to quantify impact in a way that investors can actually use. That means contributing to shared standards for measuring embodied carbon and circularity—and building transparent reporting into everyday operations. Joining open-source efforts helps accelerate this work while showing that the operator is willing to be held accountable.

Because in the end, sustainability isn’t just about what you say you’re doing. It’s about what you can prove and how well that lines up with your investors’ risk models.

Sustainability in data centers comes with real challenges—but they’re not unsolvable. Built into the business strategy, they become part of the value story. Ignored, they become a liability. These assets are no less exposed to regulatory, operational, and reputational risk than any other part of a portfolio.

Whether you own the asset or back the operator, this isn’t something to bury in the ESG appendix. It belongs on the board agenda.

So, where are the pressure points in your portfolio? And more importantly, what’s your plan to fix them?

This article was written by Maja Christenson, Marketing Manager at EVORA. Learn more about EVORA here.