Modern data center sustainability: 5 best practices to consider
Sustainable data centers are needed as energy demands rise. CIOs should consider the following best practices to reduce costs, manage resources and minimize their carbon footprint.
Sustainability is becoming increasingly important for organizations, particularly in the data center industry, where energy needs are rising due to AI and high-performance computing. As data center managers and CIOs face increasing energy costs and regulatory challenges, reevaluating sustainability strategies is crucial.
This article highlights the current state of data center sustainability, why it matters and best practices to help organizations navigate these challenges while upholding their commitment to environmental responsibility.
Why sustainability matters
Enterprises recognize that sustainability remains important, but they remain uncertain about how easy it will be to achieve, as AI data centers now compete for energy resources. Grid experts know when a new data center comes online because they see a noticeable energy spike in their distribution system. These energy spikes then become the new norm for power use as the data center continues to draw from the grid.
Straining the grid has increased energy costs, particularly in data center hubs like Virginia. AI workload demand will continue to raise energy costs if data centers keep straining the grid. Sustainable energy methods are more crucial than ever to integrate into data center infrastructure to achieve energy independence, manage energy costs and shrink the carbon footprint.
What sustainability means for the data center
Modern data centers are equipped with sustainability tools, such as power usage efficiency (PUE) monitors and water usage efficiency (WUE) systems. These tools track energy and water consumption and offer live statistics.
All are impressive achievements in data center sustainability made in recent years. However, as large AI data centers across the country intensify the competition for energy, prices will rise regardless of how well they are managed. Corporate data center managers will continue to see sustainable improvements in reducing their carbon footprint, but the accessibility of affordable energy will diminish.
While this is happening, the definition of a data center is evolving as edge computing rises. Currently, the data center is responsible for centralized IT assets as well as the growing number of micro data centers and networks appearing at enterprise edge locations. Most of these edge micro data centers support end-user and IoT devices and sensors.
In 2022, McKinsey reported, "End-user devices -- laptops, tablets, smartphones, and printers -- generate 1.5 to 2.0 times more carbon globally than data centers." The report also stated that "Emissions from end-user devices are on track to increase at a CAGR (compound annual growth rate) of 12.8 percent per year."
Then, in 2025, Statista reported that globally there were 8.62 billion mobile phones, 1.73 billion laptops and PCs, 17.4 billion short-range IoT devices and 4.93 billion wide-area IoT devices. Not all these devices are managed by enterprises, but many are -- and under the emerging definition of the distributed, enterprise-wide data center, they fall under the data center's sustainability umbrella.
Sustainability practices for the modern data center
For CIOs and data center managers, sustainability has always had two main objectives: greening the data center and saving money. However, as technology evolves and the demands on IT infrastructure grow, sustainability practices must also adapt to meet new challenges and opportunities. Below are five best practices to consider.
1. Redefine the data center and what data center sustainability means
Most CEOs, boards and non-IT C-level executives still view the data center as a centralized facility, but this perception is increasingly outdated. The modern data center is no longer confined to a single location -- it now encompasses distributed IT environments, including enterprise edges, cloud platforms and hybrid infrastructures.
CIOs should lead efforts to educate stakeholders about this shift, highlighting how distributed IT can improve operational efficiency and sustainability. By expanding the definition of a data center, organizations can better match their sustainability goals with the realities of modern IT, such as partnering with cloud providers committed to green energy or optimizing edge computing to reduce latency and energy use.
2. Keep doing what you're doing
As the concept of the data center evolves, the fundamentals of sustainability remain essential. CIOs and data center managers must continue to focus on optimizing equipment, cooling systems and floor plans to achieve maximum performance and energy efficiency. This includes regularly monitoring metrics like PUE and WUE to identify inefficiencies.
Additionally, replacing aging hardware with newer, more energy-efficient options can decrease energy use and enhance overall system reliability. Regular audits and proactive maintenance can also keep sustainability efforts on track, even as the data center environment evolves.
3. Make HVAC, cooling and facilities planning an integral part of IT
In many organizations, facilities management operates independently of IT. Facility management is responsible for procuring and installing generators, HVAC equipment, cooling systems and related equipment. They also work with vendors to determine where to install new equipment. IT teams should develop a close, collaborative relationship with the facilities group, working with them on any issues that could impact the central data center or any edge micro data center.
By integrating facilities planning into IT strategy, organizations can create a more cohesive sustainability approach. This includes selecting equipment and deciding where to install it. For example, it can significantly reduce data center energy costs by installing HVAC cooling near the largest hotspots on the data center floor. Collaborative planning can also help identify opportunities to implement advanced cooling technologies, such as liquid or free-air cooling, to further enhance energy efficiency.
4. Collaborate with Citizen IT and edge locations
The rise of Citizen IT, where end users procure their own IT equipment and software, has created challenges for centralized IT management, particularly as edge computing becomes more prevalent. With the acceleration of edge computing, IT can't afford to be hands-off for the edge equipment that it will ultimately be responsible for. By fostering collaboration with Citizen IT, organizations can maintain control over edge environments while advancing their sustainability objectives.
CIOs must work closely with end-user departments and central purchasing teams to ensure that energy-efficient hardware and software are selected for edge locations. This includes setting standards for mobile devices, IoT equipment and other edge technologies to ensure they align with the organization's sustainability goals. IT teams should also take responsibility for monitoring and managing these assets throughout their lifecycle, including proper disposal or recycling at the end of their useful life.
5. Prepare for the impact of AI
AI is a major driver of energy consumption in data centers, yet many organizations remain unprepared for its impact. CIOs and data center managers should begin assessing where AI workloads are likely to be deployed, whether on-premises or in the cloud, and estimate the associated energy requirements. This includes evaluating the energy efficiency of AI hardware, such as GPUs and TPUs, and exploring options for optimizing AI workloads to minimize energy use.
Organizations should consider adopting AI-driven tools to improve data center operations, such as predictive maintenance and dynamic workload management, which can help offset the energy demands of AI. By planning for AI's energy impact now, CIOs can ensure that their data centers remain sustainable as AI adoption grows.
Questions to consider when preparing for AI's impact
In many cases, CIOs and data center managers can't yet answer the following questions, but they should start strategizing them now.
• Where in the enterprise is AI most likely to be used, and how?
• Will you be hosting AI internally or on the cloud?
• In either case, is it possible for you to predict what energy consumption will be for AI?
Mary E. Shacklett is president of Transworld Data, a technology analytics, market research and consulting firm.
