Building the data centers necessary for AI and emerging technology shouldn’t require sacrificing environmental and climate goals.


In brief
  • Data centers have significant sustainability impacts that require upfront strategy to mitigate impacts.
  • Smart operators and hyperscalers will take time to fully understand the risks and array of viable solutions in design, construction and operation of data centers.
  • A partnership mindset can help those across the data center ecosystem identify and implement innovative solutions across the full value chain. 

This article is co-authored by Shannon Roberts, Principal, Climate Change and Sustainability Services, Ernst & Young LLP; Kristina Albang, Managing Director, Ernst & Young LLP and Reagan Richmond, Senior Manager, Climate Change and Sustainability Services, Ernst & Young LLP.

As hyperscalers and other developers rush to build data centers to manage the rapid growth of artificial intelligence (AI), cloud computing and storage needs, they face a significant challenge — supporting a new era of digital productivity that requires significant energy and materials in a sustainable manner.

In a recent EY survey of 500 senior leaders across all sectors, 64% said they are concerned about the negative impact of increased AI usage on their organization’s sustainability (or emissions) goals.

Many of the world’s largest technology companies were early adopters of sustainability as a business imperative. To demonstrate their commitment, they unveiled detailed net zero strategies to stakeholders, pledging to reduce or eliminate carbon emissions in their operations and across value chains. Their data center customers — other Fortune 500 companies — now also share those same objectives and commitments.

 

A decade or so ago, however, few could predict the extent to which AI and the growth in data would create conflict with the desire to achieve a carbon-free future. Technology firms are rethinking their baseline emissions commitments given the energy consumption requirements of new or expanded data centers. Their customers, too, will quickly recognize, if they haven’t already, that their sustainability goals could be jeopardized by scope 3 emissions from the data centers they utilize.

Continuously evolving uses of AI are critical to the digital revolution’s transformation of business. But how can the resulting data centers be built in a manner that allows business, people and the planet to thrive?

1

Chapter 1

Data centers’ impact on the environment

Data center developers and operators must manage a host of potential sustainability and social impacts and risks across the entire value chain.

From construction through ongoing operations, data centers require significant resources and can potentially alter the surrounding community. Their sustainability impacts are vast, and those responsible include participants across the ecosystem.

  • The construction and expansion of data centers can lead to land use changes, potentially impacting local ecosystems and biodiversity. Construction itself can be an issue, increasing traffic and noise and creating issues for local communities. Additionally, construction materials such as concrete and steel may have high embodied carbon.
  • Data centers require significant amounts of electricity and water to power and cool high-velocity computing equipment. Utilities are currently adjusting to the sudden need to provide the large power loads that data centers require while also considering their own carbon reduction targets. Therefore, as data centers seek inexpensive power that is readily available, it is not always possible to procure renewable electricity — which cannot be quickly built due to permitting and interconnection issues. And while some regions may be attractive due to the availability of solar energy, those areas are sometimes dealing with water scarcity issues.
  • Data centers generate sizable levels of electronic waste. Servers, routers, switches, hard drives and other technology components break down often under intense usage or become obsolete quickly due to advancements. Often, that waste can’t be processed sustainably due to volume or lack of local facilities.
  • Data centers face a variety of social concerns for nearby communities. During construction, data centers have community and work impacts during rapid construction. The size of the land used by data centers is comparable to a distribution warehouse, yet data centers create relatively few jobs. In fact, data centers typically employ around 50 long-term employees after construction. Noise is another concern, as data centers’ servers, cooling systems and backup generators often produce a constant humming noise, which can be a nuisance to nearby residents.

Data center development requires a “partnership mindset,” because the environmental and regulatory demands and risks can’t be managed alone. Integrated strategy and collaboration along the value chain are increasingly important elements in development. Further, underestimating the ESG challenges inherent in data center infrastructure will slow development far more than taking the time to fully understand where and how obstacles exist.

2

Chapter 2

The path to a sustainable data center future

By strategically addressing the ESG impacts of data center development and operation, hyperscalers and operators can better position themselves and data center customers.

Today, some data center developers are taking a “bolt-on” approach to sustainability rather than intentionally integrating it into the initial design and operating strategy. The developers that create a lower-carbon, high-sustainability strategy for planning, constructing and managing data centers will win out in a competitive marketplace as companies seek to stay on track with their net zero commitments.

What can smart developers do to grow capacity sustainably? Here are four key steps to consider:

Step 1: Understand the risks — and benefits — inherent in site selection

Taking the time to conduct detailed landscape surveys and site analyses will pay dividends later. Understanding the local utility’s grid energy mix and potential for load expansion is critical, as is predicting future water availability.

There are other potential risks that should be evaluated. How willing are local and state governments to provide timely approvals? How eager is the community for economic development?

Also, how resilient is the area in withstanding natural disasters? Companies are increasingly conducting climate risk analysis to understand the most beneficial placement of their data centers and prepare for any potential physical and transition risks.

Step 2: Be proactive and creative with partners across the ecosystem

Power and utility companies are eager to capture increased load from data centers, but they often need assistance to meet higher demand, especially for clean energy. Can you partner on joint investments in clean infrastructure, sharing the capital expense and potentially speeding implementation?

Beyond the local utility, is there potential to partner with a renewable energy developer to solarize the roof? This approach can create co-benefits in cost sharing and may provide tax equity partnership benefits. There may be other state or local tax incentives for green building/on-site generation of renewable energy such as solar, wind and hydro power. Some are even seeking cooler climates to support energy reduction and natural cooling.

Many hyperscalers prefer shorter-term power purchase agreements (PPAs) with local utilities. But the utilities themselves would prefer longer-term PPAs that allow them to recapture the cost of building generation and transmission more completely. Can you think differently about your contracting approach to create opportunities for clean power?

Step 3: Measure impacts on an ongoing basis and engage with stakeholders

Assessing the environmental and community impacts of a data center is not a “one and done” effort.

Data center hyperscalers and operators will need to understand the regulatory reporting requirements in their areas as well as what their partners and suppliers will require for their own sustainability goals.

Many multinational companies are subject to increasing regulations that require ongoing reporting of greenhouse gas emissions, energy sources and water usage. Even companies that aren’t subject to global regulations must comply with state and federal rules and will be asked for data by their customers.

Community impacts could persist long after construction is completed. Ongoing efforts to monitor and respond to concerns is critical.

Step 4: Identify and deploy innovative solutions

Using state-of-the-art design, construction and operating technologies and approaches can provide levers to reduce carbon.

It starts with procurement of materials. During the build phase, Engineering, Procurement and Construction companies should procure materials to minimize carbon by seeking low embodied carbon in cement and steel. There are also opportunities to design facilities with high energy efficiency design and equipment to minimize electricity use when in operation.

For example, optimized building layouts and high-efficiency cooling mechanisms can reduce HVAC usage along with power consumption. Hot aisle/cold aisle containment, where the exhaust from computing hardware is physically separated from the cold air intake, is one method for improving efficiency. Another is air-side economizers, which utilize cold air from outside the building when temperatures are low.

Data centers can also match non-urgent computing activities to hours when access to clean energy is available, or, if needed, to off-peak hours when the grid is less stressed. Energy-efficient data analytics, components and hardware can also limit power usage.

 In areas where water is an issue, a rainwater capture and circulation system can help reduce demand, as can gray water technologies to reuse and recycle water.

Ultimately, mapping the interaction of different environmental impacts and benefits can lead to solutions and partnerships that create innovative solutions. For example, industrial co-location can allow for creative reuse of gray water or local processing of obsolete or damaged hardware.

Data center developers/operators should consider partnering with private equity/venture capital firms to invest in clean-tech solutions, driving more sustainable outcomes. Over the last three years, PE firms have invested more than US$100b in data center projects. Opportunities will exist across the value chain — from data centers themselves, to the software and hardware vendors that enable them, to the enormous amounts of power needed to run them.

Finally, staying abreast of new developments in technology, equipment and processes can deliver sustainability gains. For example, recent advances in system architecture, optimization and quantization can reduce the energy required to run models by efficiently distributing workloads.

3

Chapter 3

Data center teams must act today

By utilizing existing strategies and technologies and being creative with partners, data center operators and hyperscalers can grow sustainably.

Inevitably, customers will demand that hyperscalers effectively manage data center sustainability, primarily because tremendous power demands will drive significant scope 3 emissions that must be reported. And other environmental and social challenges will play a role, too.

The hyperscalers and developers/operators that can offer a lower emissions profile will benefit over the long term. Overlooking sustainability creates risks and leaves potential business and societal value on the table.

The good news is that utilizing existing sustainability technologies and strategies and being creative with partners can make a difference.

Summary 

Those involved in data center growth, across multiple industries, can implement viable, real-word sustainability measures to create cleaner platforms, which are critical to support AI, emerging technology and more in the digital future.

Our latest thinking

Top 10 opportunities for technology companies in 2025

Technology companies are at a pivot point and transformation is at their fingertips if they act on these top 10 opportunity recommendations from EY in 2025.

18 Dec 2024 James Brundage + 2

How to secure data centers from internal and external risks

Data center security is key; firms must tighten internal policies to shield sensitive info. Discover more.

21 Nov 2024 Bill Fryberger

Data center development: mastering the stabilized asset cycle

Avoid pitfalls and unlock value through a multidisciplinary, three-stage strategy. Learn more.

14 Oct 2024 Gordon Bell + 2

Energy transition

Discover energy transition insights, people and services and how our teams can help your business unleash sustainable value by embracing the evolution of energy.

18 Jun 2024

Why data center industry operators need to keep up with growing AI demand

Data center industry participants can develop new strategies to address the unique needs of AI as the tech goes mainstream and demand grows. Read more.

04 Apr 2023 Gordon Bell + 2
    You are visiting EY us (en)
    us en