Water is a key resource for data centers. At the same time, more and more areas are experiencing drought and other water supply pressures. Additionally, understanding that water use itself has an energy footprint is a complex issue.

In this installment of the roundtable discussion, panelists touch upon the importance of water analysis during site selection, cooling strategies that balance energy and water use, tactics to reduce water use during construction and more.

The roundtable discussion panelists included (from left to right in the video):

The roundtable series will be shared in four parts and will include segments covering energy, water, embodied carbon and what's next.

This discussion has been edited for clarity.

What's top of mind in the data center space for energy, water, embodied carbon—for transitioning the built environment into being a more sustainable and resilient resource?

Ryan: Continuing our discussion around some of the most important topics for creating more sustainable data centers. We're going to transition now to focus a little more on water. When we talk about water, we're talking about how data centers use it, the mass volume that is consumed, what we're doing to try to make that more efficient, and what goes into the actual construction process, because that can be water intensive as well. Lots of different aspects around water conservation that the data center space is using.

Greg: Yeah. It's a key piece of information that we need to look at when looking at site selections. Is water available? What type of water is available? Is reclaimed water available? Is potable water stressed? Is water in general stressed in that location? And that drives a decision for us about what type of cooling system we're going to use. But from an energy efficiency perspective, wherever water is not stressed, we do want to use it to improve energy efficiency as much as possible. Evaporating water drives down refrigeration energy cost and energy budget substantially.

Ryan: And emissions connected to that, right?

Greg: Yeah, resultant emissions of that energy consumption for chiller operation. And as we move into that liquid cooling future, and water temperatures potentially rise to serve that liquid cooling load, it's going to potentially reduce the need for refrigeration for that component of the load. Maybe we'll be able to do some of that stuff with evaporative techniques.

So water is seen as, you know, the big bad evil of consuming it in the industry. There's a lot of focus on it because I think there's been a lot of development in water stressed areas that maybe weren't water stressed at the time when the building was built, but have become water stressed since and resultantly there's a lot of focus on consumption. But in areas of the world, there are places that are not water constrained. For us, using water in those locations that are not water constrained has a significant energy advantage, but we are paying close attention to locations that are water constrained or indeed water reliability concerns exist and deploying dry solutions in those locations. But we're actively looking at water solutions and where we can get it in new site selection.

Josh: Yeah, I think again, starting with site selection, when you're going somewhere that's water stressed there's a different perspective than when you're going somewhere where there is adequate water supply. In the case where there is water supply and you're a relatively dry climate, you know, some of the most efficient data centers that have the lowest cooling energy are driven by evaporative water based solutions and that is possible and it's smart where you have available water supply to do it. I think as you kind of go up further in being concerned about water in certain areas or not having access to non-potable water, since it certainly isn't a potable use, layering in different technologies to really help still achieve the efficiencies that you want, while using less water.

I think it's also important to understand that energy has a water footprint too. So, by going water free, there still is a water footprint of that energy. There's a balance and it's not a simple issue. It's sensitive to the climate, it's sensitive to the water scarcity. I think that equation is changing. The climate is changing and water scarcity is changing—in some areas more rapidly than others. I think really understanding that, understanding where we're headed, climate-wise and water-wise, and how to make sure that we're making smart choices and investments that are going to last multiple decades to serve these critical functions is really important.

We've seen a lot of thoughtfulness in where we place data centers when we have that choice, for certain types of compute and storage, and when we have to put one here because of business needs or customer demand, how we best respond to that individual circumstance of: how humid, how hot, how available, how constrained water is.

Greg: Water is not without energy intensity either, the production of water to deliver it to a building does consume energy and when we do our total operational cost energy equation, we need to be considering the energy intensity of the water delivered and how that water was made. I mean, if it's a location in southern California, for example, it's going to have a higher energy intensity than a location in northern California.

Ryan: Yeah. And that's that early analysis that really takes collaboration from this team to drive. To think about where are we located, what are our resources, how do we break those down? And then how do we actually apply those to make a more efficient system? That's the beauty of how we actually work through design and specific situations.

Denis: If we take the question of water in the context of LEED, for example, we also have more or less technical initiatives to make or to pursue—such as landscaping—no irrigation, reclaiming water, and yes, each time you reclaim water, you want to reuse it. There is stringent regulation, health regulations attached to that, that leads you to have to treat it before reusing it. But I wanted to bring also LEED water credit type, low flow fixtures, irrigation, reclaiming water, that's also part of conserving water and using less potable for the building, for the data center.

Ryan: Yeah, xeriscaping is another strategy depending on where you're located in a climate zone. So continuously thinking about those cistern systems for that rain capture and reuse. We're even exploring a technology right now that has hydro panels, and it's basically generating water from these PV panels, that's bringing it from air molecules. So there's continuous development to think about how we actually use rainwater capture or capture from just air molecules. I mean, this is even being applied in desert areas.

Greg: We've been studying water treatment systems quite closely and investigating how to improve our use of water on existing sites as well. So we've been looking at techniques that take salts out of water for reuse and being able to reuse that on site without necessarily sending it to drain, which helps with the overall impact of reducing consumption.

Ryan: Yeah and to think about that, it translates into the construction process. If we have municipality restrictions on reclamation and reuse, and we're starting to continue to put in our own batch plants for concrete at big data center developments, because of the sheer volume that it has over multiple years, it deserves its own plant. In those situations, we're able to sometimes put in water reclamation centers for the concrete washout or the water that's actually being used for the mix, and using rainwater capture and filtering to have good new water streams that didn't come from potable water sources. That continues to go down in line of how the data center owners are thinking about their scope three.

I mean, part of that calculation is how the building gets built too, and not just the operational piece on the back end. I think it's an important subject to continue to think about where are those right situations where we can have those types of solutions in place? I think of one of those situations where we put one of those water reclamation centers in and saved 2 million gallons of potable water over the life cycle of that project. If we can continue to apply those in numerous situations, that's a big win.

Josh: I think we have a lot of ground to catch up on water, at least relative to energy. Energy historically has been priced relatively smartly. It's priced based on how much you use. It's priced based on the hour you use the most of it. If you're a larger user, you pay access fees to have that much use, you're an industrial rate class. It's a relatively sophisticated market for supply and demand in energy.

Water is historically largely underpriced by a significant amount. The risk of it not being there is not priced in—best use and where it's used. There are some different rates for agricultural customers and residential customers, but water is significantly undervalued relative to its scarcity and appropriate use. So I think that the higher level thinking we're seeing in data centers and really being smart about water, unfortunately still comes with the cost of not really getting the cost savings with it or having the right mechanisms to support that thinking for other customers too.

Greg: I think in some locations, water is too cheap, fundamentally, and I think you have a point that it needs to be priced according to its value in some ways. It needs to be priced according to drought conditions and respond to timing of what's going on where, so if you're in a period of drought, the pricing potentially would go up. Now that doesn't give stability to large consumers of water to make great decisions about how they run their cost models if they don't know what the price is going to be, but it would create an incentive to be more careful about how you use it.

Greg: We are driving to lower PUE numbers and to drive to those lower PUE numbers, we see a future where we are going to have to continue using water where it's not necessarily constrained. It reduces energy consumption for compute by enabling evaporation. It does have an impact, and we need to carefully consider where we use it, how we use it, and what technology we use it in.

To give a couple of examples: we consider whether a site should be based on air cool chillers, water cool chillers with cooling towers, water cool chillers with hybrid towers that can run dry in low temperature conditions. It really depends on the climate in that particular location as to which one makes sense. Some climate models make it attractive just to use open cooling towers, some models make it more attractive to use hybrid towers.

Ryan: So one of the important aspects here is to continuously talk about how data centers are making investments outside of what's happening on campus to drive change as well because of the mass amounts of infrastructure that it takes to feed these data halls and the valuable product that they produce on the back end.

Josh: Yeah, I think that after careful understanding of how scarce water is where you're locating and being really efficient with the water that you do use and making that choice carefully based on the energy efficiency, I think that they're also recognizing that there still is an impact from the water use, and there are some methods that they are taking—some companies are taking—to own that impact and try to make a difference.

Similar to renewable energy certificates that are available to support renewable energy markets, there's a similar product—not as well used or known—called water restoration certificates that are offered by the Bonneville Environmental Foundation, and I think there are some other places you can get them as well. It's not one for one, you're not buying a kilowatt hour of renewable energy and putting it on the grid. In this case, the money is going towards restoring streams and waterways that may have been impacted or degraded by development or by other forces that are happening in the world, or by investing in infrastructure that really helps save water in the supply and distribution. For example, often water's being taken from streams, lakes, rivers, where we get water besides where it falls from the sky, where it's really hot that a lot of water can be lost through evaporation. Some projects are investing in pipeways to reduce the evaporative losses to not lose as much to the atmosphere and better provide it to drinking, and other uses.

Again, just understanding the impact is the first thing. Being really efficient with it is the next thing. And then still taking responsibility to investing in and making it better for everyone else too.