Zero water in our data centers by 2030: mission impossible?
Published in Data Center Water, Environment.
On June 3rd I participated in a panel discussion on the topic of data center water organised by Scaleway. I was one of four participants, with the other three being Marie-Laure Vercambre – Director General of the French Water Partnership, Arnaud de Bermingham – President & Founder – Scaleway, and Manuel Mateo Goyet – Deputy Head of the Cloud and Software unit at the European Commission, DG Connect.
Panel discussions are often boring, but I like the format where everyone gets 5-10 minutes to make a few remarks before going into questions from the audience. With panel members from several stakeholder groups (industry, government and non-profit advocacy), this ended up being a good discussion.
I’m publishing my prepared remarks below (with some links added for sources/further reading), and you can watch the full discussion on YouTube.
My remarks on data center water consumption
When we think about sustainable data center and the environmental impact of IT, most people start with energy. A lot of progress has been made over the last decade, both in terms of the transition to renewable electricity generally, but also specifically within the IT industry.
Large data center operators like Amazon, Google, and Microsoft, consistently hit the top 10 list of renewable energy buyers each year. There are still challenges about what “100% renewables” actually means – most data center operators claiming that today are actually just greenwashing if they’re only using RECs – but the momentum is there.
The same cannot be said for water consumption, which I would say is a decade behind energy in terms of understanding, transparency, and progress. However, we have to be careful when comparing resources like energy and water because the sustainability goals are not the same.
With energy, the goal is zero carbon, but zero water is not the right approach. Data center water consumption is highly dependent on the region because that dictates which cooling technologies can be used. In cooler regions, zero water can be achieved, and should be the goal, because free air flow cooling is possible for some, if not all, of the year. But in many hot regions, this is simply not viable. That’s important to understand because not only is a large part of future demand growth going to come from hot regions like Africa, South America, India and Asia, but global temperatures are going to increase due to global warming in general. Regions that can currently use free cooling may find that it becomes more challenging in the future.
It is also important to understand the general context for water consumption, because direct consumption from data centers is only a small part of global water consumption. For example in the US, 1.7 billion litres per day is directly consumed by data centers, but when you consider that total water consumption is over 1200 billion litres per day, the data center footprint is insignificant.
However, that only tells part of the story because most of that focus is on direct consumption. Over half of the daily consumption, some 500 billion litres per day in the US, supports thermoelectric power consumption. Power plants generate heat using fossil fuels such as coal and gas, or nuclear fission, to convert water into steam which rotates a turbine, thereby generating electricity.
With the majority of generation still from fossil fuels, the transition to renewables is important for both carbon and water intensity. Only solar and wind energy do not involve water in generation. Estimates suggest that by 2030, moving to wind and solar energy could reduce water withdrawals by 50% in the UK, 25% in the USA, Germany and Australia and 10% in India.
This means that the true data center water footprint is linked to the source of energy, but currently this is not considered the responsibility of data center operators, and so it doesn’t get reported.
And reporting is really limited. Data center operators are familiar with the Power Usage Effectiveness ratio which is often used as a measure of efficiency. A similar metric, Water Usage Effectiveness, exists, to help understand annual site water usage. Unfortunately, less than a third of data center operators track any water metrics and water consumption is ranked as low priority by most. Facebook is one of the few companies operating on a global basis who report WUE, although Google and Microsoft both publish total water consumption. Unfortunately, you can’t actually use Facebook’s infrastructure in the same way you can with Google and Microsoft – it’s all for their own applications. I note that Scaleway does also publish WUE for its 6 data centers.
Understanding the source of water is also necessary context to analyze any water metrics. Some would argue that where water is used, the goal should not be zero water, but zero freshwater. If water is recycled or reclaimed, its impact is not the same as if all the water is drawn from municipal or freshwater sources. However, this is not revealed by consumption metrics or WUE. The relative water stress of a region is also not represented in these figures. Consumption numbers in a region with easy access to abundant cool water take on a very different meaning if the same numbers represented consumption in a highly water stressed region.
All this shows how nuanced the discussion around data center water consumption is, which explains why operators have so far been reluctant to reveal much. It’s far easier for a layperson to understand zero carbon compared to what is needed to understand the context of water consumption. This means much of the media coverage around data center water focuses on highly emotive topics like access to drinking water and pollution from water treatment. It seems the current approach is to avoid publishing any numbers at all, so there is nothing to focus on or report.
This reluctance is going to have to change though, because the full environmental impact of IT is getting more and more attention. Mainstream media is starting to report on water consumption, particularly where data centers are the largest users in a particular region – this is something that has been in the news recently in Holland, and local residents groups in the US have made similar complaints in recent years. Sometimes this is about entirely new projects, but expansion of existing sites has also been challenged as they grow to meet the ever expanding growth of IT.
Now is the window of opportunity for data center operators. Regulation is still limited, particularly around disclosures, but that is going to change. Data center operators need to get ahead of reporting requirements by ensuring they have reliable metering and a plan for water efficiency that can be implemented over the coming years in conjunction with renewable energy projects. Renewable energy is the model to copy because these projects are seen as positive news stories, and data center owners have taken advantage of that to show off their sustainability credentials. The same could be done for water. Facebook talks a lot about its restoration projects for local watersheds and ecosystems, and Google’s Finland data center is quite famous for being located in the snowy north, having taken over an old mill so it can use the cold lake water.
If nothing is done then mainstream pressure will grow, and that is what politicians respond to. I’m sure data center operators would prefer to deal with this challenge on their own terms, but if they don’t, then politics will do it for them. That will probably not result in favourable outcomes.