The role of water resources for data centers and industrial users

Whether a lack of them or difficulty managing them, resources lie at the heart of slow progress on water utility digitization. Such was the message from water industry panelists during a discussion at the Bentley Year In Infrastructure event in Amsterdam Oct. 15.

Water infrastructure is aging and financial resources have not kept up with the needed investments to repair or replace those assets. Meanwhile, operational costs continue to rise, pointing again to financial resource constraints, but also to chemicals and the all-important resource, energy. On top of that, human resources are also beginning to dwindle.

“Qualified, experienced people in utilities are retiring,” said Cloe Meyer, senior research director for Bluefield Research, “and it’s becoming a lot harder for utilities to recruit new talent and also retain them.”

She said those two things, recruitment and retention, also bring up another resource issue: a shortage of technical capacity and knowledge. Nuno Medeiros, head of asset management at EPAL – Water Supply Company of Lisbon, echoed that particular statement while highlighting the tremendous degree of change for EPAL in just the past two decades.

In that time, EPAL has adopted and implemented a SCADA system, GIS mapping of its assets and integration of a CRM platform for customer support. At each stage of that process, technical skills and abilities were tested as current employees were trained on each new tool. That technological knowledge at each point became its own silo. But that didn’t deter the utility from moving forward digitally, which presented another issue entirely.

“We are applying more and more sensors in our network, more and more sensors in our dams, in our watershed,” Medeiros said, “so we have more and more data.”

That data, he said, doesn’t always turn into information for action at the utility level, and there is simly a lot of it to manage. More often however, water systems have a lack of data and information, especially when trying to work with new developments such as data centers.

Mismatched timelines bottleneck data center siting

Rod Naylor, global water lead for GHD, noted the different time horizons for data center companies and utilities as a fundamental headwind for acceleration of the AI sector.

“The water industry is slow. Things take decades. The data industry is not. Things wear out long before the decade is over,” Naylor said. “This mismatch is one of the challenges for that collaboration between the water industry and tech companies.”

Ben Townsend, head of infrastructure and sustainability for Google, said that major tech companies like Google are interested in partnering with utilities, including offering large monetary investments to upgrade the public water infrastructure.

He said utilities often have a lack of data that would inform decisions on where to site a data center or similar facility. Utilities do not always have the models for water supply of surface and groundwater to understand their long-term capabilities.

“We’ve often found the public utilities have not been given the tools or resources to truly understand the nature of their infrastructure system or the watersheds themselves,” Townsend said.

He relayed a case of a utility offering water supply to locate a data center in its service area, but Google’s own research suggested the utility could not sustain such an investment in water. As such, he said Google had to turn down the location only to later learn another facility aimed to locate there instead.

These kinds of matters are as much about the technical details as they are the political and governmental will. City councils and county boards want data centers for economic growth and local jobs, while utilities want the financial assistance for the upgrades to meet the needs. Neither of those interests matter if the industrial user sees only short-term potential for its business due to non-resilient water supply.

This has led companies like Google and Amazon to invest in leak detection programs, which ultimately aim to save enough nonrevenue water to create the supply necessary for a data center without negatively impacting communities.

“Here in the Netherlands, we’ve invested in municipal leak detection and repair programs. We’ve invested in rainwater harvesting and capturing that rainwater through aquifer storage and recovery projects to help alleviate stress within the community,” Townsend said.

Meyer and Naylor also pointed to wastewater reclamation and water reuse as options utilities have looked to as a solution, but that isn't always feasible.

‘This’ is what water utilities can do?

Townsend said water utilities have generally had two mandates: produce clean water as cheaply as possible, and spend on upgrades exactly when they are necessary. The freedom to experiment, to try something new or to invest in untested ideas is does not exist and hampers the water sector’s ability to be as nimble as the data industry.

An audience member asked the panel for concrete answers about what utilities could do, that utilities "can do this” to move the needle on progress.

“Water authorities, beyond the minimum requirements, are not clear what ‘this’ is,” Naylor said. “And then they’re not allowed or enabled because of regulation. What I’m seeing around the world is a bit of an evolution of the ambition and the manturity of the water industry … I think there’s a shift toward understanding purpose.”

He said that means utilities are getting better at understanding consumer demands and needs, which allows them to calculate the full value cycle of water, not just the asset's life cycle.

Answering the same question, Townsend said transparency from both the industrial user on exact water needs as well as transparency from the utility on what it can provide are most critical. As for action, he said utilities should also prepare information to take advantage of the spending from tech giants like Google.

“Utilities can blueprint out what their perfect system might look like that they can’t afford today,” Townsend said, “but if they were to catch a big fish, and that person wants to come to the community, be ready to tell them, ‘You’e only coming here if you build the perfect system for me.’”

But even with such investment, it isn't a guarantee that design and construction would be fast enough to meet the aggressive timelines of the burgeoning data center market.