TWO STARTUP EXECUTIVES addressed grid capacity and water usage concerns held by communities about data centers during a panel discussion at the San Francisco Bay Area Planning and Urban Research Association on Wednesday, aiming to assuage public concerns around the negative impact of the centers.
Panelists Chris Lander, vice president of XFRA at SPAN, an at-home data center developer, and Aaron Tartakovsky, CEO and co-founder of Epic Cleantec, a water recycling technology company, both agreed that it was important for the data center industry to take the public’s concerns seriously and believed it was possible to resolve them.
According to research by the Environmental and Energy Study Institute, a nonpartisan think tank based in Washington, D.C., large data centers used for artificial intelligence can consume up to 5 million gallons of water per day, equivalent to the water use of a town of 10,000 to 50,000 people.
By 2030, data centers’ projected electricity use of 130 gigawatts would represent close to 12% of total U.S. annual demand.
Earlier this month, more than 300 residents gathered at a City Council meeting in Pittsburg to voice their outrage and concerns over the long-term impacts on water and power usage over a data center that was approved in 2024 to be built on the former Delta View Golf Course site.
The proposed 347,740-square-foot data center will generate 96 megawatts of power, cooled by recycled water, and house 37 diesel-fueled backup generators, according to its developer AVAIO.
Powering data centers from homes
For Lander, it is important that companies “align incentives” between the public and industry to allow for data center construction while addressing the public concerns while providing them an opportunity to reap some of the financial benefits generated by artificial intelligence companies.
Lander spent much of his time discussing how his company utilizes existing grid capacity to power data centers by accessing spare capacity at people’s homes to power a small data center that is self-contained, requiring no outside water or additional resources to function.
“Traditional data centers are scrambling to try and build as fast as they can to meet demand that is driven by all of us,” lander said. “We’re going to need every solution that we can get to try and meet that demand,” said Lander.
At scale, Lander said, his household data center solution could achieve the same amount of computing power in six months that a centralized data center would take five years to build out, based on how fast it could be deployed.
In addition, customers would receive compensation for access to their excess energy capacity, allowing average consumers to benefit from data center growth.
Using recycled water for cooling
Tartakovsky added he was optimistic that the industry would find its way, adding that the public has historically reacted negatively toward new technology when it’s first introduced, citing the fear people initially had around indoor toilets, electricity and elevators.
“Data centers and what they are enabling are going to do a lot of good for this world,” Tartakovsky said. “My wife’s a cancer survivor, seeing now what some of these AI technologies are able to do, from identifying cancer to coming up with new drugs that our human minds couldn’t possibly conceive of … we’re so focused on all the bad that can happen that I think we oftentimes are missing just how much good these technologies are going to enable for this world.”
Tartakovsky noted that technical solutions are being developed to address public concerns at a rapid pace due to the level of money invested in AI development, and political pushback against the speed data centers are being built.
His focus has been on the development of “closed loop” water recycling technologies to minimize water necessity for cooling by recycling water used in toilets, sinks and showers of neighboring buildings directly into data centers, preventing the need for water to be pulled from other sources.
Both Tartakovsky and Lander both agreed that the role of policymakers is to incentivize the development of these solutions to bring them to scale. Despite the public push toward moratoriums on data center construction, they both thought it was possible to get policymakers to work with industry to get this right.
“I’m just an eternal optimist,” Tartakovsky said.
