California joins US states clamping down on data center gold rush
California joins US states clamping down on data center gold rush
if ( !emtpy($headline_subheadline ) ) : ?>
At least 45 US data center projects worth an estimated $68 billion were blocked or delayed in the second quarter, as communities increasingly push back on power, water, and land demands; analysts say enterprises need a different approach.
endif; ?>
Credit: IgorsPerspective | shutterstock.com
The state of California has joined the growing number of governments and communities at all levels across the US taking proactive, even preemptive, measures to keep data center development under control.
California Governor Gavin Newsom signed what he called the “most comprehensive data center laws in the nation” on Monday. While the seven bills don’t block data center development, they do reinforce requirements for reporting data centers’ energy and water use, and ensuring they pay their fair share of utility costs.
Other states — and even local community groups — have been taking more direct action. At least 45 projects worth an estimated $68 billion were blocked or delayed in the US in the second quarter, according to research by Data Center Watch. There are now 843 identified opposition groups in the US, and 49 states are scrutinizing data center development, it said.
Project disruption remains at an “elevated level” and opposition has become “more local and project-specific,” targeting permitting, infrastructure, and individual policy decisions, according to Data Center Watch.
Some communities are getting ahead of development by preemptively mandating data center development moratoriums, even before developers have filed permit applications or expressed any interest at all.
Data centers no longer ‘invisible infrastructure’
“The growing pushback is not surprising,” said Abbas Jaffery, principal advisory director at Info-Tech Research Group. “The conversation around data centers has shifted from a routine IT capacity decision to a complex energy, community, and economic issue.”
Anti data center sentiment is higher than it’s ever been, as AI gobbles up more and more resources and requires never-before-seen magnitudes of data.
State-by-state legislative activity is accelerating in response, with 30 state governments adopting or introducing legislation, resolutions, and executive actions targeting siting, cost-sharing, and electricity and water constraints, Data Center Watch reported.
Further, regulatory scrutiny is increasing across 49 states, and communities are clamoring for local, state, and national politicians to respond to data centers already in development.
California is a case in point; Newsom has just signed seven bills addressing data center costs, disclosures, and infrastructure. Operators will be required to pay their “fair share” for updates to the state’s electricity grid, comply with California energy procurement requirements, and bring on new clean energy supply.
Data center companies will also need to provide information on water use, supply, efficiency, and drought planning, and pay for any required infrastructure upgrades. Additionally, data centers will be ineligible for blanket environmental exemptions; they must demonstrate that their projects meet state energy, water, and fuel consumption standards prior to approval.
“As the innovation economy continues to grow, data centers are being presented as a solution, but with little thought or oversight of what that means for nearby communities,” Newsom said in Monday’s announcement. The goal of the bills is to protect communities from “shouldering the cost as industry reaps massive profits.”
Other regulatory initiatives can be found across much of the country: Independent electricity marketplace Electric Choice tracks 321 moratoriums and restrictions across 32 states, with pending legislation in 17 states. Just last week, Virginia Governor Abigail Spanberger unveiled a Data Center Accountability Framework, while in July, New York Governor Kathy Hochul placed a temporary moratorium on hyperscale data centers in her state.
“Data centers are no longer invisible infrastructure,” said Justin St-Maurice, technical counselor at Info-Tech Research Group. “They are being built in communities, where residents can directly experience the noise, energy demands, and other consequences.”
Enterprises must take a new approach
While some concerns may be more legitimate than others, all require a stronger understanding and partnership between builders and communities, noted Matt Kimball, VP and principal analyst at Moor Insights & Strategy.
“Dismissing fears around water consumption, for example, by showing a spreadsheet at a local planning committee meeting, doesn’t resolve concerns for a community that is already suspicious,” he said.
Community opposition “is real, and it’s everywhere,” and the new strategic pillar for data center builders and operators is social outreach, Kimball noted. Those proposing data centers must be able to provide credible answers about usage and community impacts, listen to concerns, and commit to transparency.
Most enterprises aren’t building gigawatt campuses, he pointed out, but they are paying the price downstream in colocation availability, lead times, pricing, and other factors. Predictability is the big question, supply is already tight, and every delayed project removes capacity factored into forecasts.
“IT leaders should treat power and space as a strategic constraint rather than a facilities line item,” he said.
This means extending planning horizons and committing to colocated capacity earlier on. Ask operators where their sites actually are in the permitting and entitlement process, not just their roadmap, Kimball advised. Also, he said, “squeeze every bit of productivity out of the footprint you already own.” For instance, older servers consolidated onto current-generation platforms can free up rack space and power for AI projects without “a single extra square foot of data center space required.”
Buyers should also be deliberate about workload placement and underlying infrastructure. What can sit in the cloud and what needs to be closer to the premises? What kind of inference is running? “Not every workload requires the latest and greatest GPU cluster,” Kimball said; in fact, some don’t require a GPU at all.
He pointed to companies like Furiosa AI that are building more efficient hardware expansion cards, while CPUs like Intel’s Xeon feature on-chip accelerators that can efficiently handle inference workloads. Nvidia, for its part, is claiming its latest technology advancements can provide up to 40% more GPU capacity.
He noted that other interesting technologies to watch are two-phased cooling, and heat recapture and reuse. For instance, at one data center in Switzerland, generated heat is repurposed to heat nearby homes.
‘Radical optimization’ the way forward
Info-Tech’s St-Maurice agreed that IT leaders cannot assume they can rely indefinitely on public cloud capacity at a predictable cost. Some organizations are reconsidering their compute strategies as token costs rise, repurposing internal data centers, running open-weight models, and using a tiered approach combining local, enterprise, private cloud, and public cloud resources.
“The goal is to direct each workload to the most cost-effective environment and gain more control over operating costs,” he said. While regulations will likely lead to innovation, more space, cooling, and power alone will not solve the problem.
The more sustainable path is “radical optimization,” St-Maurice said: Reducing energy cost per token and using compute more efficiently.
Indeed, when considering new data centers, enterprise leaders must ask whether the supporting power, cooling, water, and grid infrastructure actually exist, in what timeline, and at what cost, Info-Tech’s Jaffery advised.
“Organizations that integrate physical grid and power constraints into their IT architecture early will maintain speed to market,” he said, “while those that treat power as an afterthought will face costly deployment delays.”
This article first appeared on Network World.
Data Center DesignData CenterLaws and RegulationsGovernmentMarketsIndustry
About Author
