Sustainability’s legacy was sidelined by AI growth and politics, but public scrutiny and hard economics are forcing a course correction.
Data centers are where compute turns into real electrical load, heat, and water demand, and where efficiency and resilience are engineered under grid, thermal, and permitting constraints.
For operators, power, cooling, water stewardship, air quality, and building performance aren’t background considerations but instead define the operating envelope. At the megawatt scale, every kilowatt and liter translates into cost, risk, and compliance. Because data centers are the physical manifestation of the cloud – sited in communities and tied to local grids and water systems – they carry a clearer and more enforceable environmental mandate than the software stack they host. That visibility has brought sharper regulatory oversight and pressure from environmental groups over the years.
For example, Greenpeace has published a series of reports on the environmental record of hyperscalers going back more than a decade. Moratoria and permitting s have appeared in multiple markets, including [Singapore](http://datacenterknowledge.com/search?q=singapore), Amsterdam, Dublin, and, most recently, [New York State](/data-center-construction/new-york-data-center-moratorium-state-s-projects-over-50-mw), which enacted a one-year moratorium on July 14. Meanwhile, US lawmakers in multiple jurisdictions continue to push new data center regulations.
Regulatory Pressure Meets Voluntary Standards #
Beyond complying with tightening regulations, the industry has increasingly embedded environmental stewardship and broader corporate responsibility into design, procurement, and day-to-day operations. Global industry groups, such as The Green Grid, championed initiatives like Power Usage Effectiveness (PUE) about two decades ago, which became the standard measure of data center efficiency. Operators such as Microsoft and Google made commitments to slash carbon emissions through efficiency efforts and grid-related mechanisms such as power purchase agreements (PPAs) for renewable energy.
According to 2025 data from the annual Uptime Institute survey, PUEs dropped dramatically during the early 2010s. “The industry’s average dropped quickly – from 2.5 in 2007 to 1.65 in 2014 – as operators first deployed the most practical and economical changes,” the Uptime report states. “Operators favored upgrades to older electrical systems, better airflow management, and optimized cooling system controls because these changes were often straightforward and less disruptive to live workloads – and paid dividends in cost savings, as well as facility efficiency.”
Michael Winterson, secretary general at the European Data Centre Association (EUDCA), believes the industry has been proactive around self-policing (though some may argue this was an attempt to head off regulation). “Over the last two decades, sustainability has become increasingly central to how data centers are designed and operated,” he said. “Importantly, many of the industry's sustainability initiatives predate current legislation. The Climate Neutral Data Centre Pact, established in 2021, is a good example of operators coming together voluntarily to set ambitious targets around energy efficiency, renewable energy, water stewardship, and heat reuse.”
AI’s Scale-Up Sidelines Sustainability – Temporarily #
The advent of generative AI – and the rapid buildout of facilities driven by a training arms race – shifted the attention. An industry that had integrated sustainability into its core strategy (albeit under external pressure) saw that focus eclipsed by an intense emphasis on chip performance, speed, and scale. Ironically, awareness of the environmental impact of data centers has never been higher, even as many perceive a shift in focus toward performance and scale.
Following Q1 2026 earnings, analyst estimates suggest the largest US hyperscalers plan as much as $725 billion in capital expenditures this year. Against this backdrop, analyst group Gartner projects that total data center electricity consumption could approach 3.5% of global demand by 2030, with accelerated AI servers accounting for a large share of the growth.
Even before the current AI surge, a financial downturn combined with shifting political attitudes put pressure on corporate environmental, social, and governance (ESG) programs, which had become tightly woven into strategy at large cloud providers during the mid-to-late 2010s. Hyperscalers have also acknowledged that AI growth is undercutting some prior sustainability targets. A special report from S&P Global recently noted: “Big Tech firms with net-zero commitments have recently acknowledged that meeting those commitments is becoming more challenging. In Google’s 2025 sustainability report, for example, the company described its net-zero goal as a ‘climate moonshot’ and admitted that scaling carbon-free energy technology by 2030 – the target year of its net-zero goal – will be ‘very difficult.’ It also said AI-related energy demand growth has made its future emissions trajectory harder to project.”
Backsliding is not limited to hyperscalers. “Across the wider industry, data center sustainability commitments vary significantly, and net-zero ambitions are not a given,” S&P Global stated. “Data from the 2025 S&P Global Corporate Sustainability Assessment shows that 35% of assessed companies with data center operations lack a net-zero commitment.”
Meanwhile, some efficiency gains and reporting disciplines have stalled. The 2025 Uptime Institute survey reported a weighted-average annual PUE of 1.54, marking the sixth consecutive year in which this headline figure has remained virtually unchanged. “More laissez-faire governance regimes in some regions are undoubtedly contributing to this decrease in data collection,” the report adds. For example, in the US, the Securities and Exchange Commission adopted climate disclosure rules in March 2024, then voluntarily stayed them in April 2024 pending court review. As of July 2026, enforcement remains stayed, and timing is uncertain.
Community Pushback and Power Constraints Reset Priorities #
There are signs that performance, resiliency, and sustainability could be tipping back into equilibrium. Heightened scrutiny of environmental and social impacts is forcing operators – mostly large colocation providers and hyperscalers – to re-engage with sustainability or risk escalating push-back on expansion plans.
According to analyst firm STL Partners, as of 2025, public opposition had impacted approximately $77 billion worth of projects in the US. Northern Virginia alone saw about $48.7 billion in delayed capacity, and projects across at least nine US states were affected by opposition and protests.
Large operators are also signaling a pivot back to sustainability as a core mission. Frontier, structured as an “advance market commitment” to buy $1.8 billion of permanent carbon removal by 2040, announced an additional $915 million in funding for carbon removal projects. Members include Google, Salesforce, and Anthropic. Another initiative, Elemental Impact – with members including Amazon, Google, Meta, Microsoft, and Salesforce – is funding sustainable data center and climate-focused startups.
Cyre Mercedes Quiñones, CEO of global data center organization Infrastructure Masons (iMasons), told Data Center Knowledge that the industry’s stance has evolved from an initial focus on operational efficiency and energy consumption reduction to encompass renewable energy, carbon reduction, and broader environmental and societal responsibility. AI has refocused some attention on performance, she said, but sustainability is more critical than ever.
“Today, we have entered a new chapter. The rapid growth of AI is driving unprecedented demand for compute, connectivity, and power,” Quiñones said. “While this has intensified the focus on speed and scale, it has also elevated the importance of sustainability. As digital infrastructure becomes foundational to modern society, sustainability is not simply a ‘nice to have.’ It is about fulfilling our responsibility to build the infrastructure that powers the future in a way that strengthens communities, supports economies, and responsibly manages shared resources.”
Economics Make Efficiency Non-Negotiable in the AI Era #
According to Jay Dietrich, research director of sustainability at Uptime Institute Intelligence, operators have little choice but to take sustainability seriously if they want to keep expanding. “Sustainability at the facility level is now inseparable from a data center operator’s business performance,” he told Data Center Knowledge. “It is critical for securing approvals for new builds and controlling service costs in the AI era. Operators that overlook this do so at their own risk.”
Dietrich noted that sustainability has become an intrinsic requirement across multiple areas. “[The] scale of new data center projects requires operators to show that their facility design and accompanying operations strategy minimize energy and water consumption,” he said. “In arid and water-stressed regions, water use must be eliminated.”
Alex Cordovil, research director at industry analyst group Dell’Oro, agrees that the industry did refocus around performance, but says efficiency and sustainability are too business-critical to ignore, especially given ongoing concerns over power availability.
“Sustainability isn't returning as the goal; it's arriving as a by-product,” Cordovil told Data Center Knowledge. “And that's worth noting given the wider mood, where sustainability as a cause has cooled a little and the political and corporate appetite has shifted elsewhere.”
At today's AI load levels, the inefficiencies that once cost a few thousand dollars now add up to millions, Cordovil said. “What's striking is that the efficiency push is gathering pace regardless, because it's being driven by economics rather than the zeitgeist. But the more important point isn't the cost of that wasted power; it's the opportunity cost. Power is the binding constraint, so every watt not delivered to IT is a watt you can't monetize, capacity that never generates a token.”
Operators are also moving to the front foot over hot-button issues such as water consumption, though public pressure has been the forcing mechanism. For example, Amazon recently reinforced its commitment to reducing water usage, saying it was 75% of the way toward its goal of being water-neutral by 2030. The company claims that in Northern Virginia, its largest region by IT load, water use was reduced by 42% year over year, even as demand for computing continued to grow.
"This is how we innovate at Amazon," said Joern Tinnemeyer, a data center engineering leader at Amazon. "We set an ambitious target that benefits our customers, iterate relentlessly, and validate with data – in this case, proving we could cut water use in half without any impact on performance."
Raising the Bar: Net-Positive Expectations for AI Infrastructure #
The data center industry’s sustainability posture – much like that of wider business and society – has been driven by regulation, capital and operational efficiency, and periodic bursts of apparent altruism. It has also reacted to headwinds such as financial crises, alongside the tailwind of increasingly frequent extreme weather events as evidence of human-made climate change. After an initial AI-era period focused squarely on performance, the next phase will require the industry not only to reconnect with its sustainability legacy but also to set a higher bar – one where environmental stewardship, community impact, and economic pragmatism reinforce rather than impede growth.