工作负载的社会价值能否塑造数据中心政策？

Questions over grid access are becoming more politically sensitive as policymakers weigh how to allocate scarce power among competing forms of digital infrastructure.Getty Images

Data centers are facing increasing scrutiny, and, in some cases, regulation, over their energy and resource usage, especially amid public opposition to rapid expansion in some regions.

Much of that scrutiny still focuses on data centers as energy- and resource-intensive buildings rather than on the workloads they ultimately support. But as the largely AI-driven build-out of new facilities comes under greater pressure, questions are also being raised about the societal value of the services data centers provide relative to their environmental and infrastructure costs.

That is still more of an emerging policy question than an immediate regulatory reality. Today, most regulation and public criticism remain centered on the physical footprint of data centers – power demand, water use, land consumption, and grid impact – rather than the digital services running inside them. But discussions at Informa’s recent Data Center Europe event in Vienna, which ran Oct. 13-15, suggested that the distinction may become hard to maintain over time.

Could Some Workloads Be Seen as More Valuable Than Others?

Data centers, and the internet more widely, support a broad range of workloads and services, from enterprise and business applications to social media, video streaming, generative AI chatbots, agents, and other AI services. According to Sandvine’s 2024 Global Internet Phenomena Report, video streaming, social media, and related applications from companies including Meta, Apple, Microsoft, Amazon, Alphabet, and Netflix account for about 65% of total internet volume.

Some digital services are already facing greater scrutiny in their own right. Social media regulation, including Australia’s move to ban access for under-16s, illustrates how governments are becoming more willing to intervene directly in digital activity when broader social concerns are at stake. That does not amount to data center workload regulation, but it does show how the societal impact of online services is increasingly part of the policy debate.

AI is likely to sharpen that debate further. Estimates of AI’s contribution to total data center footprint vary, but Goldman Sachs has projected that AI’s share of the overall data center market will double to 30% over the next two years, taking share from conventional and cloud workloads.

At the same time, although data center opposition groups – particularly in the US – have largely focused on the environment and physical footprint of new facilities, there is also concern – often driven by a lack of transparency – about the services those facilities host and enable – most obviously, AI.

“The AI boom is being sold as inevitable progress, but the real question is not whether artificial intelligence can do useful things in theory. It is who owns it, who profits from it, what it is mostly being used for…,” environmental group Greenpeace said in a recent publication.

In response, data center operators and the wider industry have pointed to the broad range of services that data centers provide beyond AI, and on which many aspects of society depend, from banking to healthcare. In a recent blog post, AWS CEO Matt Garman made that case directly, writing: “Today, virtually everything we do each day—shopping, banking, stock trading, ride hailing, booking travel, receiving healthcare, streaming entertainment, as well as critical research in everything from preventing and curing disease to finding ways to use less energy and natural resources – is digitally driven and enabled by data centers, which manage the computing, storage, and access necessary for these activities.”

Still, the question remains whether all workloads should be treated equally if policymakers decide how to allocate scarce power, approve grid connections, or weigh the local impact of new projects. Healthcare research, for example, may be seen differently from gambling, social media, or some speculative AI use cases. To date, however, there have been no significant attempts to reflect those distinctions directly in regulation, for instance, through lighter treatment or faster approvals for facilities that support a higher proportion of socially valuable workloads.

In practice, that would be highly challenging. Cloud and colocation environments support multiple workload types, often in the same facility, and digital infrastructure is far too complex for neat categorization. Even so, some broader distinctions are possible, particularly if governments begin to place greater emphasis on sovereignty, critical services, flexibility, or public benefit.

What Industry Leaders in Vienna Said

Responding to a question from Data Center Knowledge at the Informa Data Center Europe event this week, a panel of industry experts acknowledged that greater scrutiny around workloads, services, and their wider societal value could emerge in the future.

Branko Jovanović, CTO for data centers at ADVANT d.o.o. and president of the IT infrastructure and data center section of the ICT Association of Slovenia, suggested there are already early signs of that shift.

“There is maybe a case which is already under supervision,” Jovanović said, pointing to recent amendments to grid connection rules in Finland that would prioritize smaller data centers under 1 MW. In practice, that could mean some enterprise-owned or research-oriented facilities are favored over larger AI, colocation, or hyperscale cloud development, even if the policy stops short of explicitly ranking workloads by social value. “If we take into consideration Finland’s proposal on prioritizing energy permissions, it is already touching the workloads. Those big data centers are required to prove their maturity, to prove their flexibility, also in terms of possible scheduling workloads and also other things like supporting energy production.”

Gabriele Favarò, senior policy manager at the Information Technology Industry Council, agreed that more transparency is needed around the services data centers support and said current regulation does not yet reflect the complexity of the market.

“At this moment in Europe, under the CADA [Cloud and AI Development Act], there isn't differentiation for type of workload or in other pieces of legislation,” Favarò said. “There isn't even diversification of the type of data center, if we're talking about AI data centers or we're talking about colocation.” He added that there is a need for “more awareness-raising exercises on what's beyond that black box. Regulations should somehow reflect that.”

From the customer perspective, Rajiv Dalal, managing director at Strategic eServices, said he is already seeing greater interest in the relationship between workloads and the physical facilities that support them. He pointed to one client – a Native American reservation in the US – that is developing a private cloud environment to retain sovereignty over its data. He said similar thinking can be seen in regulated sectors such as financial services, where organizations remain highly sensitive about where critical workloads reside.

A Difficult Line for Regulators to Draw

That does not mean the industry is comfortable with the idea of regulators attempting to sort digital activity into categories of greater or lesser social value. Speaking at Datacloud Global Congress 2026 earlier this year, Damir Spoljaric, founder and managing partner of investment group Gi21, warned against moving in that direction.

“There is no ‘good’ workload or ‘bad’ workload,” Spoljaric said. “For me, it is too dangerous. I would keep it as it is.”

Michael Winterson, secretary general of the European Data Centre Association, raised a related concern at the same event, asking: “Whose right is it to regulate these activities in a liberal democracy?”

For now, the idea of formally regulating data centers according to the societal value of the workloads they support remains speculative. But as opposition to new development grows, and as AI increases pressure on power and infrastructure, the issue is likely to come up more often. What is now a conference debate could, over time, become a more serious policy question.

About the Author

Reporter

Andrew Donoghue specializes in journalism, analysis, and thought leadership content for sectors including data centers, critical infrastructure, and sustainable technology.

He has held senior content, strategy, and market intelligence roles at organizations including Vertiv and Fujitsu. He is the author of several influential reports covering topics including renewable energy and the data center, IT power management, and data center cooling.

Andrew has worked for analyst companies such as 451 Research and has also held senior editorial roles at business publishing companies, including CBS Interactive and Incisive Media. He has also been involved in several European Commission-funded IT research projects on data center energy efficiency and sustainability.