Data centre backlash grows as AI demand surges

Data centre backlash grows as AI demand surges

The data centre backlash is no longer a niche concern. According to BBC reporting, there are about 12,250 data centres around the world, up sharply from around 8,000 operational sites in 2021, and the US alone has 4,767 sites. The same facilities that power web searches, emails, streaming and online banking are now at the heart of the AI boom, which is making their footprint harder to ignore. For companies, the risk is no longer only technical. It is political, local and increasingly expensive.

Key takeaways

• There are about 12,250 data centres worldwide, up from around 8,000 operational sites in 2021, according to Data Center Map and BBC reporting.
• The US has 4,767 data centre sites, while Europe trails in capacity, according to BBC reporting.
• The tech industry is expected to spend a combined USD 765bn on AI infrastructure in 2026 and more than USD 1tn next year, according to BBC reporting.
• Sixteen US states have proposed laws to ban or temporarily block new data centre facilities, according to BBC reporting.

Why is the data centre backlash spreading now?

Because data centres now strain the same resources that local communities already fight over. According to BBC reporting, these facilities can use significant amounts of water to cool their systems, which could lead to water shortages. The same report says data centres also put pressure on local grids, which some fear could push up energy prices, while noise and the land they occupy are adding to the anger.

The political response is widening. In the US, the issue is landing ahead of the midterm elections on 3 November, and the BBC says both Republicans and opposition Democrats are coming out against new development. Donald Trump has dismissed the pushback as a "sick conspiracy", even as public anger grows.

Source: BBC News

How fast is the data centre boom?

Fast enough to reshape the map of global infrastructure. According to Data Center Map, there are about 12,250 data centres worldwide, compared with around 8,000 operational sites in 2021. The US has 4,767 sites, while Europe, led by the UK and Germany, trails in capacity.

That gap matters because capacity, not just site count, determines where AI and cloud workloads can scale. The BBC says the US is ahead by a much larger margin on actual capacity, with China in second place and Europe far behind. Some observers warn Europe could miss part of the economic benefit from the AI boom.

Source: BBC News

Why are AI data centres getting so large?

Because AI workloads need enormous computing power. The BBC says data centres vary from shed-sized sites to facilities covering vast areas of land, and the world's largest operational data centre is widely reported to be China Telecom's site in Hohhot, Inner Mongolia, at about one million square metres.

The next wave is even larger. Colossus 2, part of the wider Colossus campus, will house hundreds of thousands of Nvidia chips and be used to train xAI's Grok. Elon Musk's company will use 41 methane gas turbines to help power the site, while local regulators say the turbines will meet air quality standards.

Source: BBC News

What does this mean for local economies and corporate strategy?

The business case is real, but so are the trade-offs. The BBC says supporters see data centres as important for national security, AI development, investment, tax revenue and construction jobs. A Brookings Center report found that data centres do create local jobs, but fewer than industry advocates claim, and hyperscale facilities have a bigger positive effect than colocation sites.

For companies deciding where to build or buy capacity, that means the cheapest option on paper may not be the safest option in practice. Energy access, local permitting, community response and emissions risk are now part of the location decision, not an afterthought.

Source: BBC News

What to do now

If your business depends on cloud capacity, AI training or large-scale digital operations, treat data centre access as a strategic dependency, not a procurement detail. The article shows that power, water, politics and community acceptance can all shape whether a site gets built and how much it costs to run.

• Map power, water and permitting risk before you commit to a site or a vendor.
• Stress test energy price and grid congestion assumptions in every major capacity plan.
• Compare hyperscale and colocation options on resilience, local acceptance and control, not only on headline price.
• Build a community and regulator engagement plan into capex from day one.
• Track local election cycles and planning rules where expansion could be delayed.

Frequently asked questions

What is driving the data centre backlash?

The backlash is being driven by the same things that make data centres useful: they need large amounts of electricity, water and land. According to BBC reporting, local communities are also worried about noise and energy prices. In the US, the issue has become political ahead of the 3 November midterm elections.

Why do data centres matter to AI?

Data centres house the advanced computer chips and servers that run AI services such as ChatGPT and Claude. BBC reporting says these giant buildings are at the heart of the AI boom. As AI workloads grow, the need for more computing power is pushing facilities to become larger and more energy intensive.

Should companies worry about where their data is hosted?

Yes. BBC reporting shows that site choice now affects cost, resilience and public acceptance. A location with easier access to power lines or water may still face local opposition, higher energy pressure or slower approvals. For companies, hosting decisions are becoming part of strategic risk management, not just IT architecture.