LOCAL AI ZONES: what's at stake?

<<The Government’s announcement that Blyth in Northumberland and Cobalt Park in North Tyneside will become designated AI Growth Zones has prompted debate about the trade-offs we need to consider. Durham University’s Dr Karen Lai outlines the issues raised.>>

The UK government has designated two sites in the North East as AI Growth Zones, with the potential for thousands of jobs and billions in investment. However, there is a fundamental paradox with data centre growth: the drive for digital competitiveness often conflicts with net-zero and other environmental goals. Navigating this requires a hard look at the real costs of our digital future.

This move is designed to attract companies building data centres and other AI infrastructure, offering them incentives like faster access to power and planning approvals. The news was accompanied by major investment announcements from firms like OpenAI and NVIDIA. The announcement promised more than 5,000 new jobs and £30 billion in investment, positioning the region as a future hub for AI development. While business investment in the North East is much needed, we must carefully consider the potential downsides and trade-offs that come with the rapid growth of data centres. The key concerns involve high energy demands, environmental impacts, and whether the economic benefits will truly be felt locally.

The Physical Footprint of a Virtual World

Although we think of AI and cloud computing as existing in a virtual space, the physical infrastructure that underpins them has a significant and often overlooked environmental footprint.

• The most prominent concern is intensive energy consumption. Data centres are very energy-hungry, and their projected growth in electricity demand could jeopardise climate targets. In 2023, data centres in Ireland accounted for 21% of the country’s entire electricity consumption, a figure that has increased by a staggering 400% since 2015.

• Beyond electricity, the expansion of data centres places considerable pressure on other resources like land and water.

  • AI applications require ever larger and more powerful facilities, which in turn demand more land.

  • Furthermore, while innovations like liquid cooling can be more energy-efficient than traditional air cooling, they could create another problem: increased water consumption in a country where climate change is already creating more water scarcity.

The environmental impact spans the entire lifecycle of a data centre, from construction to decommissioning. This includes the ‘embodied carbon’ in building materials and hardware, the environmental damage from mining critical minerals for components, and the challenge of managing electronic waste when equipment reaches the end of its life.

Innovation, Regulation and Financial Challenges

How can we manage the tension between digitisation and decarbonisation? The challenge requires strategic decisions that balance economic imperatives against environmental costs, from a data centre's location and technology to its financing and regulation.

Technological innovation offers potential solutions, but the sector itself presents barriers. The data centre industry is characterised by a conservative and risk-averse culture, where the primary focus is on avoiding downtime, which is financially and reputationally costly. This culture often slows the adoption of new technologies in favour of ‘tried and tested’—but often less efficient—methods.

One method of operating data halls at a slightly higher ambient temperature (27 degrees Celsius instead of the commonly used 20-23 degrees) could significantly reduce cooling energy, but this is often met with resistance over concerns about equipment safety. Similarly, more energy-efficient liquid-based cooling systems face slow adoption due to high upfront costs and perceived risks of failure. Many operators also rely on controversial instruments like Renewable Energy Certificates (RECs) and Virtual Power Purchase Agreements (VPPAs) to offset their carbon footprint, which have been criticised for their questionable effectiveness in actually reducing emissions.

Regulation presents its own complexities. In recent years, governments in Singapore, Dublin and Amsterdam have imposed moratoria on new data centre construction to buy time in developing new sustainability regulations. However, this is only a limited solution as it can harm a region's long-term competitiveness. A major risk of strict regulation is the potential for displacing environmental costs. When Singapore implemented higher environmental standards, a flood of new data centre investment moved to neighbouring jurisdictions in Malaysia and Indonesia, undermining the intended environmental benefits.

For regions like North East England, which are actively trying to attract data centres for economic revitalisation, this strategy needs closer scrutiny. It is crucial to examine the actual job creation and economic spin-offs to ensure that profits are not simply funnelled to offshore tech giants without tangible and substantive benefits to the regional economy and local communities.

Finally, green finance could play an important role, but it needs to be deployed more effectively. Currently, instruments like green bonds are predominantly directed toward new ‘greenfield’ data centres built to be sustainable from the start. There is a significant lack of investment and regulatory guidance for retrofitting existing ‘brownfield’ assets to make them more energy-efficient. Investors often perceive the economics of retrofitting as unfavourable due to building age, design limitations and the risk of downtime during upgrades. Investors in funds that hold data centre assets could use their leverage to push for more comprehensive sustainability standards for both new and existing facilities.

Ultimately, sustainable data centre development is shaped by a persistent tension between immediate economic priorities and long-term environmental goals. Successfully navigating these trade-offs requires collaboration among policymakers, data centre operators, technology firms, investors and local communities. Looking beyond the headlines, we must ensure that environmental priorities and local benefits are not sidelined in the race to build a national AI future.

***Dr Karen Lai is Professor of Economic Geography at Durham University (UK). Her research expertise includes geographies of money and finance, FinTech, financial infrastructure and digital economies, focusing particularly on issues of financialisation, knowledge networks and financial centre development. https://www.durham.ac.uk/staff/karen-lai/