How data centres could help the Gulf meet its climate targets

Saudi Arabia, the UAE and other Gulf countries are aggressively investing in AI as they seek to maintain their competitiveness. Getty

Data centres that power the Gulf region’s burgeoning cloud and AI systems are accounting for an ever-increasing proportion of electricity consumption. If managed well, this nominal increase in the need for power generation can be transformed into a tool for the greater use of renewable energy, helping the GCC countries to realise their ambitious climate goals.

Saudi Arabia, the UAE and other Gulf countries are aggressively investing in AI as they seek to maintain their competitiveness in a rapidly evolving global economy. One of the key issues that policymakers must grapple with is how to supply AI infrastructure with the power it needs to operate. Despite the recency of these novel technologies, they already account for almost 2 per cent of global electricity consumption; this figure is set to grow exponentially over the coming decades as AI is integrated into every walk of life.

A parallel and ostensibly independent challenge that governments are contending with is the need to decrease carbon emissions as part of the global fight against climate change.

The Gulf states have committed to realising carbon neutrality within the next four decades, driving heavy investments in renewable energies such as solar and wind. They also see this as an opportunity to develop a comparative advantage in the probable successor to fossil fuels, as reflected in Abu Dhabi’s Masdar becoming a global supplier of cutting-edge clean energies.

Although the technological advancements in renewable energy have been breath-taking, two key challenges persist.

The first is intermittent generation combined with a lack of battery infrastructure. The Sun does not shine all day, nor does the wind blow throughout it, and so electricity demand is invariably out of sync with the unpredictable supply profile that these technologies provide. This leads to significant waste and to the need to invest in backup, fossil fuel-powered generators to avoid supply disruptions.

The second problem is that renewable energies cause voltage and frequency fluctuations, which in turn lead to instability in the electricity grid. The resulting blackouts can be damaging economically, and also constitute a national security threat given the power-intensive and integrated nature of most modern defence systems.

Dubai developer Damac Properties announces $20 billion investment in the US

00:54

As a result of these two difficulties, renewables’ installed capacity has not kept pace with the advancement in generation technology. However, a new paper published by Massachusetts Institute of Technology economists Christopher Knittel, Juan Ramon Senga and Shen Wang offers policymakers a way of killing two birds with one stone.

The scholars begin by noting that one of the key advantages of data centres is the inherent flexibility of their electricity demand profile. While some baseline processes must occur on a continuous basis throughout the day, a lot of the computationally intensive – and hence electricity-hungry – operations can easily be delayed by a few hours or more without disrupting overall performance. An example would be a large language model such as ChatGPT reading and understanding a voluminous new tranche of material, such as all of the manuscripts in a collection that has recently been digitised.

As more renewable energy is integrated into the generation profile, AI data centres can be instructed to synchronise their flexible operations – and hence their electricity demand – with the renewable energy output

As long as AI data centres continue to account for a small proportion of electricity consumption, this flexibility will remain inconsequential. However, if – as expected – they transition into being among the biggest sources of energy demand, the flexibility can be exploited to mitigate the intermittency and grid instability problems caused by renewable energy.

More specifically, as more renewable energy is integrated into the generation profile, AI data centres can be instructed to synchronise their flexible operations – and hence their electricity demand – with the renewable energy output. This diminishes the likelihood of unconsumed electricity being wasted and decreases the need to invest in expensive battery technologies.

Moreover, the voltage and frequency surges that destabilise the grid are likely to decline in incidence. As a result, renewable energies become much more reliable components of the energy mix, and governments can more confidently invest in them without fearing the fragility that has so far limited the transition away from fossil fuels.

This solution is by no means a silver bullet. A key weakness is that it depends on a large growth in electricity demand occurring and being matched by a comparable expansion in electricity generation as a prerequisite to shifting the balance towards renewables. However, this challenge will almost certainly be mitigated by complementary technological advancements in other areas, such as battery technology and smart grid management. As a result, Gulf countries can exploit this mechanism as part of a multi-pronged approach to greening their energy mix.

In short, the rise of AI and the accompanying expansion of data centres need not be a burden on the GCC’s climate ambitions. On the contrary, as policymakers embrace a strategic view of electricity demand management, flexible data centres can become powerful enablers of a greener grid. By aligning the region’s digital and environmental transformations, the Gulf can turn a looming challenge into a unique opportunity – advancing both its technological leadership and its commitment to climate responsibility.