The UAE is following the global trend of building datacentres but is assessing the impact of the energy-intensive sector on its power generation.
The OPEC member has been shifting to renewables and nuclear as part of its 2050 net-zero target, and although it has the biggest renewables sector in the Mideast Gulf—with a 70% year-on-year growth in installed capacity in 2023—it is also witnessing an uptick in demand for power, for which it relies mostly on gas-fired plants.
As a result, the UAE’s energy ministry is undertaking a review of the impact of datacentres on the sector.
“The team is tasked with analysing and reviewing the impact of datacentres on energy demand, evaluating the local market and projected economic return for this key sector, identifying all datacentres in the country and classifying them according to specific standards,” the energy ministry said in a statement to Petroleum Economist.
The energy capacity of datacentres is forecast to increase from 495.7MW in 2025 to 917.7MW by 2030, with a compound annual growth rate of 13.11%, it added.
Gas demand growth
As the UAE seeks to become a regional leader in AI, its domestic gas consumption is soaring amid a fast-growing population, the expansion in energy-intensive industries such as desalination plants and aluminium smelters, and higher LNG production.
ADNOC Gas, majority-owned by state oil company ADNOC, provides about 60% of the UAE’s gas needs and has revised its CAGR for domestic demand up to 2030 to 6%, compared with 2% in its 2023 IPO prospectus.
ADNOC Gas is boosting its 10bcf/d gas processing capacity by 30% by 2029 and is a major contributor to the UAE’s ambition to become a net gas producer by the end of the decade. Currently, the UAE imports about 2bcf/d of piped Qatari gas, which also reaches Oman.
“We are seeing a real effort by the UAE to increase domestic gas production, and some of that will surely go to power demand,” said Karen Young, senior research scholar at Columbia University’s Center on Global Energy Policy. “But their nuclear power capacity and possibility to expand to four additional reactors can also service expected needs from datacentres.”
The UAE wants to generate 32% of its electricity from c clean energy sources by 2030. In 2023, the country’s installed power capacity rose by 13% year-on-year, to 43.5GW, with the share of renewables rising to 14% that year from 9.4% in 2022, energy ministry figures show. In 2023, the UAE’s clean energy capacity was almost equally split between renewables and nuclear.
“In the UAE, as in most of the Middle East region, demographic and economic growth are expected to strengthen electricity demand in buildings, both in the residential and commercial sectors,” said Nadim Abillama, Middle East and North Africa Programme Officer at the IEA. “The UAE’s ambitions regarding datacentres and AI will require an abundant and stable electricity supply, which can be offered by nuclear power.”
In its 2024 annual report, ADNOC Gas explained that “additional energy [is] required to power AI datacentres, where 5GW of power demand is approximately equivalent to an additional 1bcf/d of gas supply.”
“We did our assessment, and in the UAE, to serve the region, will need a minimum of 5GW of datacentres and we are building that as we speak,” Peng Xiao, CEO of UAE AI champion G42, told the World Economic Forum in Davos in January.
The UAE’s datacentre capacity is a third bigger than that of Saudi Arabia, according to data from market intelligence firm DC Byte.
Power demand
The electricity consumption of global datacentres is forecast to more than double, to around 945TWh, by 2030, slightly more than Japan’s total electricity consumption now, the IEA said in a report published in April.
In 2024, datacentres represented about 1.5% of the world’s electricity consumption, or 415TWh, and they will account for around one-tenth of global electricity demand growth to 2030, it added.
In the Middle East, total electricity consumption from datacentres is forecast to double to 3TWh by the end of the decade compared with 2024 based on IEA’s base case projection.
“We are also focusing on the energy efficiency of these datacentres, because 40% of the power consumed in datacentres is for cooling,” Sharif al-Olama, undersecretary of energy and petroleum affairs at the UAE’s energy ministry, told Petroleum Economist at World Hydrogen MENA in Dubai in February.
“We believe there is quite a lot of improvement we can do in terms of the efficiency of the cooling.”
Khazna Data Centres, a unit of G42 that is building multi-gigawatt expansions across the UAE and internationally, is trying to address the cooling challenges in a country where temperatures can top 45C. Currently, it operates 25 datacentres in the UAE with a total capacity of 230MW.
“In this region, the biggest challenge is undoubtedly overcoming extreme climate conditions,” CEO Hassan Alnaqbi told Petroleum Economist. “We use adiabatic-free cooling chillers that consume 70% less energy than traditional cooling systems. Additionally, our facilities recycle up to 30% of their water to promote sustainable resource management.”
One of their datacentres in the UAE is exploring using waste-to-energy power, and Khazna also wants to work with partners to use direct liquid cooling at its facilities.
“The majority of our datacentres are powered by the UAE grid, but we are constantly looking at ways to not only decarbonise the energy mix at current facilities, but [also] ensure a cleaner mix from the outset at future facilities,” said Alnaqbi.
Deciding whether to use land for datacentres or other industries is one of the issues the energy ministry wants to tackle.
"Of course, we also need to assess the comparison between having datacentres, and occupying a lot of land,” said Olama. “So, the question becomes: do you want to utilise the land for datacentres that will create maybe two or three jobs, or do you want to create a processing plant that will create thousands of jobs and have a bigger impact on the respective local emirate GDP?”
Industrial power consumption is already rising with the development of new projects and expansion of old facilities.
ADNOG Gas is forecasting sales growth mainly due to rising demand from ADNOC and its units’ operations, as well as rising power needs from utilities and aluminium, cement and steel producers targeting exports.
Striking a balance
All this local gas demand may put pressure on the UAE to find a balance between fuelling domestic economic growth and feeding its LNG trains.
Curtailing rising domestic gas consumption, a common challenge among Gulf oil and gas producers, is difficult due to lingering subsidies.
Egypt, for example, has turned from being an exporter of LNG to becoming an importer to meet rising domestic demand for gas, which is used as feedstock at power stations. The problem is exacerbated by subsidised electricity prices.
So far, the UAE has managed to avert such a scenario thanks to Qatari gas imports, which continued to flow between 2017 and 2021 despite Abu Dhabi joining several countries in a political and economic boycott of Doha.
ADNOC Gas’ domestic sales were boosted in 2024 by a planned outage at the Dolphin pipeline, which transports the Qatari gas.
The fate of an agreement between Qatar and Dolphin signed in 2016 to boost supplies of gas through the pipeline, which has a 3.2bcf/d designed capacity, remains unclear.
“Usage of the additional 1.2 bcf/d capacity is subject to a future agreement between Dolphin Energy and the Qatari authorities,” the company says on its website.
It remains unclear if Abu Dhabi and Doha have the appetite to extend or boost supplies via the pipeline amid the UAE’s drive to reach gas self-sufficiency by the end of the decade.
“Pipelines are things that are built on trust, and you have to be able to trust your trading partner for a long time because there is no other option,” said Ira Joseph, a senior fellow at Columbia University’s Center on Global Energy Policy.
AI advantage
While the UAE is consuming more gas domestically, it also wants to leverage AI to lower costs, reduce emissions and fast-track the development of its oil and gas production.
AIQ, a joint venture between ADNOC and G42’s Presight, has been working with the Emirati NOC to deploy AI applications on company data spanning 80 years to help improve operations, cut costs and lower its carbon footprint amid plans to hit net-zero Scope 1 and 2 emissions by 2045.
ADNOC is boosting its oil production capacity to 5m b/d by 2027 from about 4.85m b/d now, and is spending $23b to help decarbonise its operations.
For example, one of AIQ’s applications will help monitor and forecast ADNOC’s GHG emissions for up to five years, while another is focusing on reservoir management and production optimisation that gives 75% insight into their data and has been used in 30% of the company’s reservoirs, Magzhan Kenesbai, AIQ’s acting managing director, told Petroleum Economist.
“Everything related to time reduction and workflow operations or optimisation always impacts field development planning, which again indirectly leads to lower footprint in terms of CO₂ emissions,” said Kenesbai. “This leads to optimisation of cost, this leads to optimisation of production planning, resource planning, supply chain planning, all across the board.”
AIQ’s power needs are not big because the company is “not a large utiliser” of datacentre capacity and relies on supply from existing facilities, he added.
However, weighing the benefits of using AI in oil and gas operations to reduce costs, fast-track operations and lower emissions against the high power consumption of datacentres and their carbon footprint may be tricky.
Despite this dilemma, AI is worth the investment, according to the IEA.
“In exploration and development, AI can make the evaluation of resources more reliable and reduce predrilling uncertainty,” it said in the April report. “In operations, it is being used to optimise and automate production processes, detect leaks, predict maintenance needs, and support efforts to reduce methane emissions.”
Emissions from electricity use by datacentres are forecast to increase by about 70%, to 300mt, by 2035 from now in the IEA’s base scenario, but these emissions will be below 1.5% of the total energy sector emissions during this period. The emissions reductions from the broad application of existing AI tools will be equivalent to about 5% of energy-related emissions in 2035 as well.
“We are living in a world right now where energy realism has to be applied in terms of the power necessary to cope with the increase of the global population, the increase in AI, and the increase in datacentre development and storage and so forth,” said Kenesbai.
“AI is driving investments at an unprecedented rate back to the energy sector, and through investments it is improving all utility sectors, distribution, power generation, and power storage to new heights.”
Comments