������Ƶ

DUBAI: Once fueled by oil, the UAE is now betting on bits and bytes. 

The Gulf state is rapidly building hyperscale data centers, positioning itself as the Middle East’s central node for artificial-intelligence infrastructure. Backed by billions in sovereign wealth and global partnerships, the country is trading petroleum pipelines for digital ones.

In August, Texas-based TRG Datacenters ranked the country among the world’s top three AI superpowers, alongside the US and ������Ƶ.

While this infrastructure promises growth, it also raises environmental and geopolitical concerns around energy, water and data sovereignty in a region already strained by climate extremes. 

Backed by billions: sovereign capital fueling hyperscale expansion

In 2024, Microsoft injected $1.5 billion for a minority stake in Emirati technology firm G42, joining its board and committing to co-develop a $1 billion fund focused on AI skills and infrastructure across the Middle East, Central Asia, and Africa. 

This capital infusion has empowered G42’s subsidiary, Khazna Data Centers, to spearhead the country’s hyperscale expansion.

The firm was formerly owned by Abu Dhabi’s sovereign wealth fund Mubadala, with the majority now owned by G42. It holds over 70 percent of the national data center market share. 

This investment is part of a larger global surge in AI infrastructure. A 2025 McKinsey analysis projects $1.7 trillion in capital spending on AI-capable data centers globally by 2030. 

But with growth comes cost: the International Energy Agency estimates global data center electricity use could double by 2030, reaching 945 terawatt-hours, nearly 3 percent of total global consumption.

Khazna Chief Strategy Officer Johan Nilerud told Arab News the company is embedding sustainability into every layer of its operations. 

“Our operations rely heavily on recycled water rather than potable sources,” he said. “We’ve engineered our facilities to deliver high-density compute while maintaining a power usage effectiveness of around 1.5, even in extreme conditions … compared to the regional average of 1.8.”

Nilerud added that Khazna does not see their growth “as being at odds with sustainability.” To maintain efficiency in temperatures exceeding 45 degrees Celsius Nilerud said “we’re investing in direct liquid cooling and immersion technologies that can support the next generation of high-density AI chips.”

Beyond physical infrastructure, G42 is also expanding into cloud computing. Its other subsidiary, Core42, signed a $3.54 billion multi-year agreement this year with Microsoft and the Abu Dhabi government to develop a sovereign cloud system to modernize public sector services. 

The deal comes as Abu Dhabi aims to become the world’s first fully AI-native government by 2027, signaling a commitment to digital self-reliance. 

Private equity partnership meets Gulf capital

In one of the most high-profile deals to date, US investment firm KKR entered a $5 billion agreement with Emirati conglomerate Etisalat by e& in January this year, marking its first data center investment in the Middle East.

KKR also acquired a stake in Gulf Data Hub, one of the region’s largest independent hyperscale platforms.

The partnership aims to support data center expansion across Gulf nations to meet surging demand from AI workloads, cloud services, and national digital agendas. 

Stargate is a future epicenter still in flux

The UAE’s $500 billion Stargate project, set to go live in 2026, is poised to become one of the world’s largest AI data center networks outside the US. 

The 10 sq. mile (25 sq. km) AI campus in Abu Dhabi is expected to be operated with 5 gigawatts of power and host up to 500,000 Nvidia chips yearly. Led by G42 and backed by OpenAI, Nvidia, Oracle, Cisco, and Japan’s SoftBank Group, Stargate represents a new frontier in Gulf-led AI infrastructure.

But the project’s scale has drawn scrutiny. 

“The risk that some of the US’ most sensitive intellectual property could leak to US adversaries — or that those adversaries could access US AI systems in the Gulf ... remains very real,” Sam Winter-Levy, technology fellow at Carnegie Endowment for International Peace, told Arab News.

To mitigate such risks, Microsoft reportedly included strict safeguards: G42 is prohibited from using Microsoft’s AI chips for surveillance and must seek approval before sharing its technology with foreign governments or military entities.

Still, US export licensing remains unresolved amid lingering American concerns about the UAE’s ties to China, raised during both the Joe Biden and Donald Trump administrations.

“The US could retract or limit licenses in the future, if it wanted; it controls key parts of the AI supply chain,” Winter-Levy said. “But the Gulf has leverage too: they could freeze payments, turn back to Chinese providers, or even try to seize control of the chips.”

These geopolitical tensions cast uncertainty over the future of Stargate. 

Khazna’s Nilerud told Arab News that “in the UAE, we’re seeing a clear move toward sovereign-backed infrastructure that ensures critical data remains within national borders and under jurisdiction.”

Sovereign strategy and sustainability balancing act 

In June of this year, Sultan Al-Jaber, CEO of Abu Dhabi National Oil Co., announced plans to grow the UAE’s US investment portfolio to $440 billion over the next decade.

Calling AI a “once-in-a-generation opportunity,” he emphasized that the “US is not just a priority, it is an investment imperative.”

Winter-Levy argued that while Gulf states are amassing enough resources to develop sovereign AI capabilities, “they will still remain dependent on foreign technology for the foreseeable future ... advanced chips that, for now, only the US is capable of producing at scale.”

Yet the power demands of this AI-driven future are rising sharply. Goldman Sachs projects data center electricity use will surge 165 percent by 2030, largely due to AI workloads.

With digital infrastructure now sitting at the intersection of energy, economics, and geopolitical influence, the region’s push to lead in AI will depend not just on how fast it can scale, but on how sustainably it can grow.