Gulf AI Infrastructure: Examining the Business and Economic Case

High-tech futurism has become central to the Gulf’s post-oil vision. That vision now faces new pressures as the US-Israeli war against Iran unfolds. The business and economic case for the Gulf’s data center drive has become simultaneously more urgent and more fraught. Urgent, because the war has exposed structural vulnerabilities: kinetic attacks on data infrastructure, the limits of the Gulf’s “safe haven” reputation, and the crowding-out of Vision spending by surging defense budgets. Fraught, because sovereign wealth fund capital and investor confidence, once expected to fund the Gulf’s AI ambitions, are now under significant pressure. While the race to build the infrastructure that could power the region’s AI future has not stopped, the conditions under which it must succeed have materially deteriorated.

Nowhere has this ambition been more visible than in the United Arab Emirates, which launched its national AI strategy in 2017 and has since moved aggressively to institutionalize its goals. Other Gulf states are pursuing similar strategies. Saudi Arabia published its own national AI strategy in 2020, with the Kingdom’s sovereign wealth fund (SWF), the Public Investment Fund (PIF), playing a central role. In 2024, the PIF spearheaded the $100 billion AI initiative “Project Transcendence,” followed in 2025 by the creation of Humain to advance the Kingdom’s AI ambitions and compete with the UAE’s G42. Similarly, Qatar’s SWF, the Qatar Investment Authority (QIA), launched its newest AI company, Qai, at the end of 2025 to strengthen its position as a regional AI hub.

These ambitious AI goals require massive computing power. To meet this demand, the Gulf states are investing heavily in data centers to support advanced AI research, cloud services, and post-oil economic transformation. However, without parallel investments in industries and capabilities, these facilities will not succeed.

The Data Center Boom

As AI and cloud services grow rapidly, they drive rising demand for data storage and processing. Directing a share of the Gulf’s vast capital expenditure into data centers therefore seems an attractive proposition. New plans to build storage and computing facilities are announced constantly, with MENA data center capacity forecasted to triple between 2025 and 2030, from 1 GW to 3.3 GW.

The most notable project is “Stargate UAE,” announced during President Donald Trump’s 2025 visit to Abu Dhabi. The 5GW project is led by G42 as a UAE-based offshoot of Washington’s “Stargate” project, a collaboration involving Softbank, OpenAI, Oracle, and Tahnoun bin Zayed’s tech investment fund MGX. Saudi Arabia announced its new Cloud Computing Special Economic Zone in 2023 and set a national target of 1.3 GW capacity by 2030. In 2024, it secured a $5.3 billion commitment from Amazon Web Services, while China’s Tencent is also planning to expand its Saudi-based cloud services. Elsewhere in the Gulf, Kuwaiti company Omniva is planning to build a 1 GW facility, originally intended for cryptomining but now rebranded as AI. Qatar has taken a different approach, emphasizing colocation models instead of hyperscalers and aligning closely with U.S. and EU data regulations rather than the “soft law” approach favored by Saudi Arabia and the UAE.

These developments reflect significant ambition, even if many projects remain in early planning stages with high completion risk. Building data centers aligns with the shift toward a post-oil economy and appears more pragmatic than many of the region’s overoptimistic gigaprojects, exemplified in Saudi plans to repurpose significant portions of its ailing “The Line” from residential use to data centers. Yet, despite the ostensible pragmatism, critical challenges and vulnerabilities remain.

The Business Risk 

While data centers seem like a solid business model, the return on investment is less than certain. AI companies are currently losing money but promise substantial future profits. However, many experts and tech leaders fear a potential trillion-dollar bubble that could leave few winners standing. The concern not only for AI but also for data infrastructure business is that uncertain demand, combined with rapidly increasing supply, may result in low margins.

In fact, while demand is currently growing, there are already signs of underutilization in data centers, particularly for generative AI and LLMs—the prime application targeted by the Gulf Cooperation Council (GCC) states. Despite being widely used, demand is highly price-elastic, as customers exhibit little willingness to pay for computing capacity. On the supply side, the global data center industry is expanding rapidly. The United States alone is expected to invest a staggering $650 billion into AI in a bid to cement their global industry dominance. These investments present a pronounced risk of a “hog-cycle” dynamic due to a combination of high upfront capital expenditure and long project durations, combined with quick innovation cycles and hardware depreciation. At the same time, many applications are increasingly geographically independent. Developments in the United States and China’s “Eastern Data-Western Computing” initiative illustrate how data workloads can be distributed across vast distances.

For the Gulf, this is a double-edged sword. It offers access to a potentially global customer base but also exposes the region to a global pool of competitors in the largely commoditized market for storage and computing, making it difficult to mitigate industrial rivalry through product differentiation. Moreover, additional global competitors may soon emerge, such as China and Russia, who possess full-stack industries, ample computing capacity, and energy resources. Although both are currently cut off from Western markets, these barriers could shift if geopolitical conditions change. Closer European engagement with Beijing, for example, or a resolution to the Ukraine war and the easing of sanctions on Russia could reshape the competitive landscape.

There are also technological threats to the current data center business model, especially hyperscalers. Just as computing shifted from mainframes to personal computers in the 1980s, storage and processing could again become decentralized. At the same time, computing may become more efficient. Average server utilization is currently an abysmal 12-18 percent. Advances in network design and software could allow companies to use this idle capacity more efficiently. As a result, available computing power may increase dramatically without the need for new data centers, compounding the risk of a future oversupply.

Data Centers as Gulf Economic Policy

Business profitability is not necessarily the main driver of the Gulf’s data center strategy. Even with low margins due to elastic demand and escalating supply, governments may still view the sector as worthwhile if it delivers other socioeconomic benefits. However, the direct effect on local economies is limited, particularly in terms of employment. For example, Meta’s $1 billion data center in Kansas City generated roughly 1,500 jobs during construction and 100 for operations. In fact, the net effects on the local environment and economy are often negative, sparking local resistance.

Data centers can have positive indirect effects on the national economy by stimulating supply chain activity. In the short- to medium-term, this predominantly benefits construction, while high-value components such as advanced electronics and skilled labor continue to be imported. Downstream, data centers may stimulate domestic IT, finance, and other sectors by providing cheaper, lower-latency computing. But while fast, affordable computing is an industry advantage, its overall benefit, given sizeable drawbacks, remains debatable. To achieve broader benefits from complementarities and positive externalities, data centers need to be integrated into a comprehensive industrial policy. In the United States, AI investment, and particularly the focus on LLMs, is crowding out other industrial investment. By contrast, China’s planning views data centers as merely a part of its capability stack, and AI as augmenting, rather than replacing, the real economy. AI as an end in itself only yields mediocre automation.

Perhaps most importantly, data centers may generate important long-term effects. The economic return on investment often surpasses the time horizon of private investors, which would warrant government support. An AI ecosystem built around abundant local computing power may attract highly skilled talent and contribute to the establishment of an innovation hub. Yet the war has underscored a central lesson: the impact of data centers is, at best, complementary to higher-order efforts such as direct investment in R&D and higher education.

With sovereign wealth funds under revenue pressure, defense budgets crowding out Vision spending, and the Gulf’s hard-won reputation for stability under strain, the region can no longer afford to treat data infrastructure as a strategy in itself. Absent genuine integration into a broader innovation ecosystem, one with the R&D funding, educational investment, and indigenous talent to match the ambition, data centers alone will not serve as primary engines of sustained post-oil transformation. The current moment is decisive: the war risks diverting attention from these more abstract, longer-term choices just as the Gulf faces a narrowing window to get the sequence right.

The views and opinions expressed in this article are those of the authors and do not necessarily reflect the views of Gulf International Forum.