The two largest economies in the Gulf Cooperation Council (GCC), the UAE and Saudi Arabia, have taken a major bet on datacentres to further their economic diversification efforts.
Intuitively, the logic is compelling. Both countries have substantial financial resources, energy supplies and the capacity to mobilise investment at scale. But the regional conflict has exposed a new source of risk. Iranian strikes have damaged datacentre facilities in the UAE and disrupted infrastructure at a Bahrain facility, raising questions about the security and resilience of these investments. The conflict has also sharpened concerns about energy security.
Behind the billions being committed lies a fundamental economic question: Will there be enough paying demand to justify the capital being deployed?
The scale of the ambition is considerable. Yet it is important to distinguish between existing infrastructure and announced ambitions.
| Key metric | Saudi Arabia | UAE |
| Operating capacity (2024) | 274MW | 257MW |
| Capacity under construction (May 2025) | 127MW | 151MW |
| Estimated construction cost (2026–30) | $13m/MW | $11m/MW |
| Leading developers and operators | HUMAIN (Public Investment Fund), DataVolt, stc | Khazna Data Centers (G42), Gulf Data Hub, Moro Hub |
| Major AI infrastructure ambition | HUMAIN’s plans to scale AI infrastructure significantly by 2030 | Stargate UAE, a planned 5GW AI campus |
| Principal economic challenge | Converting ambitious expansion plans into bankable demand | Delivering large-scale projects while managing costs and security risks |
| Sources: Gulf Data Centre Association, S&P Global and Reuters. Figures relate to different reporting periods; announced capacity is not equivalent to operational capacity. The UAE campus plans have been subject to review following regional security developments. | ||
The economic rationale is clear. Demand for cloud computing and artificial intelligence (AI) should grow, while data sovereignty requirements encourage governments and businesses to keep sensitive information within national borders. Datacentres could also attract foreign investment, develop domestic capabilities and support higher-productivity industries.
But forecasts of future demand are not the same as contracted revenues. AI usage may expand rapidly without every operator earning an attractive return. Competition could intensify, prices fall and utilisation rates disappoint. The risk is not necessarily that demand fails to materialise, but that capacity is built faster than profitable demand can absorb it.
Datacentres require substantial upfront investment, consume significant amounts of electricity and depend on equipment that can become economically obsolete well before the buildings themselves reach the end of their useful lives. Investors must recover their capital in a market where technology and computing requirements are changing rapidly.
Cheap energy is not a business model
The Gulf enjoys genuine advantages: access to capital, relatively inexpensive energy, available land and government support. Its geographical position also offers opportunities to serve markets across the Middle East, Africa and South Asia.
But cheap electricity does not guarantee low overall costs. Reliable power supplies, transmission infrastructure, cooling and connectivity all require investment. Extreme temperatures and water scarcity add further economic constraints.
There is also an opportunity cost to subsidised inputs. Electricity supplied below its economic cost may improve an operator’s profitability while shifting costs onto the public sector. The same applies when governments provide land, guarantees or infrastructure on preferential terms.
Such support can be justified if it generates wider economic benefits. But those benefits must be demonstrated, not assumed. Commercial profitability and economic value are not the same thing. A datacentre might struggle to earn an adequate return for its investors yet still benefit the wider economy by attracting foreign investment, developing domestic expertise or improving productivity in other industries. Conversely, a profitable operator does not necessarily represent a good investment for the country if its returns depend heavily on public subsidies while the wider economic benefits remain limited.
The key question, therefore, is whether these wider benefits justify the resources committed. Public capital, subsidised energy and infrastructure all have alternative uses. Their true economic cost is not simply what the government spends, but the value of what it forgoes by choosing one investment over another.
This is where the Gulf’s development ambitions create a particular risk. Datacentres are expected not only to generate revenue but also to attract foreign investment, create skilled employment and stimulate domestic technology industries. These are legitimate objectives. Yet they can encourage policymakers to measure success in announced investment and installed capacity rather than economic returns.
The political rewards of announcing a major project arrive immediately; the costs of underused infrastructure emerge years later. Sovereign wealth funds can sustain investments that private investors might reject, but public financing does not eliminate economic risk. It merely changes who bears it.
Investment must earn its keep
The strongest projects will be those backed by credible customers rather than optimistic forecasts. Long-term contracts with cloud providers, government agencies and large businesses can reduce revenue uncertainty. Phased construction can limit exposure by allowing capacity to expand alongside demand.
Investors must also look beyond captive domestic markets. Data-residency rules and government procurement may provide an initial customer base, but sustainable returns will depend on competing for regional and international business on cost, reliability and connectivity.
Policymakers, meanwhile, should assess projects by their contribution to the wider economy. Do they attract additional private investment, develop valuable domestic capabilities and improve productivity elsewhere? Or do they primarily absorb public capital and subsidised energy?
The likely outcome is mixed. Facilities supported by firm demand, disciplined investment and genuine cost advantages can become productive assets. Those built on speculative AI forecasts and the assumption that customers will eventually arrive face a greater risk of disappointing returns.
The Gulf has the resources to build a major data-centre industry. But the ability to finance infrastructure should never be confused with the ability to make it pay.
The question is not whether the Gulf can afford to build the future. It is whether the future will pay for what is being built.
Amír Khan is a macroeconomist and strategist with a lengthy track record of advising international banking groups, including MUFG, Saudi National Bank and Mizuho, on global macro and financial markets.
