A USD 5–12 billion feasibility study across three UAE sites targets up to 1,280 MW of sovereign AI capacity—positioning the programme at less than half the global cost benchmark amid a USD 6.7 trillion global data center investment cycle.
Dubai, UAE — As the global digital infrastructure sector accelerates into what analysts project as a multi-trillion-dollar expansion cycle, Asprofin Bank Corporation is advancing a feasibility-stage sovereign AI infrastructure programme that could reshape how hyperscale computing is financed, built, and deployed in the Gulf region.
The initiative developed in collaboration with IDC Technologies and formally presented for preliminary review to Digital Dubai covers three strategic sites across Hatta, Jebel Ali, and the Mohammed bin Rashid Al Maktoum Solar Park. The proposed development envisions between 750 MW and 1,280 MW of sovereign AI computing capacity, with a total estimated investment ranging from USD 5 billion to USD 12 billion over a 10 to 15 year horizon.
While still under feasibility assessment and not yet approved for execution, the programme reflects a growing shift toward sovereign-controlled artificial intelligence infrastructure in the Middle East.
A Global AI Infrastructure Cycle Worth Trillions
The timing of Asprofin Bank’s exploration aligns with what major research institutions describe as an unprecedented infrastructure buildout.
McKinsey estimates that global data center demand could expand to 219 gigawatts by 2030, effectively tripling current capacity, with artificial intelligence workloads accounting for up to 70% of incremental demand. The consultancy also projects that the broader data center investment cycle could reach approximately USD 6.7 trillion by the end of the decade.
Other global forecasts reinforce the scale of the shift. Goldman Sachs expects AI-driven electricity consumption to increase by more than 220% by 2030, reaching approximately 1,350 terawatt-hours. Deloitte further projects that AI inference workloads will account for roughly two-thirds of total compute demand by 2026, signaling a long-term structural requirement for always-on, high-density infrastructure.
Within this global expansion, the Gulf region is emerging as one of the fastest-growing digital infrastructure hubs.
Market data from Mordor Intelligence values the UAE data center industry at approximately USD 1.74 billion in 2026, with projections reaching USD 3.86 billion by 2031. This growth is being accelerated by regulatory frameworks such as the UAE’s Personal Data Protection Law, which is increasing demand for in-country data storage and sovereign compute solutions across regulated industries including banking, healthcare, and government services.
A Cost Structure Designed to Challenge Industry Norms
One of the most closely watched elements of the Asprofin Bank feasibility study is its capital efficiency model.
Global hyperscale data center construction currently averages around USD 10–12 million per megawatt. In the case of AI-optimized facilities, costs can exceed USD 20 million per megawatt due to higher energy density, advanced cooling systems, and specialized hardware requirements.
By comparison, Asprofin Bank’s internal modelling suggests its proposed nano-scale modular architecture could deliver capacity at an estimated USD 4–8 million per megawatt—potentially representing a 30% to 65% reduction versus prevailing industry benchmarks.
According to Asprofin Bank's feasibility modelling, the proposed architecture could offer a significantly lower capital cost profile than many conventional hyperscale AI infrastructure projects, although the assumptions remain subject to technical and commercial validation.
This cost advantage becomes even more notable when compared with major regional hyperscale developments, including the Stargate UAE initiative. Early-stage projections for similar AI infrastructure projects indicate capital intensity levels that can exceed USD 30 million per megawatt in initial phases, driven by GPU density requirements and large-scale energy integration costs.
Engineering Model Built for Speed, Efficiency, and Sovereignty
At the core of Asprofin Bank’s feasibility concept is a modular infrastructure architecture designed to compress both construction timelines and long-term operating costs.
The first pillar is a factory-prefabricated “Nano Center Module” system. These containerized units, designed to deliver up to 500 kW per module, are engineered for rapid deployment—reducing traditional build timelines from 18–24 months to approximately 8–12 weeks per module.
The second pillar is an advanced cooling system designed to significantly reduce energy consumption. Industry-standard Power Usage Effectiveness (PUE) typically ranges between 1.3 and 1.5 for hyperscale facilities, while global averages sit between 1.55 and 1.60 according to the Uptime Institute. Asprofin Bank’s design targets a PUE of 1.15 or lower, which—if achieved—would materially reduce long-term operational energy costs.
The third pillar focuses on renewable energy integration. The programme is expected to anchor a large portion of its operations at the Mohammed bin Rashid Al Maktoum Solar Park, targeting more than 80% renewable energy usage across workloads. This positions the project within a growing institutional demand for ESG-aligned infrastructure assets.
Additional design considerations include underground sovereign-grade infrastructure at the Hatta site, featuring multi-layer reinforced construction, electromagnetic shielding, and thermally controlled environments intended for high-security AI and quantum computing workloads.
A Three-Site Strategy Covering the Full AI Value Chain
The proposed architecture is structured to serve distinct but interconnected layers of AI infrastructure demand across the UAE.
In Hatta, underground vault infrastructure is designed for sovereign workloads requiring maximum physical and digital security. These facilities are intended to support high-sensitivity compute environments, including advanced cryptographic and quantum computing applications.
In Jebel Ali, the focus shifts toward carrier-neutral hyperscale connectivity. Located within the JAFZA free zone, this site is designed to integrate with major global cloud providers such as AWS, Microsoft Azure, Google Cloud, and Oracle, creating a commercial interconnection hub for enterprise clients.
At the Mohammed bin Rashid Al Maktoum Solar Park, the programme envisions a large-scale green AI supercluster powered by renewable energy. This site is designed to support AI training and inference workloads at scale, leveraging high-density GPU clusters and next-generation networking fabrics.
Together, the three-site model is intended to create a vertically integrated sovereign AI ecosystem spanning security, connectivity, and compute-intensive processing.
Positioning Within a Rapidly Expanding Competitive Landscape
The feasibility study enters a market already defined by significant hyperscale investment commitments.
Microsoft has announced plans exceeding USD 15 billion in UAE digital infrastructure investment through 2029, while Oracle has launched its first OCI Supercluster in the Middle East. In parallel, multiple global technology firms are expanding regional cloud and AI capacity to meet accelerating demand.
Despite this influx of capital, industry analysts continue to highlight a structural supply gap between commercial hyperscale infrastructure and sovereign-grade compute requirements.
Asprofin Bank’s positioning targets this gap directly—focusing on regulated-sector workloads, national data sovereignty requirements, and energy-optimized AI infrastructure that can operate at lower cost intensity and higher deployment speed.
With sovereign wealth funds in the Gulf investing an estimated USD 66 billion into AI and digital transformation initiatives in 2025 alone, and global occupancy rates for data centers projected to exceed 95% in certain markets, demand fundamentals remain strong.
A Feasibility-Stage Vision, Not a Final Commitment
Asprofin Bank emphasizes that the programme remains in feasibility assessment and does not represent a final investment decision, capital allocation, or binding development commitment.
However, the study reflects a broader structural shift in how financial institutions, governments, and infrastructure developers are approaching AI-era compute demand—treating data centers not merely as real estate assets, but as strategic national infrastructure.
If realized, the proposed UAE programme would represent one of the most capital-efficient sovereign AI infrastructure models yet explored in the region, combining modular construction, renewable energy integration, and sovereign security architecture across a multi-site deployment strategy.
About Asprofin Bank
Asprofin Bank Corporation is an international financial institution specializing in the structuring, financing, and development of large-scale infrastructure programmes. This announcement reflects an ongoing feasibility exploration and does not constitute a binding commitment, capital allocation, or securities offering.
Media Contact:
Q. Lin — PR Director
Asprofin Bank Institutional Communications & Investor Relations
Dubai, United Arab Emirates
Email: qlin@asprofinbank.org
