The UAE continues to position itself as a regional hub for AI and cloud infrastructure, with large-scale hyperscale investments accelerating across Abu Dhabi and Dubai.
Notable transactions include ADQ’s reported USD 25 billion US data centre joint venture and KKR’s USD 5 billion investment in Gulf Data Hub to support regional expansion. Alongside these develop-ments, government-backed AI initiatives, global technology partnerships and emerging AI campus concepts are reinforcing the UAE’s role as a leading digital infrastructure hub.
But as firms increasingly explore data centre investments, they must also understand their unique due diligence requirements. While these requirements can vary greatly — for example, based on the data centre’s development stage — there are foundational elements that investors must evaluate before making an investment. Explore these considerations in BDO checklist:
BDO’s data centre due diligence checklist
Power Availability
With demand for AI increasing, investors should evaluate a data centre’s power capacity and energy requirements.
- Power supply - Assessing the age and performance of internal equipment, including trans-formers, switchboards, generators, distribution, safety devices, and back-up power supply can help firms determine whether upgrades are needed to handle AI workloads.
- Power capacity - Examine the facility’s current capacity and forecast future needs from in-creasing AI use. Is the data centre equipped to scale operations as AI workloads increase over time?
- Contract terms and structure - What existing energy contracts does the data centre have in place? Ideally, the facility should have a long-term energy contract, which reduces the risk of cost variability and service interruptions.
Cooling capabilities and environmental impacts
AI chips generate significantly higher thermal load than other types of chips and can put tremen-dous strain on local water supplies and infrastructure, negatively impacting the communities where data centres operate — especially those in already water-stressed regions.
- Cooling method - What kind of cooling system does the data centre use? Each method — liq-uid, air, immersion, evaporative, hybrid, or free cooling — has different impacts on the data centre’s ability to accommodate AI’s high-density computing power. Liquid and immersion cooling, for example, are considerably more efficient than air-based systems.
- State of systems - Review the efficiency of a data centre’s cooling equipment to help fore-cast maintenance costs, and in turn, profitability. Firms can request to review power usage effectiveness (PUE) ratios and maintenance logs for greater visibility during due diligence.
- Water consumption - Evaluate a data centre’s water use and any consumption strategies the facility has in place, such as closed-loop liquid cooling systems and alternative cooling flu-ids. Large-scale data centres typically use millions of gallons of water daily due to cooling needs driven by AI workloads, which carry both operational and cost implications.
- Noise pollution - Assess a data centre’s noise pollution from both its day-to-day operations as well as from back-up power generators. As many suburban areas are trying to keep data centres out of their communities due to excess noise and power consumption, it is crucial to identify and address noise pollution concerns before committing to an investment.
Workforce
Data centres often experience high turnover rates due to their high-pressure work environments and aging workforce. Because they are typically built in rural areas, data centres are also more suscepti-ble to worker shortages. PE investors should evaluate a data centre’s workforce needs and assess how they may evolve over time.
- Local talent - Consider the data centre’s local talent pool. Can the available workforce ad-dress the facility’s cabling, equipment maintenance, and operational needs? If not, the data centre may need to house workers from other areas on a temporary basis, mimicking similar arrangements in oil and gas operations, which can become extremely costly. Some organiza-tions are identifying and developing talent via collaborating with local education partners.
- Workforce needs now and into the future - Understand how the data centre’s workforce needs will change as its AI capabilities mature. For example, how many people does the data centre need to employ to maintain its operations? Will the facility require new security personnel? What training investments might be needed? A data centre in the early stages of AI adoption will have vastly different workforce needs than a data centre with advanced AI infrastructure in place.
- Specialised talent - Data centre workers with specialised skills — like advanced AI knowledge — are in high demand, and therefore, may expect higher compensation packages. Talent poaching has also become commonplace. Review the target data centre’s compensation and benefits offerings, as well as broader recruitment strategies to attract and retain top tal-ent. Consider the cost of turnover, availability of talent, and the competition when develop-ing the talent philosophy used at each data centre.
Security
A data centre’s security directly influences the long-term viability and value of investment, as a breach — whether through a back door or a firewall — can result in catastrophic consequences.
- Physical facility protection - Assess a target facility’s defence-in-depth capabilities to man-age access points and their vulnerabilities. Unauthorized data centre access remains a risk, especially for colocation arrangements. A comprehensive security plan should be part of the due diligence process, with contemplation of the facility’s vehicle blockades, cage systems, door hinges, and other protective measures put in place to ensure building security. Data centres that serve higher-risk industries like critical infrastructure, defence, government contractors, or technology providers may need even more comprehensive investments.
- Cybersecurity - Review a data centre’s digital defence capabilities, including the storage and access of identities and surveillance systems, along with incident response capabilities and after-action reports for recent incidents. The data centre’s business model should drive associated control levels and an effective information security management system. A facil-ity’s network monitoring practices and incident response procedures have a direct impact on tenants’ data protection and regulatory compliance.
- Disaster & recovery risks - Consider a data centre’s environmental risk factors and their po-tential impacts on facility infrastructure. For example, is the facility in proximity to a disas-ter zone? What infrastructure does the data centre have in place to withstand a severe weather event?
Regulatory requirements
Data centres are heavily regulated, namely for their resource use and connection with other sectors that handle sensitive information, like healthcare and financial services. Before investing in a data centre, understanding the facility’s regulatory requirements based on its current and prospective tenants will help maintain an accurate risk picture.
- State and local reporting requirements - Assess any state or local rules pertaining to a tar-get location, as they can significantly impact reporting requirements and compliance costs.
- AI & cybersecurity compliance - It is critical to verify that a data center maintains compli-ance with all in-scope regulations. Investors must also ensure that the data centre meets industry-standard cybersecurity requirements, such as ISO 27001 or NIST frameworks.
- Emerging regulations - Anticipate future regulations around alternative energy sources, par-ticularly nuclear power, which is gaining traction as a data centre energy solution. New compliance requirements in this area could impact operational costs and timelines.
- Third-party attestation - To verify that a data centre meets all its necessary compliance re-quirements, request third-party attestation and certifications. SOC 1 and SOC 2 are the baseline documentation for most facilities, though additional certifications may be required depending on the data centre’s tenants. For example, a data centre hosting a healthcare tenant may need documentation to demonstrate compliance with HIPAA.
Real estate leasing
It is important to understand real estate leasing terms to accurately model a target data centre’s cash flow stability.
- Lease terms - Consider prioritising data centres with longer (e.g., 15+ years) and more fa-vourable lease structures. A triple net lease (NNN) is ideal for a data centre, as the tenant pays base rent plus the “net” of operating expenses, including property taxes, building in-surance, and maintenance.
- Tenants - Verify the creditworthiness of existing or prospective tenants to ensure stable, long-term cash flows that justify the acquisition price. Investors must also consider what al-ternative uses exist for the facility should one or more tenants depart.
- Customer concentration risk - How many tenants does the data centre have? Do they have tenants in more than one industry? Tenant base composition can reveal potential cash flow vulnerabilities and customer concentration risk.
How BDO UAE can help
BDO UAE provides partner-led advisory, consulting and due diligence services for investors and oper-ators seeking to enter or expand within the UAE’s rapidly evolving data centre market. Drawing on deep local expertise and global sector insight, we support clients across the full investment lifecy-cle, from feasibility and transaction advisory to financial, tax and operational due diligence.
If you are assessing market entry or expansion opportunities, our team can help you identify risks, uncover value drivers and structure investments with clarity and confidence in a highly competitive and fast-scaling digital infrastructure landscape.
CONTACT OUR TEAM