Gulf AI Infrastructure Spending Reaches $8.4 Billion

Key Points

1. GCC investment in AI infrastructure reached $8.4 billion in 2025 as deployment accelerated across governments and businesses.
2. Autonomous AI agents introduce risks involving permissions, context and inappropriate decisions, even without a conventional cyberattack.
3. The investment surge makes embedded governance critical as AI gains access to financial, healthcare and government systems.

The latest

Artificial intelligence infrastructure spending across the GCC reached about $8.4 billion in 2025, part of a global total of $318 billion, as Middle Eastern governments and businesses expanded computing capacity, data centers and AI-enabled services. The shift from experimental tools to autonomous systems is placing cybersecurity at the center of deployment decisions, because AI agents can access software, retrieve data and execute tasks rather than merely generate information or recommendations.

Details

• Investment acceleration: Global AI infrastructure spending rose from $153 billion in 2024 to approximately $318 billion in 2025, effectively doubling within a year. The UAE and Saudi Arabia have made AI central to economic diversification, directing investment toward cloud computing, semiconductors, sovereign infrastructure, data centers and digital public services.
• New risk model: Zheng Li, co-founder, GVP and chief product officer at Abu Dhabi-based Ansen Technology, said agentic AI requires organisations to assess whether a system is authorised to perform a specific action in a particular context and at a particular time, rather than focusing only on malicious users, files or processes.
• Wrong decisions: An AI agent may hold legitimate system access yet make an inappropriate decision because it misinterprets instructions, receives incomplete information or has excessive permissions. Li said the risk may arise when a system operates as designed but reaches the wrong decision, rather than when an attacker compromises it.
• Access controls: Li advocated applying least-privilege principles to every AI agent, continuously evaluating its actions and maintaining detailed audit trails. Permissions should reflect the task, data and potential consequences, while higher-risk decisions should require human approval instead of granting agents broad access across an entire workflow.
• Infrastructure expansion: Gartner forecasts Middle East and North Africa IT spending will reach $169 billion in 2026, up 8.9% from 2025. Software spending is projected to increase 13.9% to $20.4 billion, while expenditure on data-center systems is expected to rise 37.3% as computing requirements grow.
• Embedded governance: Li said security leaders should treat each AI agent like an employee with an identity, defined objectives, permissions and potential consequences. Governance must be incorporated into system architecture before deployment, establishing trusted boundaries without preventing autonomous operation.

Background

Agentic AI differs from conventional tools by acting directly within connected systems. Its expansion into critical infrastructure, finance, healthcare and government services increases both the number of accessible systems and the potential consequences of an incorrect autonomous action.

What’s next

The next measurable indicator is 2026 spending: Gartner expects regional data-center systems expenditure to rise 37.3%, alongside total MENA IT spending of $169 billion. Organisations will face decisions over permission limits, human approval thresholds and audit requirements as new autonomous systems enter operation.