Dubai 
ParisThe UAE as an AI Hub: When Substance Matters More Than Structure
China’s recent decision to block Meta’s acquisition of AI agent company Manus and to require the deal to be unwound has underlined that geography still matters in artificial intelligence. What began as a review has now been framed explicitly as a response to concerns over technology exports, national security and the relocation of AI activity away from China, and as a signal that regulators are prepared to intervene where offshore structures and re‑domiciliation do not reflect where substantive control and risk continue to sit.
Against that background, the question for jurisdictions that wish to act as AI hubs is whether they are attracting activity on a substantive basis or merely as a legal or regulatory label. The UAE has placed artificial intelligence at the center of its long-term economic agenda. Through the National Strategy for Artificial Intelligence 2031, the creation of a dedicated ministerial portfolio and the establishment of federal and emirate level AI and advanced technology bodies, it has indicated that it intends to play a meaningful role in this field rather than simply adopting technologies developed elsewhere. That direction is reinforced by continuing investment in digital infrastructure and public sector AI programs, particularly in Abu Dhabi and Dubai.
The UAE is also increasingly visible in international conversations about AI risk and governance, including through participation in multilateral initiatives and bilateral engagement with technology supplying states. The UAE is not in the position of Singapore in the Meta Manus narrative, but that story provides a useful lens. It raises the question of when a jurisdiction can credibly present itself as an AI hub and when it risks being seen as a venue for regulatory arbitrage, particularly where activity is relocated in form but not in substance.
From a legal perspective, the UAE’s answer to that question is still developing, but the direction of travel is increasingly clear. There is no single comprehensive AI statute. Instead, AI activity sits within a framework that brings together Federal Decree Law No 45 of 2021 regarding the Protection of Personal Data (“PDPL”), the data protection regimes of the Dubai International Financial Centre (“DIFC”) and Abu Dhabi Global Market (“ADGM”), cyber security and content rules and sector specific regulation in areas such as financial services and healthcare. AI innovation is encouraged, but within a structure that is increasingly shaped by these regimes and by emerging expectations around AI governance.
This article is the first in a three-part series on what that means for AI businesses that choose to base operations in the UAE. The series starts with the UAE’s positioning as an AI hub and the extent to which it offers substance rather than a light touch label. It then considers data and cross border flows, and finally looks at compute, export controls and AI governance, with a particular focus on whether a UAE base can act as a credible organizing center for group level AI risk.
For founders and investors, three questions underpin the analysis across this series.
- The extent to which the UAE provides a platform for substantive AI operations, or primarily a jurisdictional and branding advantage?
- The extent to which UAE-based structures can support multi-jurisdictional data use and governance without creating regulatory friction?
- The extent to which the UAE can credibly anchor decision-making around infrastructure, export controls and AI risk in a manner that withstands scrutiny from regulators, investors and commercial counterparties?
These questions are particularly salient in the context of cross border AI investment and technology flows, where the UAE may serve as a neutral platform for structuring AI‑related operations across multiple jurisdictions. In such scenarios, the credibility of a UAE‑based structure will turn on whether it embodies genuine operational substance, including real decision‑making authority and governance functions, or whether it is instead perceived as an intermediary layer for activities carried out elsewhere. The Meta / Manus unwind illustrates that, for higher value AI transactions, regulators are now willing to look past corporate form and ask where technology, talent and effective control are really located when deciding whether a structure is acceptable.
These considerations have moved beyond the abstract and are now directly influencing how AI‑related investments are structured and negotiated in practice.
