What Building AI Infrastructure in the UAE Actually Looks Like, in the Words of the People Doing It
This was supposed to be a Founder Friday interview. We reached out for one, on the theme of what actually differs about building AI infrastructure in the UAE compared with elsewhere. The conversation did not close in time for today's slot. Rather than run a piece built on assumptions or dressed-up desk research, we did something more useful. Three people who are actually building AI infrastructure in the UAE, at three different layers of the stack, have already said on the record what makes the job here different. We pulled their words together, checked every quote against its original publication, and laid them side by side.
Why Sovereignty Is the Word Everyone Uses
Every figure in this roundup uses some version of the word sovereign to describe what they build. That is not a coincidence of branding. It reflects a specific commercial pressure. In an interview with Inc. Arabia, AVELIN AI founder and CEO Yury Akinin explained the underlying dynamic in blunt terms. Most AI technology, he said, sits in the hands of large technology companies based in the United States and China, and more than 120 countries have already introduced regulations pushing for local AI development. That regulatory pressure, not marketing, is what makes sovereign infrastructure a genuine commercial category rather than a slogan.
The Founder View: AVELIN AI's Yury Akinin
Akinin, who previously co-founded and exited life sciences AI company Quantori at a valuation above 100 million dollars, closed a 3.7 million dollar pre-seed round for AVELIN AI in July 2026, a raise DSD covered at the time. Speaking to Inc. Arabia, he framed the practical difference this way: an open-source model on its own is not a platform, since it does not connect to business applications and usually falls short of the intelligence frontier labs offer, yet roughly half of all business activity sits in regulated sectors that need exactly that kind of on-premises deployment. AVELIN's answer is an enterprise license that installs fully air-gapped on a client's own cluster, with no data leaving the premises, something Akinin points out neither OpenAI nor Anthropic can offer since their models remain closed.
On what makes the UAE specifically workable for a startup like his, Akinin was equally direct in the same interview. The intensity of interest from the region's leadership gives young AI companies visibility and momentum that is hard to manufacture elsewhere, though he noted that same momentum is drawing in more competitors, most of which build application layers that end up needing his infrastructure rather than competing with it.
The Infrastructure Major View: Core42's Talal M. Al Kaissi
One layer up the stack sits Core42, the G42 company that runs sovereign cloud and high-performance compute infrastructure across the UAE and beyond. In comments to Inc. Arabia's Gamechangers: AI feature, interim CEO Talal M. Al Kaissi described the central problem Core42 is built to solve as giving governments and regulated industries access to frontier-scale AI without asking them to give up control of their data or infrastructure. Its stack includes the Core42 Sovereign Public Cloud, built on Microsoft Azure with more than 173 sovereign technical policies embedded directly into the architecture, and the Core42 AI Cloud, which gives customers on-demand access to NVIDIA, AMD, Cerebras, and Qualcomm accelerators deployable within minutes at national scale.
Al Kaissi's read on the region's long-term position is the clearest articulation in this roundup of what is actually different about building here. Over the next decade, he said, the MENA region will be among the few places in the world where a company can deploy sovereign-grade AI, access frontier-level compute, and still stay plugged into global capital and technology flows. That combination, sovereignty without isolation, is the specific bet the UAE's AI infrastructure ecosystem is making.
Core42 Sovereign Public Cloud: built on Microsoft Azure, more than 173 sovereign technical policies embedded in the architecture Core42 AI Cloud: silicon-agnostic access to NVIDIA, AMD, Cerebras, and Qualcomm accelerators UAE customer base includes Abu Dhabi's Department of Government Enablement, First Abu Dhabi Bank, and Alef Education Core42's Maximus-01 supercomputer reached a top-20 global ranking on the TOP500 list Core42 has expanded operations into France, Italy, and the United States
The Telecom Infrastructure View: du's Jasim Al Awadi
The third layer is domestic telecom infrastructure. du's National Hypercloud, built on Oracle Alloy, positions the operator as what its own materials call a sovereign hyperscaler rather than a cloud tenant of a foreign provider. In comments reported by MIT Sloan Management Review Middle East, du Chief ICT Officer Jasim Al Awadi described a recent sovereign industrial AI platform launch, developed with Bosch SDS, as a way to combine du Tech's sovereign cloud and digital infrastructure with industrial expertise so organisations can turn operational data into actionable intelligence while maintaining security, transparency, and control.
Al Awadi has made a similar point in the context of government-linked enterprises. When du migrated Dubai Taxi Company's core systems onto the National Hypercloud, du CEO Fahad Al Hassawi described the project as creating a blueprint that other government-linked organisations across the mobility and logistics sectors could reuse, built specifically to satisfy the UAE's Cloud First mandate and full data sovereignty requirements, a standard that differs materially from what global hyperscalers based outside the UAE can guarantee.
What Is Actually Different, Put Together
Read side by side, these three voices are not describing three unrelated companies competing for the same customers. They describe three layers of one stack, each responding to the same underlying pressure. Akinin's answer is about model deployment: give regulated enterprises a frontier-level model they can run entirely on their own hardware. Al Kaissi's answer is about compute and cloud: give governments silicon-agnostic access to the world's leading accelerators without losing control of the data that touches them. Al Awadi's answer is about the network layer underneath both: give UAE-based organisations a domestically hosted cloud that satisfies a national Cloud First mandate no foreign hyperscaler can fully meet on its own.
None of the three frames this as a story about cheaper GPUs or faster permits, the kind of advantage founders sometimes assume defines the UAE's pitch. The consistent theme across all three is regulatory and political alignment: a national mandate for data sovereignty, government-level investment and visibility, and infrastructure built specifically to let regulated industries adopt AI without triggering compliance problems elsewhere. That is a structurally different pitch from the one most AI infrastructure hubs make on cost or raw compute density alone, which is worth understanding for any GCC founder deciding where to build.
What This Means for Founders Building Here
- If your product touches regulated data: the sovereignty argument these three figures make is not abstract. Financial services, healthcare, and government buyers in the UAE increasingly ask for on-premises or nationally hosted deployment as a baseline requirement, not a differentiator.
- If you are building an application layer: Akinin's point about competitors becoming clients is a genuine pattern. Infrastructure providers at every layer of this stack are actively positioning application-layer startups as customers rather than competitors.
- If you are evaluating where to host, the UAE's pitch is regulatory and political alignment more than raw price. Founders comparing the UAE against other hubs should weigh compliance and government procurement access alongside cost per GPU hour.
READ ALSO
