Post

The UAE's Sovereign AI Compute Play: A Template for National AI Infrastructure

The UAE's Sovereign AI Compute Play: A Template for National AI Infrastructure

On 20 July 2026, e& UAE (formerly Etisalat) and Core42 — the sovereign cloud and AI infrastructure arm of G42 — announced a partnership to launch Sovereign AI Compute. The service promises on-demand GPU infrastructure where data never leaves UAE borders, with zero egress fees and no capital expenditure required from customers. It is a direct challenge to the hyperscaler model and a signal that the UAE intends to own its AI destiny.

The Announcement

Sovereign AI Compute delivers GPU clusters — including NVIDIA H100 and B200 Blackwell systems — as a managed service inside the UAE. Customers get dedicated compute, high-speed interconnects, and full control over their data residency. The “no CapEx, no egress” structure is particularly striking: enterprises and government entities can spin up AI workloads without building data centres, and they can move data in and out of the service without incurring the punitive transfer fees that characterise most public clouds.

The infrastructure sits inside e& UAE’s existing data centres, operated and secured by Core42. It is available immediately to UAE-based organisations and is expected to expand to other markets in the MENA region.

What Makes It “Sovereign”

The word “sovereign” does heavy lifting here. It means three things:

  1. Data sovereignty — all processing and storage occurs within the UAE’s national borders, subject to UAE law and regulation.
  2. Operational sovereignty — the infrastructure is owned and operated by UAE entities (e& and Core42), not by a US or Chinese hyperscaler.
  3. Supply chain sovereignty — the hardware supply and software stack are managed end-to-end by Core42, which runs the Condor Galaxy supercomputer network and has deep relationships with NVIDIA, AMD, and Cerebras.

This is not simply a resale of AWS or Azure inside the UAE. It is a purpose-built national AI cloud, designed to meet the security requirements of sectors like defence, finance, healthcare, and government — where data leaving the country is legally or strategically unacceptable.

The Bigger G42 Picture

Core42 is one piece of a much larger G42 machine. The Abu Dhabi-based AI holding company has been systematically assembling a full-stack national AI capability:

  • InceptionClaw, launched in May 2026, is G42’s enterprise AI super assistant — a multimodal agent platform for workflow automation, document intelligence, and decision support.
  • Inception42, a joint venture with Microsoft, is moving agentic AI from pilot projects into production operations across UAE government and enterprise customers.
  • The G42-NVIDIA Condor Galaxy network already provides some of the largest publicly accessible AI compute clusters outside the US.

Sovereign AI Compute completes the picture: it provides the infrastructure layer that InceptionClaw and Inception42 run on, creating a vertically integrated national AI stack — compute, platform, and applications — all under UAE control.

The US–UAE AI Relationship

Days before the Sovereign AI Compute announcement, the United States upgraded the UAE’s export classification, easing restrictions on advanced AI hardware transfers. Core42 CEO Kiril Evtimov described the timing as significant, telling media that the move “strengthens AI security” for both nations and reflects US confidence in the UAE’s governance framework.

The upgrade is widely seen as a recognition that the UAE is a trusted intermediary in the global AI supply chain — capable of hosting cutting-edge hardware and managing the associated security protocols without becoming a leak point. It also positions the UAE as a neutral ground between the US and China, a status that the country has cultivated aggressively through its foreign policy and technology partnerships.

For US AI companies, the UAE is increasingly the gateway to the Middle East, Africa, and South Asia — markets where American hyperscalers face regulatory friction, data residency requirements, and competition from Chinese alternatives like Huawei Cloud and Alibaba Cloud.

Lessons for Africa

The Sovereign AI Compute model is directly relevant to Kenya and other African countries pursuing digital sovereignty.

Kenya’s Digital Transformation Programme and the forthcoming National AI Strategy both emphasise data localisation, digital public infrastructure, and domestically controlled compute. The challenge has always been execution: building sovereign cloud infrastructure is capital-intensive and requires specialised talent that is scarce locally.

The UAE model offers a pragmatic template:

  • Partner with a national telecom — e& already had the data centres, fibre, and regulatory permissions; Core42 brought the GPU expertise and supply chain. Kenya could replicate this with Safaricom or Telkom Kenya as the infrastructure anchor.
  • Start with on-demand, not build-to-suit — Sovereign AI Compute does not require customers to fund data centres. A Kenyan sovereign AI cloud could offer the same: GPU cycles as a service, removing the CapEx barrier for local startups, universities, and government agencies.
  • Zero egress as a policy statement — data transfer fees are a hidden tax on digital sovereignty. Committing to zero egress is both a commercial differentiator and a sovereignty signal.
  • Bundle platform services — just as G42 bundles InceptionClaw on top of its compute, Kenya could layer local LLMs (like the East African LLM initiatives), AI translation for Swahili and indigenous languages, and government service automation on top of a sovereign compute layer.

Kenya’s Window

Kenya is already further along than most African countries. The Kenya Cloud initiative, the Ajira Digital programme, and the Konza Technopolis data centre campus provide foundations. What is missing is the integrated vision — compute, platform, and applications — that the UAE has executed with G42.

The UAE’s Sovereign AI Compute shows that it is possible to build national AI infrastructure that is both secure and commercially viable. The question for Kenyan policymakers is not whether to do it, but whether they are ready to move with the same speed and strategic clarity.

The UAE placed its bet. Kenya should take notes.

This post is licensed under CC BY 4.0 by the author.