e& enterprise has announced a sovereign AI inference platform in the UAE, timed to GITEX GLOBAL 2025 in Dubai. The offer runs on Intel’s Gaudi 3 hardware with Dell’s infrastructure and is operated locally by e& enterprise. The pitch is simple: keep data in-country, hit real-time speeds, and meet residency rules across government and sensitive sectors. If you’re a CIO trying to ship AI in production without tripping over compliance, this is squarely aimed at you.

What e& enterprise actually launched

e& enterprise is rolling out an in-country, “sovereign” inference stack for enterprises and public agencies. It’s positioned as the first dedicated inference solution of its kind in the region.

Inference platform, not training cluster
Operated and supported locally by e& enterprise
Built on Intel Gaudi 3 plus Dell infrastructure (incl. Dell AI Factory)
Data stays in the UAE; compliance is a core design choice

The company is leaning into a reality most IT leaders know: shipping AI at scale is more about dependable inference than raw model training. By keeping the stack on-shore and managed by a single accountable operator, the platform aims to reduce latency and regulatory overheads for production apps in sectors like healthcare, finance, energy and telecom.

Hardware and software: the stack at a glance

The platform pairs Intel Gaudi 3 accelerators with Dell’s scalable systems. e& enterprise runs the service, integrating with customer environments and providing the compliance wrappers.

Intel Gaudi 3 for AI-optimised inference performance and TCO
Dell infrastructure, including the Dell AI Factory offer
Inference endpoint software, with support for multiple open-source LLMs
Deployable on-prem or in cloud footprints managed in-country

The Gaudi 3 + Dell combo gives customers a familiar enterprise backbone with the cost dynamics of an inference-first GPU/accelerator. Intel says this enables “inference-as-a-service” with model flexibility and better TCO, while e& enterprise wraps it in local operations and SLAs.

Why sovereignty and residency are front and centre

Sovereign AI isn’t just a buzzword in the Gulf—it’s becoming table stakes. e& enterprise’s message: data never leaves national borders, and the platform is aligned to residency and regulatory frameworks.

Data residency and sovereignty mandates are tightening
Vendor cites forecasts that inference will dominate AI compute by ~2030–2032
Gartner is quoted as projecting three-quarters of organisations will mandate data sovereignty for AI by 2026

We’ve seen the same direction across the UAE this year. For context on local moves around sovereign cloud and in-country AI, see our coverage of du’s National Hypercloud and UAE data-sovereignty shifts. These help frame where e& enterprise’s inference play fits.

Who it’s for, and what jobs it does

The launch is aimed at CIOs, CTOs and Chief Data Officers who are trying to move from pilots to stable production services without breaking compliance.

Government services needing low-latency citizen apps
Hospitals and insurers with sensitive medical data
Banks and fintechs under strict regulatory control
Energy and telecom operators with real-time workloads

In practice, this means standing up APIs and agents that answer citizens, process documents, route tickets, summarise calls, or score risk—at scale, in real time—with logs and datasets stored on-shore. The platform’s “single accountable partner” pitch also matters for procurement: one operator to integrate, monitor and support across environments.

What is “sovereign inference”?

Running AI models inside national borders with full control over data storage, access and operations, focusing on production inference rather than model training. e& enterprise’s platform keeps workloads and data in-country.

Is this only for government?

No. It’s built for government and enterprise across sectors such as healthcare, financial services, energy and telecoms—anywhere latency, scale and compliance matter.