Dell: AI in Saudi Arabia enters production test, not experiment

In Las Vegas, on the sidelines of Dell Technologies World 2026, the question is no longer about the size of investment in infrastructure or the speed of building data centers, but about the difference the Kingdom can make in a global market shifting from AI experiments to production within enterprises. This shift in tone is what Michael Dell, chairman and CEO of Dell Technologies, summarized in a private interview with Al-Sharq Al-Awsat: "We see in Saudi Arabia a deep commitment to modernising the Kingdom," pointing to large energy resources, Dell’s work with HPE and other companies, and a regional facility that aggregates these capabilities and builds infrastructure for customers in the region.

Dell described Saudi Vision 2030 as "extremely ambitious" and the AI ambition underneath it as "impressive". Yet the real test lies between ambition and operational infrastructure. Mohamed Amin, senior vice-president for Central and Eastern Europe, the Middle East, Turkey and Africa at Dell Technologies, frames this shift clearly: the biggest obstacle for organisations in Saudi Arabia and the Gulf as they move from AI experiments to production is not a single factor but an inter-connected system that includes infrastructure, governance, skills, cyber-resilience, cost and operating models, though he counted "data readiness" as the first hurdle.

Data before the model

This point appears fundamental in Dell’s reading of the Saudi stage. The company’s AI and innovation survey indicates that 96 percent of Saudi organisations now view AI as a core part of their strategy. Yet, despite its importance, this metric does not mean the path to production has become easy. Many organisations still operate on legacy, fragmented systems, distributed data, inconsistent governance, and limited access to reliable real-time data.

Amin defines "AI-ready data" as data that is "discoverable, governed, trustworthy and usable by AI systems in real time". This definition shifts the discussion from a narrow technical question to an organisational one: does the organisation know where its data resides, who can use it, and can it be trusted when fed into a model or intelligent agent?

In Saudi banking, this could mean linking customer, transaction and risk data across different environments while maintaining compliance and governance. In hospitals, it concerns securely organising clinical and imaging data so AI can assist diagnosis or optimise workflows without jeopardising patient privacy. In government agencies, it means unifying citizen and process data while preserving sovereignty and security controls. In energy companies, it entails combining operational data, sensor feeds and geospatial data to support predictive maintenance and performance optimisation.

Local operating economics

As AI moves to more sensitive and continuous workloads, another question arises: when does a private or on-premise infrastructure become more suitable than public cloud. Amin does not present the issue as a stark choice; he sees public cloud remaining valuable for experimentation, flexibility and rapid access to AI services, but adds that there is a stage where a controlled infrastructure is "strategically better", especially when workloads involve national or financially sensitive data, or when response-time requirements are critical.

This aligns with Dell’s presentation of "Deskside Agentic AI", a solution that aims to run some AI agents locally on high-performance workstations instead of relying entirely on cloud APIs. The company says the solution can, in some cases, break even with cloud API costs within three months and cut spending by up to 87 percent over two years. Amin reads these figures more broadly: Saudi technology leaders should evaluate AI economics "over the full lifecycle, not just by focusing on upfront infrastructure cost". Cloud may appear attractive at launch, but it can become more expensive when running generative or agent workloads continuously across a large organisation.

Cooling, energy and sovereignty

This discussion cannot be separated from cooling and energy, as AI raises rack density and power requirements inside data centres, making cooling a strategic as well as an operational factor. The Dell PowerCool C7000’s ability to support water at temperatures up to 40 C means data centres can operate more efficiently in hot climates while reducing reliance on energy-intensive cooling. In Saudi Arabia, where the public and private sectors invest in sovereign AI infrastructure, Amin notes that cooling "is no longer just an operational issue", but is tied to scalability, energy efficiency and long-term viability.

Cyber-resilience is part of AI readiness, as an intelligent system cannot be reliable if its data can be corrupted, its models can be compromised, or its infrastructure cannot recover. Amin points out that an AI system "is no more reliable than the data and models it runs on", and that a cyber-attack that spoils data or damages a model can have serious consequences. Here Dell offers tools such as "Cyber Detect", which it says can detect data corruption from ransomware attacks and pinpoint the last known clean version with high accuracy.

What that means for organisations in the region

The picture that emerges from Michael Dell’s response on Saudi Arabia and Mohamed Amin’s view of the region sketches a different upcoming phase. The Kingdom is not entering the AI race solely from a consumption or experimentation angle, but from building institutional capability. Yet true capability will not be measured merely by the number of data centres or investment size, but by organisations’ ability to ready their data, choose where to run workloads, control costs, protect models and data, and expand usage without losing control or governance. For a Gulf reader, the practical message is single: infrastructure investments are necessary but not sufficient, and the next test lies in data, governance and daily operating economics.