Energy is the only sector in this market where artificial intelligence appears on both sides of the ledger at once.
It is a demand driver, because the data centres being built to run models need power and water and a grid connection. It is also an operating tool, because the same models forecast load, predict failures and sequence restoration inside the network that serves them.
In Dubai both sides sit inside one organisation, and that organisation was given a Chief AI Officer by the government before most private companies here had finished the argument about who owns AI.
What the public side has already committed to
In June 2024 the Crown Prince of Dubai appointed 22 Chief AI Officers across Dubai government entities under the Dubai Universal Blueprint for Artificial Intelligence. DEWA is one of them, alongside Dubai Police, the Roads and Transport Authority and the Department of Economy and Tourism.
DEWA operates Digital DEWA as its digital arm, with the stated ambition of becoming the world's first digital utility using autonomous systems for renewable energy and storage, and describes its use of the internet of things, artificial intelligence, nanosatellites and data analytics for efficiency, predictive maintenance and customer service.
The energy targets around it are published. Clean energy is more than 21.5% of Dubai's mix, expected to reach 36% by 2030, against the Dubai Clean Energy Strategy 2050 goal of 100% clean energy production capacity by 2050. The Mohammed bin Rashid Al Maktoum Solar Park is being expanded beyond 8,000MW by 2030.
For a supplier or a private operator, that is not background. It is the specification the buyer is working to.
The three use cases that pay, in order
Predictive maintenance. High-value rotating assets, transformers, pumps, turbines and compressors. The data is operational technology telemetry the company already collects, the failure history exists, and the number is unplanned outage hours or maintenance cost per asset. No personal data is involved, which means the governance burden is a fraction of what it would be in finance or health.
Generation and load forecasting. In a grid absorbing a growing share of solar, forecast error is money. This is a well-understood problem with published methods and a clear benchmark, which makes it one of the few AI projects where a board can be told in advance what good looks like.
Outage prediction and restoration sequencing. Where the customer experience actually lives.
Notice what is not on the list. Generative assistants for the corporate functions are useful and cheap and belong in the register, and they are not the reason a utility hires an AI owner.
The operational technology problem
Energy is the sector where the data problem is different in kind rather than in degree.
The information technology estate and the operational technology estate were built by different teams, on different networks, under different security rules, and in many cases with a deliberate air gap between them. That separation exists for good reasons and it is the single largest obstacle to every use case above.
An AI owner in this sector who does not understand why the operational technology team will not simply hand over the historian will fail in month two. This is the clearest example in any sector of why the seat needs someone who has done the work before rather than someone who has read about it.
The governance position
There is no Gulf energy AI regulation as at August 2026. The federal Personal Data Protection Law applies to the customer side: billing, consumption data, service interactions, and any model that decides something about a named account holder.
The operational side sits largely outside the personal data perimeter and inside safety and critical infrastructure practice, which in this sector is already mature. The useful move is to write the register in two halves, put the customer-facing systems through assessments, and put the operational systems through the safety and change control process that already exists rather than inventing a second one.
Suppliers into government-owned utilities have one more layer. On 20 January 2025 the Dubai Centre for Artificial Intelligence launched the Dubai AI Seal, and it is treated as a prerequisite for companies wanting to be selected as partners on Dubai and UAE government projects. If you sell AI into a Dubai government entity, that certification is a commercial gate rather than a compliance nicety.
What the seat holds in energy
The two-half register. Assessments on the customer half. A vendor and hosting position, which in this sector runs into sovereignty questions faster than in most. A working relationship with the operational technology function that survives disagreement. The staff policy. And one use case with a number that already appears in the operations report.
In a private generator, a distribution utility or an oil and gas services business here, that is a two or three day a week seat for the first twelve to eighteen months. It becomes full time when the company starts operating its own models on its own assets at scale, which is the point at which it also knows what to ask for.
Nothing on this page is legal or regulatory advice.
Where to go next
For the procurement gate, read AI and Dubai government suppliers. For the order of work, read AI transformation in the UAE.
If you have run AI inside a Gulf utility or energy operator, claim a page. If the seat is open, read the register.